Alex has been trying to hire an engineer for three months. Not just any engineer. A specific person who works at a company that ships features in days while his 47-person team takes weeks.
He’s watched them launch three features in the time his team shipped one. Same complexity. Same domain. They just move faster.
He knows why. They don’t wait. They use AI agents to generate complete specifications, implementation, tests, and deployment configs in a single pass. They ship with one other person in two to three days.
His team? Each engineer needs eight other people to ship one feature. Product manager writes specs. Designer creates mocks. QA tests separately. Security reviews separately. DevOps deploys separately. Engineering manager coordinates all of this.
Eight people. Seven weeks. One feature.
He wants to hire engineers who eliminate that entire coordination chain.
HR said no. Three times.
First Try: Market Rate
He identified the engineer in May. They were interested. Compensation expectations: $250,000 base plus 25% bonus.
His Head of Talent pulled benchmark data. “Senior Software Engineer in our market averages $145,000 total comp. Your budget is $100,000 base plus bonus. That’s already above market for our region.”
He tried explaining this wasn’t a normal senior engineer role. This person eliminates the entire SDLC coordination overhead. They ship complete features with one other person, not eight.
She looked confused. “What do you mean eliminates the SDLC? Everyone goes through our process.”
That’s the point. They don’t need the process.
“We need to maintain internal equity,” she said. “If we pay one engineer $312,000, everyone will expect the same.”
He tried the math. We have 47 engineers at $100,000, that’s $4.7M. Add product managers, QA, security, DevOps, managers and the total delivery cost is about $7.7M annually. We shipped 24 features last year. That’s $320,000 per feature.
With 12 of these engineers at $312,000, total delivery cost would be $4.25M. They’d ship 240 features. That’s $17,000 per feature. We’d save $3.5M and ship 10x more.
She wrote it down. “I’ll take this to the compensation committee.”
Two weeks later he presented to the committee. They said the role doesn’t exist in market data. How do we know this is real?
Because startups are paying it. The company this engineer works at just raised a Series B. Twelve people. $23M ARR. That’s $1.9M revenue per employee. We’re at $750K per employee with 800 people.
“We’re not a startup. We need to maintain internal equity and sustainable cost structures.”
Decision: Denied. Counter-proposal: hire three engineers at $100,000 each.
He took it.
What Happened
Six months later those three engineers have shipped six features total.
Each one needs a product manager to write requirements for 2-3 weeks, a designer for mockups taking 1-2 weeks, code review with 2-3 day turnaround, QA cycle for 1-2 weeks, security review for a week, and DevOps deployment with 3-5 day wait for scheduled windows.
The coordination overhead consumes 40% of their time. They’re productive. They’re good engineers. They’re just trapped in a system that requires eight people and seven weeks to ship anything.
The engineer he couldn’t hire? Joined a competitor.
Last week that competitor launched a feature that’s been on his roadmap for four months. They built it in 19 days.
Second Try: Better Data
He went back to the compensation committee in August with numbers he’d been tracking all quarter.
His three new engineers cost $300,000 in salary. Their allocated share of PM, design, QA, security, DevOps, and management added another $400,000. Total cost $700,000 to ship 6 features. That’s $116,666 per feature with an average cycle time of 7.2 weeks.
For agent-native engineers he calculated $312,000 per engineer with minimal supporting costs. Three engineers would be about $1M total to ship a projected 60 features based on competitor velocity. That’s $16,666 per feature with 3-4 day cycle time.
“This is still theoretical,” the committee chair said. “We don’t have proof these engineers would perform this way in our environment.”
Let me run an experiment, he said. Hire one. Give me 90 days. Track features shipped, cycle time, cost per feature, and people required. If it doesn’t work we spent $78,000 on a definitive answer.
“What if it works and then everyone wants the same compensation?”
Then we have a model that ships features at $16,000 instead of $116,000. That’s a problem I want to have.
They escalated it to the CFO.
The CFO Meeting
Two weeks later he sat down with the CFO. She was concerned about precedent. If we pay one engineer $312,000 we’ll have retention issues with the others.
We have retention issues now, he said. I’ve lost two senior engineers in the past quarter to companies paying $250,000 and up. We’re losing people because our delivery process is frustrating. Engineers don’t want to spend 40% of their time waiting for approvals.
She asked what happens if we don’t do this.
We continue shipping 24 features per year at $320,000 per feature while competitors ship 200 plus features at $20,000 per feature. Our cost structure becomes unsustainable and our product falls behind. In two years we’re explaining to the board why we’re losing market share despite having 10x the engineering headcount of our competitors.
