65 min read
Reader Signal
Lower current readership momentum right now.
EDWARD
Tuesday, October 7, 2025 – 10:00 AM – Corner Coffee Shop, Brickell
The second wave of pilot selection.
Edward and Rachel had taken over a corner booth at the coffee shop on the first floor of their office building. It was nice to get out of the building to review the team applications. The morning light came through the windows, and the ambient noise gave them privacy without the sterile feeling of a conference room. Between them were twelve applications from engineering teams who wanted to participate in what Rachel had started calling “the pilot program.”
“Twelve teams,” Rachel said, spreading the applications across the table. “All of them want in. Only three got selected in the first round. The word is spreading.”
“What’s the criteria?”
“Same as before. They have to map their value stream first. Show us where work gets stuck. No vague answers, no aspirational thinking, no ‘we want to adopt AI because everyone else is.’ They need to know exactly what’s slowing them down.”
Edward picked up the first application. The Security team. Their stated outcome: “Reduce vulnerability response time from 14 days to 3 days.” Their value stream map showed a detailed flowchart: where reviews got stuck, which handoffs added delays, how much was actual work versus waiting.
“This one’s good,” he said.
“I agree. They’ve done the work. They know what’s blocking them.”
The second application was from the Mobile team. Their goal: “Improve mobile app quality using AI.” Their value stream: “Standard Agile with two-week sprints.”
“This one’s not ready,” Edward said.
Rachel nodded. “Too vague. ‘Improve quality’ doesn’t mean anything. Which quality? Measured how? And their process description is the PowerPoint version, not the real one. They haven’t actually mapped where work gets stuck.”
They went through all twelve. Four were ready. These teams had genuinely mapped their process and knew their bottlenecks. Three were close but needed more work on the value stream, which turned out to be the hardest part. Five weren’t ready at all. They wanted AI because AI was exciting, not because they had a specific bottleneck to fix.
Rachel set down two applications. “These two said they didn’t have time to create a value stream map.”
Edward chuckled. He pulled out his phone and drafted an email to their VPs. “Hi. Your team applied for the AI pilot program but said they don’t have four hours to map their value stream. I’d like to help. Can we talk about what’s consuming their time?”
He hit send.
“That’ll fix it,” he said, picking up his coffee.
Rachel laughed. “You’re terrible.”
“I’m curious. If a team doesn’t have four hours to understand their own process, that tells me something important about their process.” He shrugged. “Either they’re genuinely slammed and need help, or they’re not prioritizing this. Either way, the VP should know.”
By the time they finished their coffees, both VPs had replied. One apologized and said the team would have the value stream map by Friday. The other asked if Edward could meet with the team directly to help them find the time.
“You know what I am doing?” Edward said, watching Rachel stack the approved applications. “I’m using the AI vibes from the engineers to get my middle management to fix the process.”
“What do you mean?”
“Everyone wants AI tools. The engineers are excited about it. So I created this application process for deep AI investment, and now I’m measuring my leaders in a kind way.” He smiled. “The ones whose teams can map their value stream and articulate their bottlenecks? Those are the leaders who actually understand their operations. The ones who can’t? Now I know where to focus.”
Rachel raised an eyebrow. “So the AI pilot program is secretly a management assessment tool.”
“Not secretly. It’s both. And to be clear, I’d never block my ICs from using AI tools – do I want them to quit? Everyone gets the tools. That’s table stakes now.” He leaned back. “But the pilot program is about deep investment. Dedicated support. Process redesign. Training budgets. The teams that can articulate their bottlenecks get the extra resources. The ones that can’t?” He shrugged. “Their leaders get extra help from me.”
“Help.”
“Coaching. Mentorship. Whatever you want to call it.” He finished his coffee. “I’m not punishing anyone. I’m identifying where the organization needs development. Some leaders have their houses in order. Some have been hiding behind activity metrics for years. Now I know which is which.”
Rachel nodded. “AI is the carrot. Process improvement is the meal.”
“And leadership development is the side dish nobody ordered but everyone needs.”
“Here’s what I’m learning,” Edward said, setting down the last application. “The value stream mapping isn’t just a filter. It’s a forcing function. Teams that can’t map their process have problems they don’t know they have. The mapping surfaces the problems.”
“The Mobile team doesn’t know where their time goes. That’s why they can’t map it.”
“Exactly. And until they figure that out, no amount of AI is going to help them. They’d just be automating confusion.”
Rachel smiled. “Outcome before tools.”
“Outcome before tools.”
THE CRISIS
Thursday, October 16, 2025 – 6:47 AM – Edward’s Phone
Three weeks after the medication feature shipped.
Edward was making coffee when his phone exploded with notifications. Slack messages. Emails. A missed call from Rachel at 6:32 AM.
He called her back.
“We have a problem.” Rachel’s voice was tight. “St. Augustine Memorial. The medication tracking system.”
Edward’s stomach dropped. St. Augustine was one of their first pilot hospitals. The feature they’d shipped on Day 87. The proof that the approach could work.
“What happened?”
“A nurse accessed the system during a code blue. The medication reconciliation module locked up. Spinning wheel for eleven seconds. She had to abandon the system and do manual verification.” Rachel paused. “The patient was fine. But if she’d trusted the system and waited…”
“Eleven seconds during a code blue.”
“Eleven seconds. In a situation where every second matters.” Rachel’s voice cracked. “Edward, this is exactly what Johnny warned about. ‘When something goes wrong, and it will, remember I warned you.’“
Edward set down his coffee. His hands were steady, but his mind was racing. Forty-five days of work. The 90-day bet with the CEO. The board approval he’d fought for. All of it suddenly fragile.
“I’m coming in. Get the team together.”
Thursday, October 16, 2025 – 8:00 AM – Emergency War Room, Riverton Health Systems
The war room was silent except for the hum of laptop fans.
Edward looked at the faces around the table. Rachel. Priya Sharma, whose team had built the feature. Kevin Nakamura, who’d reviewed every line of code. Dr. Amira Patel from Compliance, looking pale.
“Walk me through what happened,” Edward said.
Priya pulled up the logs. “6:14 AM Eastern. Nurse Bridget Okafor at St. Augustine Memorial accessed the medication reconciliation screen for a patient in cardiac arrest. The system initiated a drug interaction check against 23 current medications. The query took 11.3 seconds to complete.”
“Why?”
“The HIPAA audit logging.” Priya’s voice was hollow. “When we streamlined the compliance review, we removed the three-week manual audit. We replaced it with automated logging. But the logging query wasn’t optimized for real-time access. It was designed for batch processing.”
“So we fixed one bottleneck and created another.”
“We didn’t test under load. Under normal conditions, the logging adds 200 milliseconds. Under code blue conditions, with multiple systems accessing the same patient record simultaneously, it compounds. The nurse hit a perfect storm.”
Edward closed his eyes. He’d pushed to streamline the compliance review. He’d celebrated when they cut three weeks from the process. He hadn’t asked what they were trading away.
“What’s the patient status?”
“Stable. Recovered.” Dr. Patel spoke up. “But the near-miss report is already filed. The hospital’s risk management team is asking questions. If this had gone differently…”
She didn’t finish the sentence. She didn’t need to.
Thursday, October 16, 2025 – 10:00 AM – Edward’s Office
Edward closed his door and sat alone.
The feature had shipped three weeks ago. The approach had been validated. And now, in its first real-world test, it had nearly failed at the worst possible moment.
He thought about what he would tell David Aldridge. What he would tell Sandra Williams. What he would tell the board.
The approach failed. The value stream mapping identified bottlenecks correctly. But we didn’t validate the fixes under real-world conditions. We optimized for speed and created a new risk.
His phone buzzed. Text from Jennifer.
Heard something went wrong. Are you okay?
He typed back: Not sure yet. Might have made a serious mistake.
You’ll figure it out. You always do.
He wished he shared her confidence.
Thursday, October 16, 2025 – 2:00 PM – CEO’s Office
David Aldridge listened without interrupting.
Edward laid out everything. The incident. The root cause. The near-miss. The risk that a patient could have been harmed because they’d moved too fast.
When Edward finished, David was quiet for a long moment.
“What do you want to do?”
“I want to pause the rollout to additional hospitals. Fix the logging issue. Re-validate under load conditions. Add a pre-flight checklist for any process change that touches patient-facing systems.”
“That sounds like the responsible thing to do.”
“It is. It means the expansion slows down. The board was expecting us to be in fifty hospitals by Q4. This could push that back months.”
David leaned back. “Edward, the 90-day bet was about proving an approach could work. It worked. The feature shipped. Does a bug in production mean the approach doesn’t work?”
“It means the approach has gaps. Gaps I didn’t see.”
