Evidence and economics

External results are useful for selecting hypotheses. A local comparison is needed to establish the value of a particular workflow.

Forecast

Estimates a future state. Useful for strategic scenarios; cannot establish current productivity.

Survey

Describes reported experience or associations. Useful for adoption barriers; selection and measurement bias remain.

Experiment

Can isolate an effect within its design. Transfer still depends on tasks, people, tools and the environment.

Case study

Shows implementation mechanics and local outcomes. Usually lacks a comparable untreated group.

Product docs

Describe available functions and integration contracts. Do not establish comparative effectiveness.

Apparently conflicting results can both be valid

The field experiments reported by Cui and colleagues found a pooled increase in completed tasks. METR’s early-2025 trial found experienced developers took longer on familiar repositories. The studies differ in task setting, participant experience, tool configuration and outcome definition. Pooling those percentages would create a metric that neither study measured. [E01][E03]

METR estimates · Change in task completion time with AI · Dots show estimates; lines show reported confidence intervals.

Early 2025: 16 developers, 246 issues in familiar repositories. This setting showed a slowdown; it does not represent all software work.

Different cohorts and tool periods; no pooled estimate. The follow-up intervals cross zero and selection bias limits interpretation.

Primary sources: July 2025 RCT ↗ · February 2026 follow-up ↗

The 2026 METR update does not provide a clean before/after series: recruitment selection and time measurement changed. Treat its estimates with the uncertainty and limitations visible in the chart. [E02]

Industrial QE cases show the importance of filtering

Meta’s test-generation work combines candidate generation with checks before developer review. Its mutation-guided work tests whether a candidate detects a relevant modeled fault. Google’s AutoDiagnose illustrates contextual failure diagnosis; Uber’s flaky-test work illustrates how repair depends on understanding the execution context. These are useful engineering patterns rather than universal business-case inputs. [E04][E05][E06][E07]

CandidateBuilds

Valid syntax
Available dependencies

RepeatRuns reliably

Stable outcomes
Relevant environment

ChallengeFinds a fault

Meaningful assertions
Targeted mutation tests

ReviewEarns acceptance

Maintainer judgment
No weakened oracle

↓ Failed checks reject or return the candidate for correction
Authored synthesis of filtered test generation and mutation-guided testing. Stages are not a numerical funnel. [E04][E05][E07]

A test that merely builds and passes can still preserve a defect or weaken an assertion. Review test meaning, fault detection and the independence of the expected result.

The value model must include the work around generation

Measure net effort per eligible task, including context preparation, generation, review, correction, repeated execution and control administration. Pair it with escaped defects, false dismissals, rework and elapsed delivery time. Keep active human effort separate from agent runtime and queue time.

Released capacity becomes recognized value through an explicit mechanism: additional delivery, reduced external spend, avoided hiring or another Finance-approved route. An improvement in task throughput is not the same unit as a reduction in labor cost.

$45.0killustrative net economic impact3.3% capacity released · 0.45% of addressable spend

Annual value bridge $ thousands · per $10M addressable QA spend
Base assumptions, illustrative annual values
Capacity equivalent$330000
Uncaptured capacity−$165000
AI and pilot cost−$100000
Quality allowance−$20000
Extra human effort$0
Net economic impact$45000
Net benefit vs. adoption All other current assumptions held fixed · $ thousands

Illustrative assumptions, not a forecast. Capacity is effort equivalent; cash savings need Finance-approved capture.

Activity share 55% · eligibility 60% · AI/pilot cost 1.0% · quality allowance 0.2%.

Formulas, assumptions and accounting treatment ↗

The simulator is an authored scenario model. Its defaults are assumptions, not industry benchmarks. See the full savings model and pilot measurement design for definitions and decision gates.

What the current evidence does not settle

There is no general bank-wide QE savings rate in this review. Evidence remains limited on sustained escaped-defect reduction, long-term test maintainability, simultaneous-agent human effort, rare operational failures and the economics of assurance itself. Those are explicit measurement questions for the pilot and subsequent production validation.

Generated tests: the quality filter in practice

Pan horizontally to explore the full diagram. A text description is available below.

Meta TestGen-LLM evaluation: among 86 target test classes, 75 percent had a generated test that built, 57 percent had one that passed reliably, and 25 percent had one that increased coverage A PASSING TEST IS AN INTERMEDIATE RESULT Share of target test classes with at least one qualifying generated test Builds correctly Syntax and dependencies 75% Passes reliably Builds + non-flaky execution 57% Adds coverage Builds + passes + new line coverage 25% 0% 50% 100% Meta / FSE 2024 · §3.3 · 86 Kotlin components with existing test classes · reported percentages rounded Denominator is target test classes, not individual generated tests. Coverage is a proxy for test improvement.
Observed industrial evaluation · Meta TestGen-LLM, FSE 2024, §3.3. The three bars use the same test-class denominator. [E04][E05]
Research behind this design & diagram description
Finding

In Meta’s 86-component Kotlin evaluation, 75% of target test classes gained at least one generated test that built; 57% had one that also passed reliably; 25% had one that also increased line coverage.

Design implication

Treat generation as a candidate factory. Require independent checks for stability and added test value; use fault detection where possible because coverage alone does not establish semantic quality.

Diagram in words

The bars have a common zero-to-100-percent axis. The qualifying share drops from 75% for build success to 57% for reliable passing and 25% for added coverage. These percentages describe test classes, not individual generated tests.

Inspect methods and source documents →

Canonical TestGen claim and economics

75% of target test classes had at least one generated test that built; 57% had one that also passed; 25% had one that also increased coverage. These are cumulative class-level yields, not percentages of individual generated tests. Cohort: 86 Kotlin test classes (31 Stories, 55 Reels), §3.3. In separate Instagram/Facebook test-a-thons, engineers accepted 73% of recommendations; 11.5% of all classes to which the tool was applied were improved. The cohorts and denominators differ. Neither acceptance rate is a predicted result for a bank pilot. [E04]

The canonical scenario model separates exact calculations, additional human effort and Finance-approved cash capture.