What’s the ask?
Authorize one hire at $250,000 to $350,000 base plus 25% bonus. Bypass the normal compensation process. Give me 90 days to prove this works with data. If it doesn’t I’ll own it.
She looked at the quarterly numbers. They’d missed three delivery commitments in Q2.
“One hire. Ninety days. You report results to me weekly. If this doesn’t work we don’t do it again.”
Approved.
Getting It Done
He didn’t go through recruiting. He called an engineer he’d identified at a 15-person startup. They’ve been shipping AI-powered features consistently for 18 months.
How are you shipping complete features in 2-3 days?
“We use agents to generate complete specifications with all the edge cases, security requirements, and test scenarios up front. Then the agent implements everything in one pass. Code, tests, docs, deployment configs. I validate the spec is complete, run the tests, deploy. Most features take 12-16 hours of actual work across 2-3 days.”
How many people do you need to ship a feature?
“Me and one other person to review the specification before implementation. That’s it. No separate QA, no security review meetings, no deployment tickets. It’s all in the spec.”
What would it take to do that here?
“You’d need to let me work directly with product on specifications. No user stories, no sprint planning. We generate the complete spec together with the agent in 2-4 hours. Then I ship it. If your process requires eight handoffs this won’t work.”
What if we create a two-person team that bypasses the normal process?
“Then I’m interested.”
He offered $325,000 base plus 25% bonus. They started in September.
Sixty Days In
The agent-native engineer has shipped 14 features. Average cycle time 2.9 days. Two people involved per feature. Cost per feature around $19,000.
His traditional team of 46 engineers shipped 4 features. Average cycle time 7.8 weeks. Eight to ten people involved per feature. Cost per feature around $340,000.
Same feature complexity. Same production standards.
But here’s what he didn’t expect. After the agent-native engineer shipped their seventh feature, one of his senior engineers came to his office.
How is this person shipping so fast?
They’re using agents to generate complete specifications that include everything. Requirements, edge cases, security, tests, deployment. Then the agent implements all of it at once. No handoffs. No waiting.
Can I try that?
Yes but you’ll need to work differently. Instead of waiting for product to write user stories you’ll work with product to generate complete specs with the agent. 2-4 hours of spec work instead of 2-3 weeks of story refinement.
What about code review?
The agent generates tests with the code. If the tests pass ship it. Code review is for the specification before implementation, not the code after.
What about QA?
That’s what the tests are for. If you need QA your specification wasn’t complete enough.
Two weeks later that senior engineer shipped their first feature this way. Three days, one other person involved. They came back and said this is how we should be working.
Now he has three engineers working this way. Four more want to try.
What He’s Learning
The compensation committee keeps asking for market data for this role. There isn’t any yet. The role is emerging right now.
It’s like 2008 when DevOps Engineer didn’t exist in HR databases. Systems Administrators were benchmarked at $85,000. Early DevOps engineers were getting $140,000 to $180,000. HR said that’s 65% above market. Companies that paid it deployed 100x more frequently and won their markets. Companies that waited for HR benchmarks to catch up were 18 months behind.
He’s not waiting 18 months.
The agent-native engineer isn’t 10x more productive at coding. They’re productive because they eliminated the waiting. His traditional engineers spend 2-3 weeks waiting for product to write requirements, 2-3 days waiting for code review, 1-2 weeks waiting for QA, 3-5 days waiting for deployment windows. That’s 40-50% of their time waiting for other people.
The agent-native engineer generates complete specs with product in 4 hours using Claude. The agent writes code and tests simultaneously. They deploy continuously. Zero waiting time. That’s the difference.
The compensation committee sees $325,000 versus $100,000 equals 225% more expensive. The actual cost comparison is different.
Traditional engineer gets $100,000 salary plus their share of PM, design, QA, security, DevOps, and management which is about $130,000. Total system cost $230,000 per engineer. They ship about 4 features per year. Cost per feature $57,500.
Agent-native engineer gets $325,000 salary plus minimal support costs of about $20,000. Total system cost $345,000 per engineer. They ship about 60 features per year. Cost per feature $5,750.
The agent-native engineer costs 50% more per person but 90% less per feature. That’s what the compensation committee can’t see.
He can’t hire 20 of these engineers through the normal process. The compensation committee will never approve it. But he can transition existing engineers to this model. Four of his current team are already working this way. They’re shipping faster, they’re less frustrated, and they’re proving this isn’t about special people. It’s about working in a different system.