“Then name the gap correctly,” David said. “You proved you could get a real feature out the door inside ninety days. Now you’re learning what it takes to harden it for production. Those are different proofs.”
“So now you see them. What are you going to do about it?”
Edward thought about Johnny Morrison. About the warnings he’d dismissed. About the arrogance of thinking he’d fixed problems that had defeated other transformation attempts.
“I’m going to add a step to the approach. Before any feature goes to additional hospitals, stress testing under production conditions. Not just does it work, but does it work under the worst conditions we can imagine?”
“That sounds like learning.”
“It’s learning that almost cost a patient’s life.”
“Almost.” David’s voice was firm. “Almost. The nurse caught it. The patient recovered. The system failed gracefully enough that manual backup was possible.” He leaned forward. “This is what scaling looks like, Edward. Not perfection. Learning. Fast learning. Learning from the first hospitals before we expand to fifty more.”
“Johnny would say I was reckless.”
“Johnny would say you were reckless because he’s risk-averse to the point of paralysis. You were aggressive. There’s a difference.” David paused. “The question isn’t whether you made a mistake. You did. The question is what you do next.”
Friday, October 17, 2025 – 9:00 AM – Engineering Floor
Edward gathered the team.
“I owe you an apology.”
The room went silent.
“I pushed to streamline the compliance review. I celebrated when we cut three weeks from the process. I didn’t ask hard enough questions about what we might be missing.” He looked around the room. “The logging issue was foreseeable. If we’d stress-tested under real conditions, we would have caught it. We didn’t, because I was too focused on shipping fast.”
Priya spoke up. “We all missed it. The code review didn’t catch it. QA didn’t catch it. It’s not just on you.”
“The approach is on me. The culture that said move fast, we’ll fix it later. That’s on me.” Edward walked to the whiteboard. “Here’s what changes.”
He wrote: STRESS TESTING REQUIREMENT
“Every process change that touches patient-facing systems gets stress-tested before deployment. Not synthetic load tests. Realistic scenarios. Code blues. Multi-system access. Edge cases we can imagine and edge cases we can’t.”
He wrote: PRE-FLIGHT CHECKLIST
“Before any value stream optimization ships, we answer three questions: What could go wrong under pressure? What’s our manual fallback? Have we tested the fallback?”
He wrote: NEAR-MISS REVIEWS
“Every near-miss gets a full retrospective. Not to assign blame. To find the gaps in our thinking.”
Kevin Nakamura raised his hand. “This is going to slow down the expansion.”
“Yes. It will.” Edward met his eyes. “We might not hit fifty hospitals by Q4. But every hospital we do add will be safe.” He paused. “I’d rather explain a slower rollout to the board than explain a patient death to a family.”
The room was quiet.
“Questions?”
Dr. Kapur spoke up. “What do we tell St. Augustine?”
“The truth. We identified a performance issue during their pilot. We’re fixing it. We’re implementing additional safeguards. We’re grateful the nurse followed protocol and used manual backup.”
“They might pull out of the pilot.”
“They might. That’s their right.” Edward’s voice was steady. “But I’d rather lose a pilot hospital than lose their trust.”
Friday, October 17, 2025 – 4:00 PM – Phone Call with St. Augustine
The call with St. Augustine Memorial’s CIO, Dr. Marcus Greene, lasted forty-five minutes.
Edward explained everything. The incident. The root cause. The fix. The new safeguards.
Dr. Greene listened, asked questions, and was silent for a long moment.
“Most vendors would have buried this,” he said finally. “Blamed the nurse. Called it user error. Quietly patched it and hoped nobody noticed.”
“That’s not how we operate.”
“I can see that.” Dr. Greene paused. “We’re staying in the pilot. But I have conditions.”
“Name them.”
“I want weekly updates on the fix. I want to review the stress testing protocol before you redeploy. And I want you to present this incident at our patient safety committee. Not to shame you. To show our staff what good vendor behavior looks like.”
Edward closed his eyes for half a second. “Done. All of it.”
“One more thing. The nurse, Bridget Okafor. She’s been worried she’ll get blamed for ‘not trusting the system.’ I want you to send her a personal note thanking her for following protocol. For doing the manual backup instead of waiting for the screen to load.”
“I’ll do it today.”
“Good.” Dr. Greene’s voice warmed slightly. “Most CIOs in my position would have pulled out. Most vendors in your position would have lawyered up. We’re both choosing to do this differently. Let’s see if it works.”
Saturday, October 18, 2025 – 6:00 AM – The Dock, Edward’s House
Edward sat on his dock in the pre-dawn darkness, coffee in hand, watching the first light touch the water.
The crisis had been close. Closer than he wanted to admit. A patient could have died because he’d been too focused on shipping, too confident in his approach, too eager to scale.
But the patient hadn’t died. The nurse had followed protocol. The system had failed in a way that allowed recovery.
Learning that almost cost a patient’s life.
Almost. The word hung in the air.
He thought about what Dr. Greene had said. Most vendors would have buried this. Most CTOs would have too. Blamed the team. Hidden the report. Moved on without examining their own role in the failure.
He’d been tempted. For about thirty seconds in that war room, he’d been tempted to point at the code, at the logging design, at anything except his own push to ship faster than was safe.
But he hadn’t. And that choice, the choice to own it publicly, to apologize, to change the approach, felt like the most important decision he’d made in months.
His phone buzzed. Text from Priya.
Stress testing framework ready for review. Also, Bridget Okafor at St. Augustine sent a message. She said your note made her cry. The good kind.
Edward smiled in the darkness.
The expansion would be slower now. But it would be safer.
And he’d learned something that all the value stream mapping in the world couldn’t teach: The approach doesn’t prevent failure. It makes failure visible. And what you do when failure is visible, that’s the real test.
Monday, October 20, 2025 – 10:00 AM – Riverton Boardroom
Johnny Morrison had called an emergency session of the Technology Steering Committee. Except the Technology Steering Committee no longer existed. Edward had dissolved it months ago.
So Johnny went to the board instead.
Edward found out about the meeting from Rachel, who’d heard about it from Maria Santos in Legal. By the time he reached the boardroom, Johnny was already presenting.
“The incident at St. Augustine represents exactly the kind of risk I’ve been warning about.” Johnny stood at the head of the table, a slide deck open on the screen behind him. Victoria Hale, the independent director, was present, along with David Aldridge and Sandra Williams. “Edward dissolved governance structures that existed for a reason. He pushed changes without proper oversight. And now we have a near-miss that could have killed a patient.”
Edward took a seat at the far end of the table. He didn’t interrupt.
“I’m not saying Edward’s approach is wrong in theory,” Johnny continued. “But the execution has been reckless. Moving fast is one thing. Moving fast without adequate safeguards is something else entirely.”
“What are you proposing?” Victoria asked.
“Pause the pilot program. Reinstate the governance committees. Create a proper review process before we make any more changes to patient-facing systems.”
“And the 90-day bet?”
“The 90-day bet was always unrealistic. What happened at St. Augustine proves it.” Johnny looked at Edward. “I’m not trying to destroy what you’re building. I’m trying to make sure we don’t destroy ourselves in the process.”
Victoria turned to Edward. “You’ve heard the concerns. What’s your response?”
Edward took a breath. He’d been expecting something like this ever since the St. Augustine incident. Johnny had legitimate concerns. The question was whether the response was proportionate.
“Johnny’s right that the incident at St. Augustine was serious. A patient could have been harmed. That’s unacceptable.” Edward stood and walked to the screen. “But the incident also proved something important: the system failed in a way that allowed recovery. The nurse followed protocol. The patient was fine. The near-miss process worked.”
“You’re saying the failure was acceptable because nothing bad actually happened?”
“I’m saying the failure was visible. That’s different from the old approach, where failures were hidden in process complexity until they became catastrophic.” Edward pulled up his own data. “Under the old governance structure, we had three patient-facing incidents last year. Two of them weren’t discovered until customer audits found the problems. One took four months to resolve because nobody could figure out which committee owned the decision.”
“And under your approach?”
“One incident. Discovered immediately. Root cause identified within hours. Fix deployed within a week. The nurse who caught the problem received a personal thank-you for following protocol.” Edward looked at Johnny. “The old governance didn’t prevent problems. It hid them. My approach doesn’t prevent problems either. But it makes them visible. And visibility is how you actually improve.”
The room was quiet.
Victoria leaned back. “I’m going to ask a direct question. Should we pause Edward’s pilot program?”
Johnny spoke first. “Yes. The risk is too high.”
David Aldridge spoke next. “No. The incident was handled well. The learning was real. Pausing now sends the wrong message.”
