01
Account creation
Auth, KYC, identity verification, 2FA, session management, password reset, lockout policies. A full domain before anyone applies for anything.
Slide 01
For fifteen years, the pyramid told you to write fewer end-to-end tests because humans could not maintain them at scale. That was the right call given the constraint. The constraint no longer exists. If your testing strategy still looks like a pyramid in 2026, you are carrying risk you do not need to carry.
Slide 02
One engineer writes them alongside the code. They run in milliseconds. They break cleanly. A human can maintain hundreds without losing their mind.
Test environments. Data setup. Cross-service coordination. When they break, a human has to investigate whether the fault is yours or someone else's. That takes time.
Full environments, synthetic users, browser automation. One flaky E2E test burns a day of engineering time. Ten burn a sprint.
"Why does the pyramid look like a pyramid?" Because of what things cost. The entire ratio — units to integration to E2E — is a human-capital ratio.
That is not a testing strategy. That is a budget allocation model disguised as engineering wisdom.
Slide 03
E2E tests most closely mirror what your users actually do. If you could write one for every permutation, every user path, every edge case, you would have near-perfect software.
That is not a technical limitation. That is a financial one. The pyramid exists because human beings are expensive and E2E test maintenance is a human-capital sinkhole.
Slide 04
Auth, KYC, identity verification, 2FA, session management, password reset, lockout policies. A full domain before anyone applies for anything.
Income, employment, document upload, automated underwriting. Approved, denied, or conditional. Each branch is its own workflow with its own edge cases.
ACH pulls, late fees, grace periods, principal-vs-interest application, extra payment handling. One missed payment triggers a cascade of business rules.
Monthly statements, 1098 tax docs, credit bureau reporting, lien release, e-title transfer. Each with its own regulatory requirements and delivery mechanism.
Slide 05
Slide 06
The part where a human opens a test file, reads the failure, traces it to a code change three PRs ago, updates the selector, regenerates the fixture, reruns the suite, watches it fail again for a different reason, fixes that too, and commits the whole mess? That part is gone.
You still pay for compute. You still need engineers who understand what to test and why. That judgment is not going anywhere. But the manual labor of test maintenance is no longer the binding constraint.
Slide 07
Agents generate and maintain the complete unit suite for every service. Same as before, but fully automated regeneration on every code change.
Agents maintain cross-service integration tests. Payment-to-ledger. Decisioning-to-payments. No more mocked boundaries hiding contract drift.
Agents verify that the payment API still matches what reporting expects. Every commit. Not once a quarter by a QA engineer who knows where the bodies are buried.
Agents walk a synthetic borrower through the complete 36-month lifecycle. Account creation to e-title. A suite that would have taken a human team a week to build and a month to keep from rotting.
Agents simulate ten thousand borrowers all making payments on the first of the month. Every release. Not just before the annual audit.
Slide 08
Slide 09
If you mock the database, you never find out your ORM generates a query that locks the table for thirty seconds under load. If you stub the payment processor, you never discover their API changed the error response shape and your retry logic now swallows failures silently.
You do not have to choose between isolation and realism anymore. Mocks where isolation genuinely matters (pure business logic without network noise), real integrations everywhere else. The ratio stops being a religious war and becomes an engineering decision.
Slide 10
The QA engineer reads the acceptance criteria, builds a test plan around what was written down, and misses the thing that was never written down because the developer thought it was obvious.
The knowledge and the verification live in the same process. The feedback loop is measured in seconds, not days. Engineers who understand quality deeply are more valuable than ever — but they are embedded in engineering teams, shaping what agents build and how agents test.
Slide 11
Agent orchestration. Full test suites across every domain. Engineering time to design the test architecture — deciding what agents should assert and how domains get segmented. Real work requiring real judgment.
Production outage during month-end close. Compliance finding from wrong interest on six months of 1098s. Security incident from an unvalidated session token exposing 1,200 borrowers' data.
Every commit runs through unit, integration, contract, E2E, and performance tests at equal depth. You ship because the system told you it was safe. Not because someone said "I think we're good."
Your release cadence stops being a negotiation. It becomes a function of how much change your customers can absorb. That is a different problem entirely, and a better one to have.
The square is dramatically cheaper than the risk you are carrying right now — with a testing strategy shaped by a budget constraint from 2012.
Slide 12
Given the reality that human engineers had to write, maintain, and debug every test by hand, the pyramid was brilliant. It told you how to allocate a scarce resource across test types with wildly different maintenance costs.
Human capital was the constraint. Agents removed it. The pyramid becomes a square.
Slide 13
Integration failures. Contract drift. End-to-end workflow breaks. Performance degradation under load. These are the defects that take down production, trigger compliance reviews, and erode customer trust. And you are deliberately writing fewer tests for them because fifteen years ago someone told you they were too expensive to maintain.
They were. They are not anymore.