He’s not asking to hire 20 engineers at $325,000. He’s asking to let his existing team work without the eight-person coordination chain. That costs nothing.
Next Week
He’s presenting to the board. One slide.
We currently spend $7.7M annually to ship 24 features per year. Cost per feature $320,000. Time to market 7-8 weeks. For $4.25M annually we can ship 240 features per year. Cost per feature $17,700. Time to market 2-5 days. The constraint is not capital. It’s our delivery process and the compensation framework that prevents us from hiring people who can work outside it.
Request: Authority to transition 12 more engineers to the agent-native model and hire 3 more at market rate, $250K to $350K base plus 25% bonus, without compensation committee approval.
If they say yes he’s scaling this. If they say no he’s updating his résumé. Because he cannot keep explaining to customers why they missed another deadline while startups with 15 people are shipping their roadmap faster than they can.
What This Means
If you’re trying to hire engineers who can ship faster using AI agents here’s what matters.
Stop calling them AI engineers. HR hears that and thinks senior engineer who uses a tool. They’ll benchmark it at $145,000. That’s not the role. Call them engineers who eliminate delivery dependencies or engineers who ship complete features with minimal coordination. Make HR see the system cost difference not the tool difference.
Calculate the total delivery system cost. Don’t compare salaries. Compare cost per feature shipped. Include product managers, designers, QA, security, DevOps, and engineering managers in your calculations. Show what it actually costs to ship one feature through your current system. Then show what it costs with someone who eliminates six of those eight dependencies.
Run a 90-day experiment not a hiring initiative. Don’t ask to hire 10 agent-native engineers at $300,000 each. Ask to hire one engineer at $300,000 for a 90-day experiment to prove we can ship features at $20,000 instead of $320,000. Make it an experiment with clear metrics and a personal commitment to track results. That’s approvable.
Work with your CFO not your comp committee. Compensation committees optimize for internal equity. CFOs optimize for business outcomes. Show your CFO we’re spending more to ship less, here’s a way to spend less and ship more. That’s a capital allocation conversation not an HR policy conversation.
Build the proof with one engineer then scale. Alex isn’t asking to transform his entire org on day one. He hired one engineer, proved it works, now four more are doing it. In 90 days he’ll have 10-15 working this way. In six months it’ll be 30. He’s not asking for permission to scale anymore. He’s showing the results and scaling into them.
Prepare to bypass process. The traditional hiring process will not work for this. Job descriptions optimized for keyword matching won’t find these people. Compensation bands built for 2023 roles won’t approve 2025 compensation. You’ll need to identify specific people, call them directly, and get executive approval to bypass normal process. That’s uncomfortable but it’s faster than waiting for HR systems to catch up to a role transformation that’s happening right now.
The Real Problem
Alex doesn’t blame HR. They’re doing their job. Maintain internal equity, control costs, use benchmark data to make defensible decisions.
The problem is their job was designed for a world where roles changed slowly, productivity scaled linearly with headcount, and market data reflected current roles.
That world doesn’t exist anymore. Roles are transforming faster than salary surveys can capture. Productivity is non-linear when one person can do what used to take eight. Market data is 12-18 months behind role transformations.
HR’s frameworks worked perfectly for 20 years. They don’t work now. And Alex can’t wait for them to catch up.
Where He Is Now
Sixty days into his 90-day experiment. The agent-native engineer has shipped 14 features. His traditional team of 46 has shipped 4. Four of his existing engineers are now working the same way. They’re shipping 2-3 features per week instead of 1 feature per 8 weeks.
The compound effect is starting to show. They’ll ship more features this quarter than they did in all of last year.
But he’s still frustrated. Because he knows exactly who he wants to hire. He knows what they’d deliver. He knows the ROI. And he still can’t get a compensation committee to approve it without months of benchmark analysis and internal equity studies.
Meanwhile startups are hiring these engineers at $300,000 to $400,000 and shipping products that make his look slow.
So he’s bypassing the process entirely. Going directly to his CEO and board with data. We can spend less and ship more if we hire people who eliminate coordination overhead.
If they approve it he’ll scale fast. If they don’t he’ll know this company isn’t capable of adapting to how software gets built now.
Either way he’ll know in two weeks. And he won’t spend another three months in compensation committee meetings debating whether $312,000 is above market while competitors pay $400,000 and capture his market.
The constraint isn’t capital. The constraint isn’t talent availability. The constraint is that HR frameworks built for 2005 can’t evaluate roles that emerged in 2023. And he can’t let that constraint determine whether his company is competitive in 2025.
So he’s going around it.
You should too.