Sandra Williams hesitated. “I want to say yes. The CFO in me hates uncontrolled experiments. But…” She looked at Edward. “The old committees weren’t preventing anything. They were just generating paperwork. If Edward’s approach actually works, it’s worth the discomfort.”
Victoria nodded. “Then we continue. But with conditions.” She turned to Edward. “Johnny joins the AI-SDLC Board. His concerns get heard, not dismissed. And any changes to patient-facing systems go through a pre-flight checklist that Johnny approves.”
Edward felt his jaw tighten. Johnny on the board meant friction on every decision. It meant slower approvals, more documentation, endless debates about acceptable risk.
But it also meant Johnny’s concerns were addressed. It meant the board had confidence in the approach. It meant the 90-day bet could continue.
“Agreed,” Edward said.
Johnny looked surprised. He’d expected a fight.
“Agreed,” Johnny said. “But I want it noted that I still have concerns about the pace of change.”
“Noted.” Victoria stood. “Meeting adjourned. Edward, Johnny, I expect you to work together. Not against each other.”
They filed out of the boardroom. In the hallway, Johnny caught Edward’s arm.
“That wasn’t what I expected.”
“What did you expect?”
“You to fight. To argue. To dismiss my concerns the way you’ve been dismissing them for months.”
Edward stopped walking. “Johnny, I’m not dismissing your concerns. I never was. What happened at MedFirst was real. The consequences were real. You carry that, and you should.” He paused. “But the lesson isn’t ‘move slow and nothing bad happens.’ The lesson is ‘make failures visible before they become catastrophic.’ We’re trying to learn the same thing. We’re just learning it differently.”
Johnny was quiet for a long moment.
“I still think you’re moving too fast.”
“I know. And I still think you’re moving too slow. That’s why we need each other.” Edward started walking again. “First AI-SDLC Board meeting is Thursday. I expect you there. With concerns, with questions, with pushback. That’s your job now.”
“And if I think a decision is wrong?”
“Then you say so. Loudly. That’s how this works.”
Johnny nodded. He uncrossed his arms.
“Thursday,” he said. “I’ll be there.”
Thursday, October 23, 2025 – 2:00 PM – Engineering Floor, Riverton Health Systems
The Kevin Nakamura conversation.
Priya Sharma, team lead for Claims Processing, had been sending increasingly urgent Slack messages. One of her senior code reviewers was threatening to quit. He thought the AI pilot was going to replace him.
Edward found Kevin in his cubicle, staring at his screen with the expression of someone contemplating a funeral.
“Mind if I sit?” Edward pulled up a chair without waiting for an answer.
Kevin was fifty-two, a fifteen-year veteran of code review who could spot a security vulnerability in his sleep. He’d been at Riverton longer than Edward had, and he’d earned the respect of every engineer who’d ever submitted code to his queue.
“I know why you’re here,” Kevin said, not looking away from his screen. “Priya told you I’m being difficult.”
“Priya told me you have concerns. I wanted to hear them directly.”
Kevin finally turned to face Edward. His expression was a mixture of anger and fear.
“I’ve been doing code review for fifteen years. Fifteen years of learning patterns, building intuition, understanding what makes code good or bad. And now you’re telling me a machine can do it?”
“That’s not what I’m telling you.”
“That’s what it sounds like. AI-assisted code review. What does that even mean? The AI does the review, and I assist? Is that the future?”
Edward leaned back in his chair. He’d been expecting this conversation, not necessarily with Kevin, but with someone. The fear was real. The fear was valid. Dismissing it would be a mistake.
“How much of your time goes to obvious stuff? Syntax, style violations, common patterns?”
Kevin shrugged. “Maybe forty percent.”
“And the hard stuff? Architecture, business logic, healthcare-specific problems?”
“The rest.”
“What if AI handled the forty? So you could focus on the sixty?”
Kevin was quiet. “You’re saying it’s a filter, not a replacement.”
“I’m saying the best code reviewer I know is still Kevin Nakamura. But Kevin with AI handling the routine stuff catches twice as many real problems.”
“And what do I do?”
“Teach it what good looks like for us. Not generic good. Riverton good. The AI knows code. You know healthcare.”
Kevin’s eyes narrowed. “You’re offering me a job teaching an AI?”
“I’m offering you a promotion.”
The anger in Kevin’s expression faded slightly, but not the skepticism. Edward could see him processing, calculating.
“I’d need to understand how it works,” Kevin said. “The AI. How it makes decisions. What it catches and what it misses.”
“That’s exactly what we need. Someone who can audit the AI’s work and identify the gaps.”
“And I’d still review code?”
“The hardest stuff. The problems only a human can solve.”
Kevin was quiet. “One week. If it’s as bad as I think, I reserve the right to say I told you so.”
“Fair.”
“And I veto anything that compromises quality. No questions.”
“That’s not a condition. That’s the point.”
“One week.”
Monday, November 3, 2025 – 9:00 AM – Engineering Floor, Riverton Health Systems
Kevin Nakamura was still skeptical.
Edward heard it from Priya first: Kevin was finding every flaw in the AI’s suggestions. He’d sent three emails in the first week documenting cases where the AI missed obvious problems. One email had fourteen screenshots.
“He’s not wrong,” Priya said. “The AI misses context. It doesn’t understand our business rules. But Kevin’s treating every mistake like proof the whole thing is broken.”
Edward found Kevin at his desk, surrounded by printed code snippets covered in red ink.
“You’ve been busy,” Edward said.
“Your AI approved a null pointer dereference on Thursday. It would have crashed in production.” Kevin pointed to a highlighted section. “It also suggested removing a validation check that’s required for HIPAA compliance. The AI has no idea what it’s looking at.”
“You caught both issues.”
“I caught them because I was looking. If this goes to production without me reviewing everything, you’ll have lawsuits.”
Edward sat down. This was the conversation he’d been expecting. “How many reviews did you do last week?”
Kevin checked his notes. “Sixty-three.”
“How many before the AI pilot?”
A pause. “Maybe forty.”
“So you did fifty percent more reviews. How much of that extra capacity came from not having to flag the obvious stuff?”
Kevin didn’t answer immediately. Edward could see him running the numbers.
“The AI caught thirty-two style violations I would have flagged. Nineteen common security patterns. Eight syntax issues.” Kevin’s voice was grudging. “That’s maybe three hours of work I didn’t have to do.”
“And you used that time to catch the null pointer and the HIPAA issue. Things the AI would have missed.”
“I would have caught those anyway.”
“Maybe. But you had more time to look. More time to think. The question isn’t whether the AI is perfect; it’s whether the combination of AI plus Kevin Nakamura produces better outcomes than Kevin Nakamura alone.”
Kevin was quiet for a long moment. Then he picked up one of his printed screenshots.
“The AI got this one wrong. Suggested refactoring a loop that would have broken the medication reconciliation logic.” He set it down. “But it also caught six issues I missed on first pass. Things I would have caught eventually, but didn’t on the first review.”
“So it’s not replacing you. It’s…”
“Making me slower at some things and faster at others.” Kevin’s expression was complicated, frustration mixed with something that might have been curiosity. “I still don’t trust it. But I’m not ready to shut it down yet.”
“That’s all I’m asking. Give it another two weeks. Keep documenting the failures. But also document what’s working.”
Kevin nodded. “Two weeks. Then we talk again.”
Wednesday, November 5, 2025 – 4:30 PM – Edward’s Office, Riverton Health Systems
Rachel closed the door behind her. Never a good sign.
“Kevin’s threatening to quit. For real this time.”
Edward looked up from his laptop. “What happened?”
“You happened.” Rachel sat down across from him, her expression careful. “He told Priya you’re pushing too hard. That you’re more interested in proving the pilot works than in whether it actually works.”
“That’s not fair. I’ve been listening to every concern he raises.”
“You’ve been answering every concern he raises. That’s different.” Rachel paused. “Edward, I warned you about this two weeks ago. Kevin doesn’t need answers. He needs to feel like his expertise matters. And every time you explain why the AI is actually helping, you’re telling him his fifteen years of experience are less important than your ninety-day experiment.”
Edward felt his jaw tighten. “So what am I supposed to do? Let him sabotage the pilot because his feelings are hurt?”
“His feelings aren’t hurt. His identity is threatened. There’s a difference.” Rachel leaned forward. “And frankly, you’re not listening to me either. I told you to give him space. You scheduled three check-ins in two weeks. I told you to let him document failures without commentary. You sent him a four-paragraph email explaining why each failure was actually a learning opportunity.”
“I was trying to help.”
“You were trying to convince. It’s not the same thing.”
Edward was quiet. The feedback stung because it was accurate.
“What do you suggest?”
“Back off. Completely. Let Kevin run the evaluation his way for the next two weeks. No check-ins unless he asks. No explanatory emails. No helpful suggestions.” Rachel stood. “And maybe apologize. Not for the pilot. For not listening when he told you what he needed.”
“And if he quits anyway?”
“Then you’ll have learned something about how not to manage change. Which is still more valuable than pretending you got it right.”
She left. Edward sat alone, replaying the last two weeks. Every conversation with Kevin. Every email. Every moment where he’d been so focused on proving the approach worked that he’d stopped asking whether Kevin was okay.
You’re becoming the thing you fired the consultants for being. Confident you have the answer. Not curious about whether you might be wrong.
He picked up his phone. Typed a message to Kevin.
Can we talk? Not about the pilot. About how I’ve been handling this. I think I owe you an apology.
The reply came an hour later.
Tomorrow. 8 AM. Coffee shop on Brickell. Not the office.
Thursday, November 6, 2025 – 8:00 AM – Coffee Shop, Brickell
Kevin was already there when Edward arrived. Two coffees on the table. The gesture said something.
“I ordered for you. Black, right?”
“Thanks.” Edward sat down. “I appreciate you meeting me.”
“I almost didn’t.” Kevin stirred his coffee, not looking up. “I had my resignation letter drafted. Was going to send it yesterday afternoon.”
“What stopped you?”
“Your text. The part about owing me an apology.” Kevin finally met his eyes. “In fifteen years, I’ve never had a CTO apologize to me. Not once. Not for anything.”
“Then I’m long overdue.” Edward took a breath. “I’ve been so focused on making this pilot work that I stopped asking whether I was making it work for you. I treated you like an obstacle instead of a partner. That was wrong.”
Kevin was quiet for a moment. “You know what the hardest part has been?”
“Tell me.”
“Not the AI. The AI is just a tool. I can learn to work with tools.” He set down his coffee. “The hardest part is feeling like my judgment doesn’t matter anymore. Like the decision’s already been made, and I’m just here to validate it.”
“That’s not what I intended.”
“I know. But intention and impact aren’t the same thing.” Kevin leaned back. “Here’s what I need. I need to run this evaluation my way. No check-ins. No helpful emails. Just me, doing what I do, and documenting what I find. And at the end, if I say it’s not ready, you listen. Really listen. Not ‘listen and then explain why I’m wrong.’“
“What if you’re right and it’s not ready?”
“Then we shut it down. Or we fix it. Or we try something else.” Kevin shrugged. “But the answer has to come from looking honestly at the data, not from deciding in advance what we want the answer to be.”
Edward recognized his own words being reflected back at him. The approach he’d been preaching to everyone else, applied to himself.
“Fair,” he said. “Two weeks. Your way. And whatever you find, I’ll listen. Really listen.”
“Okay then.” Kevin picked up his coffee. “And Edward?”
“Yeah?”
“Thank you. For actually hearing me this time.”
They finished their coffee. Talked about Kevin’s daughter’s wedding, about the Dolphins’ chances next season, about nothing related to AI or pilots or transformation. Just two people remembering they were more than their roles.
Walking back to the office, Edward called Rachel.
“You were right. I almost lost him.”
“I know.”
“How did you know?”
“Because I’ve watched you for six months. You’re good at asking questions about process. You’re terrible at asking questions about people.” She paused. “The consultants you fired? They had the same blind spot. All frameworks, no empathy. Don’t become them.”
“I’m trying not to.”
“Try harder.”
She hung up. Edward walked the rest of the way in silence, thinking about the difference between being right and being effective.
Friday, November 14, 2025 – 3:00 PM – Rachel’s Office, Riverton Health Systems
The early results came in through November.
Claims Processing saw 40% faster code review cycles in the first month. But the number that mattered more to Edward was the quality metric: production bug rate stayed flat. The AI wasn’t just making things faster; it was maintaining the standard Kevin Nakamura had spent fifteen years building.
Kevin himself had become something unexpected: the pilot’s most rigorous quality auditor. He still documented every AI failure, but now he also documented the successes. His weekly reports to Edward showed a system slowly earning his trust, one review at a time.
“It’s not perfect,” Kevin said during their third check-in. “But it’s getting better. And I’m catching things I never had time to look for before.”
It wasn’t a conversion. It was something more durable: a skeptic becoming a critical collaborator.
And then there was Priya Sharma’s “toil repayment” metric, the ratio of time saved on repetitive work to time reinvested in high-value work. Her team had moved from 60% toil to 35%, and that freed-up 25% was going straight into the backlog.
Into the medication tracking feature.
The one that had been sitting there for eighteen months while hospitals kept asking for it. The one that could flag potential drug interactions before discharge. The one Priya’s sister, a nurse, had told her about, a patient at her hospital who’d had a bad interaction because the discharge paperwork didn’t catch a conflict.
That feature was already live in three hospitals. Real people would be safer because Priya’s team had time to build what mattered.
Friday, November 21, 2025 – 2:00 PM – Engineering Floor
A month after the St. Augustine crisis, the expansion had resumed.
The stress testing framework was in place. Every hospital now went through a rigorous validation before going live. It added a week to each deployment, but the team understood why.
Edward found Kevin Nakamura at his desk, reviewing the latest batch of AI-generated code suggestions.
“How’s the new protocol working?” Edward asked.
Kevin looked up. “The stress testing caught two issues this week. Both would have caused problems in production.” He shrugged. “I still don’t love the AI. But I trust the process.”
“That’s all I’m asking.”
Kevin turned back to his screen. “Also, the feature is in seven hospitals now. One of the nurses at Tampa General sent a message to Priya. Said the system flagged a potential interaction she might have missed.”
Edward nodded. That was why they did this. Not the productivity metrics. Not the AI adoption rates. Real outcomes.
“Keep documenting everything,” he said. “The good and the bad.”
“Always do.”
Monday, November 24, 2025 – 10:00 AM – Riverton Health Systems
Provider Portal finished their process mapping and discovered something unexpected. Three of their review stages were redundant, holdovers from a compliance requirement that had been changed two years ago. They eliminated them before deploying any AI tools, cutting their delivery time by 30%.
Documentation reduced their turnaround from six days to three, without implementing a single new tool. Just by answering the third question honestly, they’d discovered inefficiencies nobody had noticed.
Rachel brought the results to Edward’s office on a Friday afternoon.
“The pattern is consistent,” she said, spreading the data across his desk. “Teams that map their value stream are seeing gains. Teams that can’t map it are getting stuck, but they’re getting stuck on problems they needed to address anyway.”
“The mapping is the intervention.”
“Exactly. The AI is accelerating what they already know how to do. But understanding the process is what surfaces what they need to know.”
Edward went through the numbers. They were good. Maybe too good. Good enough that people would start noticing.
“Sandra Williams wants a meeting,” he said. Sandra was Riverton’s CFO. a by-the-numbers finance executive who kept close tabs on anything that affected budget or risk.
“About the pilots?”
“About ‘unauthorized technology experimentation.’ Those were her words.” Edward smiled grimly. “The board asks her about AI strategy. She doesn’t have an answer. She hears rumors about pilots happening without board approval. She puts two and two together.”
“Are you worried?”
“I’m cautious.” He looked at the data again. “But I’m also holding a stack of results that are hard to argue with. If Sandra wants to talk about AI strategy, I can show her what works. Not theory, evidence.”
Wednesday, November 26, 2025 – 11:00 AM – Sandra Williams’s Office, Executive Floor
The Sandra Williams meeting.
Sandra’s office was on the executive floor, with windows overlooking Biscayne Bay. She kept it sparse: a desk, two chairs, a single framed photo of her MBA class at Northwestern. She was notoriously direct.
“Close the door,” she said when Edward arrived.
Edward closed it and sat down.
“The board meeting is next week,” Sandra said. “They’re going to ask about AI. They always ask about AI now. It’s the only thing anyone wants to talk about.” She folded her hands on her desk. “I need to know what you’re doing.”
“I’m running pilots. Small scale, low risk, documented results.”
“Without board approval.”
“Within my discretionary authority as CTO. Nothing I’ve done requires board approval.”
Sandra’s expression didn’t change. “That’s a fine line.”
“It’s a line I’ve been careful not to cross.” Edward leaned forward. “Sandra, let me show you what I have.”
He spent the next twenty minutes walking her through the results. Claims Processing, 40% faster, no quality degradation. Provider Portal, 30% improvement from process optimization alone. Documentation, turnaround cut in half. All of it documented, all of it measurable, all of it tied to a repeatable approach.
Sandra listened without interrupting. When Edward finished, she was quiet for a long moment.
“Value stream mapping,” she said finally.
“That’s where every team starts. Before we talk about tools, they have to map their actual process. Where does work wait? Where does it get stuck? What’s really slowing them down?”
“And the teams that can’t map it?”
“Don’t proceed. Not as punishment, but as protection. If you don’t know what’s blocking you, no amount of AI will help you.”
“What’s your goal, Edward? What’s Riverton trying to accomplish?”
Edward didn’t hesitate. “Build better software, faster and safer, so we can actually impact patient lives. We had a medication tracking feature sitting in the backlog for eighteen months. Hospitals kept asking for it. It could prevent drug interaction errors at discharge. Real patients, real safety.” He paused. “That feature is finally being built. Not because of AI. because our team has time to work on what matters now.”
Sandra raised an eyebrow. “The goal doesn’t mention AI at all.”
Edward laughed. “I know. Kind of ironic, isn’t it? I might be kissing my consulting career goodbye by admitting it.” He shrugged. “But that’s the point. The AI is just a means. The goal is patient lives.”
Sandra spread the data sheets across her desk. Her expression was unreadable.
“The board is going to ask what you’ve been doing,” she said. “They’re going to want to know why you didn’t ask permission.”
“What would you like me to tell them?”
Sandra looked up. For the first time, approval flickered across her face.
“Tell them you were proving it works. Tell them you have evidence, not theory. Tell them the process improvement is the innovation, not the AI.” She gathered the data sheets and stacked them neatly. “And then tell them you’re ready to scale.”
Edward blinked. “You’re supporting this?”
“I’m supporting results. These results.” She handed the stack back to Edward. “Don’t make me regret it.”
Monday, December 8, 2025 – 9:00 AM – CEO’s Office, Executive Floor
The Johnny Morrison confrontation.
Edward walked into the CEO’s office expecting a one-on-one with David Aldridge. Instead, Johnny Morrison was already seated, binder open, expression triumphant.
“Edward, please sit,” David said. His tone was neutral, carefully so. “Johnny has raised some concerns about your pilots.”
Johnny didn’t wait. “You’ve been running unauthorized AI experiments for four months. You dissolved the Architecture Review Board. You eliminated the Security Council. You created some kind of rogue governance structure without my approval.”
“The AI-SDLC Board. It’s not rogue. It’s faster.”
“You didn’t go through my transformation office. You didn’t follow the governance process I put in place.”
“I replaced it. One board. Twenty-four-hour decisions. Any engineer can ask for an answer and get one. That’s not undermining governance. That’s making it work.”
Johnny’s jaw tightened. “What’s not interpretation is that you’ve undermined a company-wide transformation initiative. My initiative. The one the board approved and funded.”
David held up a hand. “Let’s keep this constructive. Edward, what results are you seeing?”
Edward spent fifteen minutes walking through the data. Claims Processing, Provider Portal, Documentation. The toil repayment metrics. The medication tracking feature finally moving. He watched David’s expression shift from neutral to interested.
Johnny countered with his own presentation. Prompt coaching workshop attendance. AI Readiness Assessment completion rates. Center of Excellence membership growth.
“These are participation metrics,” Edward said. “Not outcome metrics. How many features shipped faster? How many bugs prevented? What’s the actual impact on the business?”
Johnny’s face reddened. “We’re building organizational capability. That takes time. You can’t measure transformation in weeks.”
“You can if you’re doing it right. My pilots have been running for four months. Johnny’s program has been running for eight. I have outcome data. He has attendance sheets.”
David was quiet for a long moment. Then he turned to Johnny.
“Johnny, I appreciate what you’ve built. The governance is important. But Edward’s raising a fair question. Where are the outcomes?”
“The outcomes are coming. We need to complete the assessment phase before. “
“Your assessment phase started eight months ago. Edward’s pilots started four months ago. He’s shipping results. What are you shipping?”
Johnny looked like he’d been slapped.
“David, you can’t compare. “
“I can. I have to. The board is going to ask me about AI strategy next week. I can show them Edward’s results or your attendance metrics. Which one do you think they want to see?”
The meeting ended with David asking both of them to prepare board presentations. But Edward could see the writing on the wall. Johnny’s $400,000 program had eight months of workshops and zero measurable outcomes. Edward’s $50,000 pilots had four months of experiments and a 40% improvement in delivery time.
The data spoke for itself.
Wednesday, December 24, 2025 – 2:00 PM – CEO’s Office, Executive Floor
The reassignment meeting.
Edward walked in expecting a debrief. Instead, David Aldridge was seated with Johnny Morrison on one side and the CHRO on the other.
“Gentlemen,” David began, “the board has reviewed both approaches to AI transformation. They’ve made a decision.”
Johnny straightened in his chair, confident.
“Effective immediately, all AI initiatives, pilots, governance, and strategy, will report to the CTO. Johnny, your transformation office will be reassigned under Edward’s organization.”
The silence lasted three seconds. Then Johnny exploded.
“You’re putting me under the CTO?” His voice cracked with disbelief. “I was brought in to lead transformation. I have a direct line to the board.”
“You had a direct line,” David said. “The board wants results. Edward is producing results. This is about alignment, not demotion.”
“Alignment?” Johnny stood up, face red. “I’ve given PowerPoint decks to half the Fortune 500. I’ve led transformations at companies three times this size. And you’re telling me to report to someone who’s been running unauthorized experiments in a back room?”
Edward stayed silent. There was nothing to say.
“This is a mistake,” Johnny continued. “You can’t just let engineers run wild with these tools. There needs to be governance. There needs to be control. When this falls apart, and it will fall apart, you’ll regret this.”
David stood. “Johnny, I think you should take some time to consider your options.”
“My options?” Johnny laughed bitterly. “My option is to watch someone with no transformation experience dismantle everything I built. My option is to explain to my network why I’m suddenly reporting to the CTO instead of the CEO. My option is to pretend this isn’t a humiliation.”
“That’s not what this is. “
“That’s exactly what this is.” Johnny gathered his binder, the same 52-page AI Readiness Assessment he’d been carrying for eight months. “I’ve given PowerPoint decks to half the Fortune 500. I’ve been in rooms you’ll never see. And this is how you treat me? You’ll regret this. All of you.”
He walked out. The door slammed behind him.
David turned to Edward. “I’m sorry you had to see that.”
“He’s not wrong about the need for governance.”
“He’s not. But he’s wrong about what governance means.” David rubbed his temples. “HR will handle the transition. I expect his resignation by end of week.”
Friday, January 2, 2026 – 10:00 AM – HR Conference Room
Johnny Morrison’s resignation was formalized that morning.
Edward wasn’t in the room when it happened, but Rachel told him later. The severance package was generous; the CEO didn’t want a lawsuit or a viral post about “toxic leadership.” Johnny took the money and left quietly.
His parting words, delivered through the HR director: “You can’t just let engineers run wild with these tools. There needs to be governance. There needs to be control. When something goes wrong, and it will, remember I warned you.”
Edward thought about that for a long time. Johnny wasn’t wrong about the need for governance. He was wrong about what governance meant.
Johnny’s version of governance was stopping everything until a perfect plan existed. Edward’s version was understanding what was already happening and building guardrails around reality.
They’d both wanted top-down change. They’d both wanted control. They’d just had fundamentally different ideas about how to get there.
Johnny wanted to stop the flood. Edward wanted to build channels for the water.
The flood always wins.
But Johnny’s warning stayed with Edward: When something goes wrong, and it will, remember I warned you. The man had seen transformation failures up close. He’d watched careers end and companies bleed. His solution was wrong. You can’t stop technological change by banning it. But his fear wasn’t irrational.
Edward made a note to himself: Don’t let winning make you forget what Johnny got right. Ungoverned experimentation kills. Governance has to be real, not just a box to check.
Thursday, January 8, 2026 – 2:00 PM – Riverton Boardroom
The board meeting went better than Edward expected.
He presented the pilot results with Sandra standing beside him, a visual signal that Finance was on board. He walked them through the approach, the selection criteria, the early wins. He showed them Kevin Nakamura’s transformation from skeptic to champion. He showed them the medication feature, shipped in 87 days after sitting stuck for 18 months.
“This isn’t a technology initiative,” he concluded. “It’s a process improvement initiative. The AI accelerates some bottlenecks. But most of what we fixed had nothing to do with AI. We just finally looked at where work was getting stuck and fixed it.”
The questions were skeptical but not hostile. How do we scale? What are the risks? What happens when something goes wrong?
Edward had answers for all of them. Not because he’d planned for every scenario, but because he’d done the work. He’d lived through the Kevin Nakamura conversation. He’d seen the Documentation team eliminate redundant processes. He’d watched Provider Portal discover problems they didn’t know they had.
By the end of the meeting, he had approval to expand. Six more teams in Q3. A dedicated budget line. A mandate to document the approach for enterprise-wide adoption.
Walking back to his office, Edward loosened his tie. He pulled up the medication tracking dashboard and watched the numbers tick.
The medication feature was live. Priya’s sister was using it. Nurses across the country would be using it soon. Real patients, safer because they’d finally asked: what’s stopping us from shipping?
Monday, January 12, 2026 – 10:00 AM – Infrastructure Team War Room
The Infrastructure team’s pilot failed.
Edward found out from Rachel, who found out from the team lead, who had spent the weekend watching their AI-assisted deployment pipeline corrupt three staging environments.
“They mapped their value stream,” Rachel said, standing in Edward’s doorway. “Their goal was clear: faster deployments with fewer rollbacks. Their bottlenecks were documented. They knew where work got stuck.”
“So what happened?”
“They skipped the validation step. The approach was to pilot small, gather data, iterate. They got excited about early results and pushed to production infrastructure too fast.” Rachel set an incident report on his desk. “Twelve hours of downtime on the staging systems. The CFO already knows.”
Edward felt his stomach drop. This was exactly what Johnny had warned about. When something goes wrong, and it will, remember I warned you.
The next three days were damage control. Meetings with Sandra Williams. A presentation to the CEO explaining what happened and why. A retrospective with the Infrastructure team that revealed the real problem: the team lead had been so eager to show results that he’d ignored the “start small” guidance.
“The mapping worked,” Edward told Sandra. “They knew what they were trying to accomplish. But knowing the bottlenecks doesn’t prevent human error. It doesn’t prevent people from getting overconfident.”
“So the approach failed.”
“The approach surfaced what went wrong. Without it, we’d still be guessing.” Edward pulled up the incident timeline. “Look, they documented everything. The moment they deviated from the pilot parameters, it’s visible. We can learn from this.”
Sandra flipped a page. “The board isn’t going to see it that way.”
“Then help me frame it correctly. This is proof that the approach works, not because it prevents all failures, but because it makes failures visible and learnable.”
The emergency board update happened that Friday. Edward presented the failure alongside the successes, full transparency, no spin. Three board members pushed back hard. One suggested shutting down the pilots entirely.
But Sandra surprised him.
“Every transformation has setbacks,” she said. “The question isn’t whether we have failures; it’s whether we learn from them. Edward’s approach didn’t prevent this incident. But it did make the incident understandable. That’s more than most transformation programs achieve.”
The pilots continued. But Edward went home that weekend knowing the easy wins were over. Now came the hard part: scaling something that could fail, with real consequences, and learning fast enough to stay ahead of the failures.
He sat on his dock that Saturday night. 5 AM, his usual time now, notebook in hand, and wrote: The approach doesn’t prevent failure. It makes failure visible. That’s the difference between transformation theater and actual process improvement.
Wednesday, January 14, 2026 – 3:00 PM – Analytics Team Review
The Analytics team’s pilot results came in that week. Not a failure. Not a success. Just… underwhelming.
Rachel presented the numbers in Edward’s office. “They mapped their value stream. Goal was clear: faster report generation for hospital administrators. They identified the bottlenecks. Implemented AI-assisted data analysis.”
“And?”
“Twelve percent improvement. Cycle time down from eight days to seven.”
Edward read the report. The Infrastructure team had been a disaster. Kevin’s team had been a triumph. But this was something harder to categorize: a team that had done everything right and gotten mediocre results.
“What happened?”
“Nothing happened. They just… didn’t have that much waste to remove.” Rachel set down the report. “Their process was already pretty good. The bottlenecks they found were real, but small. AI helped with some tasks, but the gains were incremental.”
“So the approach failed?”
“The approach worked perfectly. It just revealed that this team didn’t have a dramatic transformation waiting to happen.” Rachel paused. “Not every team is sitting on eighteen months of stuck features. Some teams are just… okay. And ‘okay’ doesn’t become ‘amazing’ just because you map the value stream.”
Edward thought about this for a long time after Rachel left. The medication feature had been a home run precisely because so much waste had accumulated. Kevin’s team had been a success because years of dysfunction had created massive improvement potential.
But the Analytics team was a reminder: process improvement only finds what’s there to find. If a team is already reasonably efficient, the gains will be reasonable, not revolutionary.
Not every transformation is dramatic, he wrote that night. Sometimes you do everything right and get 12%. That’s still 12%. But it won’t make anyone’s keynote speech.
Monday, January 12, 2026 – 2:00 PM – Post-Board Debrief, Edward’s Office
An industry analyst named Rebecca Harmon had been observing the board meeting as part of a healthcare IT benchmarking study. She caught Edward in the hallway afterward.
“Impressive results,” she said, falling into step beside him. “But I have to be honest. I’m skeptical.”
“Most people are. What’s your concern?”
“Value stream mapping. Process improvement. Looking at where work gets stuck.” She shook her head. “This is elementary, Edward. Every Lean Six Sigma workshop teaches this. Every DevOps maturity model includes it. I’ve seen a hundred companies do exactly what you’re describing.”
“And?”
“And most of them fail. Not because the approach is wrong, but because organizations can’t sustain simple practices.” Rebecca stopped walking. “So what makes you different? What makes Riverton succeed where everyone else has failed?”
Edward thought about the question. It was a fair one.
“Maybe nothing makes us different,” he said. “Maybe we’ll fail too. But here’s what I’ve learned: the reason consultants sell complex frameworks isn’t that transformation is complicated. It’s that doing simple things consistently is hard. The framework is a crutch for discipline.”
“You’re saying the entire transformation industry exists because organizations lack discipline?”
“I’m saying the transformation industry has an incentive to make simple things seem impossible without expensive help. ‘Map your value stream and fix what’s broken’ doesn’t justify a $5 million consulting engagement. ‘Implement our proprietary 36-month AI Readiness Framework’ does.”
Rebecca smiled slightly. “That’s a cynical view of my industry.”
“It’s an honest one. Look at Johnny Morrison. He had eight months and $400,000 for workshops about AI readiness. Zero features shipped faster. We had four months and $50,000 for actual pilots. The medication feature shipped in eighty-seven days after being stuck for eighteen months.” Edward shrugged. “The approach is elementary. The discipline to execute it isn’t. That’s the only secret.”
“And if the discipline fails? If Riverton backslides into old patterns?”
“Then we’ll have failed at something simple. Which is more honest than failing at something complex.” He paused at his office door. “Write whatever you want in your report. But don’t call this innovation. It’s just process improvement with the courage to actually look at the process.”
Rebecca left looking thoughtful. Edward wondered if she’d understood what he was really saying: that the hardest part of transformation isn’t the frameworks or the tools. It’s having the discipline to do what you already know you should do.
Elementary. But not easy.
Tuesday, January 13, 2026 – 9:00 AM – Leadership Team Meeting, Riverton Health Systems
Edward looked around the conference table at his leadership team. Twelve people. VPs, directors, senior managers. The people who would make or break the scaling effort.
“I have a new requirement,” he said. “Starting this week, every leader in this room needs to start building with AI tools. Not reviewing. Not approving. Building.”
The silence was immediate.
“I don’t mean you need to become engineers,” Edward continued. “But you need to understand, viscerally, not just theoretically, what these tools can do. You need to experience the capability for yourself. Otherwise you’re making decisions about something you don’t understand.”
Linda Park, VP of Platform Engineering, nodded. She’d been coding for twenty years. This wouldn’t be a stretch for her.
But others looked uncomfortable.
“Edward.” It was Patricia Okonkwo, Director of QA. She’d been with Riverton for eleven years, had built the testing organization from five people to forty. “I need to be honest with you. I’ve never written a line of code in my life.”
The admission hung in the air.
“My background is quality management,” Patricia continued. “Process design. Test strategy. I know how to build a QA organization. But I’ve never been a developer. I don’t know Python or JavaScript or any of it.”
“Then this is your chance to learn.”
“I’m fifty-two years old. I’ve spent thirty years building expertise in quality management. And now you’re telling me that expertise doesn’t matter?”
“I’m telling you the expertise matters more than ever, but only if you understand the new context.” Edward leaned forward. “Patricia, your knowledge of testing strategy is invaluable. But AI is changing what’s possible. If you don’t understand the tools, you can’t lead a team that uses them.”
“So learn to code or get out?”
“Learn to build or get left behind. There’s a difference.” Edward leaned forward. “I’m not asking you to become a software engineer. I’m asking you to spend two hours a day for the next month using these tools. Build something small. A script. A prototype. Anything. So you understand what your team is experiencing.”
Rachel Torres spoke up. “I’ve been doing this for six weeks. It’s uncomfortable at first. But Patricia, the testing applications are incredible. AI can generate test cases, find edge cases, write automation scripts. Your expertise in knowing what to test becomes even more valuable when the how gets easier.”
Patricia didn’t look convinced.
Mike Donovan, the Scrum Master lead, raised his hand. “Edward, what about us? The Agile coaches, the scrum masters? We don’t write code either. We facilitate. We coach. We remove blockers.”
“What blockers are you removing that AI couldn’t handle faster?”
Mike opened his mouth, then closed it.
“I’m not trying to eliminate your jobs,” Edward said. “I’m trying to help you see where you add value that AI can’t replace. But you can’t see that clearly if you don’t understand what AI can do.”
The meeting ended with assignments. Every leader would spend two hours daily for the next month building something, anything, with AI tools. Edward would check in weekly. Those who couldn’t or wouldn’t engage would need to have a different conversation.
Friday, January 23, 2026 – 4:00 PM – Edward’s Office
The first two weeks were revealing.
Linda Park thrived. She’d started using Copilot for her own coding projects and reported that it felt “like having a senior engineer looking over my shoulder, but one who never gets impatient.”
Tom Bradley, the Chief Architect, had automated three documentation tasks he’d been procrastinating on for months. “I finally understand why the junior engineers are so excited,” he told Edward. “This isn’t about writing code faster. It’s about removing the friction that makes you not want to start.”
But others struggled.
Mike Donovan and the Agile coaches were lost. They’d spent years perfecting facilitation techniques, retrospective formats, sprint planning rituals. Now they were being asked to write Python scripts, and the dissonance was profound.
“I don’t know what I’m supposed to build,” Mike admitted in their one-on-one. “My job is helping teams work together. That’s not something you can code.”
“What if it is?” Edward asked. “What if AI could handle the scheduling, the note-taking, the action item tracking, all the logistics you currently manage? What would be left?”
“The human stuff. The conversations. The coaching.”
“Exactly. That’s where you add value. But you need to understand what AI can automate so you can focus on what it can’t.”
Mike left looking thoughtful, if not convinced.
Patricia Okonkwo was the hardest case.
She’d tried. Edward could see she’d tried. But after two weeks, she came to his office and closed the door.
“I can’t do this.”
“Can’t or won’t?”
“Can’t.” Her voice was tired. “I’ve spent thirty hours over the past two weeks staring at tutorials, watching videos, trying to write basic scripts. I managed to get a ‘Hello World’ program to run. That’s it. That’s all I have to show for thirty hours.”
“That’s a start.”
“It’s not enough. And we both know it.” Patricia sat down heavily. “I’m good at what I do, Edward. I’ve built quality processes that have caught thousands of bugs. I’ve trained dozens of testers. I’ve designed test strategies for products that handle patient data for millions of people.”
“I know. Your work has been essential.”
“But I can’t learn to code in a month. I can’t learn to code in a year. Not at the level where I’d actually understand what my team is doing.” She met his eyes. “So what happens now?”
Edward was quiet for a long moment.
“What do you want to happen?”
“I want to keep doing what I’m good at. But I can see where this is going. The manual testing team is already shrinking. The automation engineers are doing things I don’t understand. My QA directors are talking about AI-generated test suites, and I nod along like I know what they mean, but I don’t.”
“Then let’s find a role where your expertise matters and the coding doesn’t.” Edward pulled up an org chart. “Customer Success. Implementation. Training. You know our product better than almost anyone. You know quality better than anyone. Those skills translate.”
“You’re moving me out of QA.”
“I’m moving you to where you can succeed. The alternative is watching you struggle with something that doesn’t play to your strengths while the organization moves in a direction you can’t follow.”
Patricia was quiet for a long time.
“I appreciate you being honest,” she finally said. “Most executives would just wait for me to fail and then fire me.”
“That’s not how I operate. And for what it’s worth, this isn’t a demotion. Customer Success is critical. We need someone who understands quality, who can help hospitals implement our software correctly, who can train nurses to use features like the medication tracker. That’s you.”
Patricia nodded. “When?”
“Let’s take a month to transition. Find the right person to take over QA. someone technical who can grow into the role. And let’s make sure Customer Success is set up for you to succeed.”
After she left, Edward sat alone in his office, staring out at Biscayne Bay.
This was the part nobody talked about. The human cost. Patricia Okonkwo had given eleven years to Riverton, had built something real, and now the ground had shifted beneath her. She wasn’t wrong or bad or lazy; she was just in a role that required skills she’d never needed before.
How many Patricias were there across the organization? Across the industry? Good people, experienced people, who’d built careers on expertise that was suddenly insufficient?
He thought about Harold again. The tape library guy. Harold hadn’t failed either; the world had just changed faster than he could adapt.
The difference was that Edward could see it happening and try to find soft landings. Not everyone would be that lucky.
Saturday, February 7, 2026 – 9:00 PM – The Dock, Edward’s House on the Intracoastal
Edward sat on his dock with Jennifer, watching the stars emerge over the Intracoastal.
“The board approved expansion,” he said.
“I know. You texted me. Three exclamation points.” She smiled. “You never use exclamation points.”
“It felt like an exclamation point kind of moment.”
“So what happens now?”
Edward was quiet for a moment, listening to the water lap against the pilings.
“Now I prove it scales. Now I document the approach so other teams can use it. Now I take what started as a quiet experiment and turn it into something that changes how the whole organization thinks about process improvement.”
“That’s a big responsibility.”
“I know.” He looked at her. “I’m also scared. Every morning, I wake up wondering if this is the day something goes wrong. If this is the day I find out the results were a fluke, or the approach doesn’t work at scale, or I’ve been fooling myself the whole time.”
“But you keep going.”
“But I keep going. Because the alternative is worse. The alternative is watching the world change and being too afraid to change with it.” He thought about Marcus. confident Marcus, who was preparing for an IPO while the ground shifted beneath him. “I’d rather fail trying than succeed at standing still.”
Jennifer didn’t reach for his hand. Her expression had shifted, something harder in it now.
“You missed Emma’s recital.”
Edward blinked. “What?”
“Last Tuesday. Her fall piano recital. The one she’s been practicing for since August.” Jennifer’s voice was flat. “You said you’d be there. Then Rachel called about another crisis, and you said you’d catch the second half, and then you didn’t come at all.”
“I told her I was sorry. I explained.”
“You explained to me. You texted her.” Jennifer turned to face him fully. “She played Chopin, Edward. The piece she’s been working on for three months. She looked for you in the audience. Twice. Lily told me.”
He’d known. He’d known when he made the choice to stay at the office. He’d told himself it was temporary. Emergency. That Emma would understand.
“The medication feature is saving lives.”
“I know. You’ve told me. You’ve told everyone.” Jennifer’s voice cracked. “But you used to say the same thing about being present for your family. About not becoming the kind of executive who sacrifices everything for work.”
“This is different.”
“Is it?” She stood up. “You were up at 5 AM every day this month. You missed three dinners last week. You didn’t even know Lily had a math test on Friday. You asked her about her ‘spelling test’ and she just looked at you.”
Edward opened his mouth to defend himself. Closed it.
“The transformation is working. That’s wonderful. Really.” Jennifer turned her wine glass by the stem. “But Marcus isn’t the only one protecting something without examining it. When’s the last time you mapped the value stream of this family? Where does work wait? Where does it get stuck?”
The question landed like a blow.
“That’s not fair.”
“No, it’s not. None of this is fair.” She sat back down, but further away now. “I’m proud of you. I am. But I’m also tired of being proud of someone who’s never here. Emma’s going to be sixteen next year. Lily’s growing up. And you’re shipping features.”
They sat in silence. The water lapped against the pilings.
“I don’t know how to do both,” Edward finally said. “That’s the truth. I don’t know how to be the CTO this organization needs and the father my kids deserve. I keep thinking I’ll figure it out after the next milestone. After the board approval. After the expansion. After.”
“‘After’ sounds familiar.”
Edward winced. “That’s not fair either.”
“Maybe not.” Jennifer reached over and took his hand, finally. “But I’d rather have this conversation now than watch you become someone you don’t recognize. The way Marcus is becoming.”
“I’m not Marcus.”
“No. But you’re closer than you think.” She squeezed his hand. “Promise me something. Not grand promises. Just one thing. Emma’s winter concert is December 12th. Put it in your calendar. In ink. And be there.”
“I’ll be there.”
“Not ‘I’ll try.’ Not ‘unless something comes up.’ Be there.”
Edward thought about the St. Augustine crisis. About the twelve-hour days. About the board meeting and the conference keynote and the scaling plan that consumed every waking hour.
“I’ll be there,” he said. “I promise.”
Jennifer leaned her head on his shoulder. “That’s the Edward I married. The one who remembers what matters.”
They sat in silence, watching the lights of boats move on the Intracoastal. The night was warm, the air thick with the salt smell of low tide.
“I love you,” Jennifer said. “Even when you’re obsessed with process improvement.”
Edward laughed, but it was softer now, more tired. “Even then?”
“Especially then. Just don’t let it eat you alive.”
Friday, February 13, 2026 – 4:00 PM – Riverton Health Systems
By mid-November, Edward’s pilots had produced consistent results. The value stream mapping approach was spreading organically. teams who hadn’t been selected were doing it anyway, mapping their own processes, preparing themselves for the next round.
Word was starting to spread beyond Riverton. Rachel had been invited to speak at a healthcare IT conference. A vendor had asked for an interview about “the process improvement approach that’s changing Riverton’s culture.”
And the medication tracking feature was now in 23 hospitals. Already catching drug interactions that might have been missed. Real patients, safer.
Something was building. Edward could feel it.
Monday, February 16, 2026 – 10:00 AM – DevOps Team War Room, Riverton Health Systems
The CI/CD breakthrough happened almost by accident.
Rachel Torres brought Edward to the DevOps team’s war room. a converted conference room covered in architecture diagrams and deployment pipeline visualizations.
“Remember the environment provisioning problem we fixed in February?” Rachel asked. “Two-week wait time reduced to two hours?”
“That was one of our first wins.”
“Well, it opened a door nobody expected.” She gestured to Marcus Okafor, the DevOps lead. a different Marcus, younger, more technical than the one Edward had watched dismiss AI tools. “Tell him what happened.”
Marcus Okafor pulled up a screen full of metrics. “We mapped the entire deployment pipeline. From code commit to production. Everything.”
“How long?”
“Forty-seven days. Average. For a single feature to go from ‘code complete’ to ‘running in production.’“
Edward stared at the number. “Forty-seven days?”
“And that’s after the development work is done. That’s just the deployment pipeline. CI builds, test environments, staging, compliance gates, production rollout.” Marcus Okafor shook his head. “We’ve been talking about CI/CD for eight years. Leadership approved a ‘continuous deployment initiative’ in 2017. Never shipped. Approved again in 2019. Partially shipped. The joke on the team was that we had ‘continuous meetings about continuous deployment.’“
“What changed?”
“The value stream map.” Marcus Okafor pulled up the diagram. “When we actually mapped where time went, we found thirty-two separate handoffs between code commit and production. Thirty-two. Some of them were automated. Some were manual. Some were just… waiting. Waiting for someone to click a button. Waiting for a meeting. Waiting for approval from someone who’d been on vacation.”
“And AI helped?”
“AI helped with maybe a third of it. We used AI to generate the pipeline-as-code configurations we’d been hand-writing for years. We used AI to auto-generate test fixtures and deployment scripts. We used AI to document the runbooks nobody had ever documented.” He paused. “But the bigger wins were just removing waste. We found approval gates that existed because of incidents from 2018 that nobody remembered. We found manual checks that were duplicating automated checks. We found entire staging environments that hadn’t been used in two years but were still part of the ‘official’ deployment process.”
“So what’s the number now?”
Marcus Okafor smiled. “Seven days. Average. From code complete to production.”
Edward did the math. “That’s an 85% reduction.”
“And it’s not magic. It’s not some revolutionary new tool. It’s just actually looking at the pipeline and fixing what was broken.” He pulled up another screen. “The really embarrassing part? Half of what we fixed was stuff we’d been complaining about for years. Stuff everyone knew was broken. We just never had permission to fix it because ‘the deployment process is too critical to experiment with.’“
“Permission from whom?”
“Ourselves, mostly. Leadership would say ‘focus on features, not infrastructure.’ DevOps would say ‘we can’t change the pipeline during feature freeze.’ And the pipeline kept getting slower, and nobody looked at why.” Marcus Okafor shrugged. “The value stream mapping gave us permission. Once we could show the actual numbers. forty-seven days, thirty-two handoffs. suddenly everyone wanted to fix it.”
Edward thought about all the DevOps initiatives that had been proposed and rejected over the years. The consulting engagements. The maturity assessments. The “transformation roadmaps” that promised CI/CD in eighteen months.
“You did this in three months,” he said.
“Twelve weeks. With a team of four.” Marcus Okafor looked almost embarrassed. “We’d been told for years that true CI/CD was ‘too complex’ for healthcare IT. Too many compliance requirements. Too much risk. Needed a ‘phased approach’ with ‘proper governance.’” He shook his head. “Turns out it just needed someone to map the actual process and fix what was broken.”
Rachel caught Edward’s eye. This was the pattern. over and over again. Problems that had seemed impossible for years, solved in weeks, once someone actually looked at where the work was getting stuck.
“What’s next?” Edward asked.
“Continuous deployment to production. Right now we still have a manual approval gate before prod. HIPAA requirement, need human sign-off. But we’re working on automating the compliance checks so the human just has to verify the AI’s work instead of doing it from scratch.” Marcus Okafor grinned. “Eight years of ‘continuous meetings about continuous deployment.’ Twelve weeks of actually doing it.”
Saturday, February 21, 2026 – 2:00 PM – The Dock, Edward’s House on the Intracoastal
Sophia came home for the weekend.
She was in her spring semester at Georgia Tech. She had a few free days, and time to catch up with her father. Edward picked her up at the airport on Friday night and spent Saturday afternoon on the dock with her, fishing poles in hand, catching up.
“Tell me about what you’ve been doing,” Sophia said. “Mom says you’ve been obsessing over it.”
“Obsessing is a strong word.”
“Dad. She said you were up at 5 AM every morning for three months. That’s obsessing.”
Edward smiled. “Fair. It’s process improvement. Value stream mapping. Teams map their actual process. where work waits, where it gets stuck, where time gets wasted. Then we fix the bottlenecks. Some fixes involve AI. Most don’t.”
“That’s it?”
“That’s it. The medication tracking feature. the one that had been stuck for eighteen months. shipped in eighty-seven days. Not because of AI. Because we finally looked at what was slowing us down and fixed it.”
Sophia was quiet for a moment, reeling in her line. “At spring break, I told you and Uncle Marcus about the hackathon. About how the teams that struggled were the ones who didn’t know what they were building.”
“I remember.”
“You listened. He checked his phone.”
Edward looked at his daughter. Nineteen now, a sophomore in computer science, already seeing things he’d missed for decades.
“I almost didn’t listen,” he admitted. “My first instinct was the same as Marcus’s. ‘That’s great for school projects.’ But something you said stuck with me. About clarity being more important than technical skill.”
“And now you have an approach that works.”
“Now I have an approach. But it’s not really about the approach. It’s about looking at reality before you jump to solutions.” He paused. “The Infrastructure team failed back in January. Twelve hours of downtime. Three board members wanted to shut everything down.”
“What happened?”
“They got overconfident. Mapped their process, saw early wins, and pushed too fast.” Edward set down his fishing pole. “The approach didn’t prevent the failure. It just made the failure visible. Learnable.”
Sophia turned that over. “That’s still valuable. Most failures are invisible. People don’t even know what went wrong.”
“That’s what I told the board. They weren’t convinced.”
“But the pilots are continuing?”
“They’re continuing. Sandra. the CFO. backed me up. She said every improvement effort has setbacks. The question is whether you learn from them.”
They fished in silence for a while. The afternoon sun was warm, the water calm.
“Dad,” Sophia said finally, “can I tell you something?”
“Always.”
“I’m worried about Uncle Marcus. He’s so certain about everything. Every time I see him, he has answers. Confidence. But he never seems to wonder if his process is actually working.”
Edward felt a chill despite the warm air. “I’ve been thinking the same thing.”
“Have you talked to him?”
“I’ve tried. He’s focused on the IPO. Everything is ‘after the IPO.’ Like the IPO is the finish line and nothing exists beyond it.”
“But the IPO isn’t a finish line. It’s a milestone. What comes after? What’s he actually trying to build?”
“That’s what I keep wondering. And he’s not.”
Sophia reeled in her line. empty hook, no fish. “Maybe he’ll start asking after the IPO.”
“Maybe.” Edward didn’t believe it. “Or maybe by then it’ll be too late.”
They packed up the fishing gear as the sun started to set. Walking back to the house, Sophia slipped her arm through his.
“I’m proud of you, Dad. For actually looking at your process. For fixing what was broken. For building something that matters.”
“I’m proud of you too. For being the one who started this whole thing. At that dinner in March, with your hackathon stories.”
“I didn’t start anything. I just talked about what I was learning.”
“That’s how most important things start. Someone talks about what they’re learning. Someone else listens.”
They went inside, where Jennifer and the girls were waiting with dinner.