Updated

Updated · InfoWorld · Sep 7

R&D Organization Triples Software Output, Cutting Defects 65% With 18-Month AI Overhaul

Updated

Updated · InfoWorld · Sep 7

R&D Organization Triples Software Output, Cutting Defects 65% With 18-Month AI Overhaul

2 articles · Updated · InfoWorld · Sep 7

Output per engineer roughly tripled over 18 months after the R&D group rebuilt its software development life cycle around AI rather than limiting adoption to coding pilots.
Four-hour feature turnarounds replaced a 15-day path from spec to working pull request after the organization removed stage-by-stage handoffs and had one team carry work end to end.
68% of pull requests now use AI assistance, up from 3%, while deployments rose from 82 to more than 155 and customer-reported defects fell 65% per million lines of code.
76% lower vulnerability density and broader engineer adoption followed governance changes that put code review, security scanning and confidence-based approvals inside the delivery pipeline before scaling AI use.
99% of new tests are now AI-generated, versus 47% a year earlier, as the company extends the same agent-driven model to requirements, testing and even go-to-market support.

Sources

Could eliminating specialized security and QA handoffs actually be the counterintuitive secret to dropping software defect rates by 65 percent?

If AI writes the code and generates the tests, who ensures the machine isn't simply validating its own hidden mistakes?

What happens to a tech company when its engineers become so fast that the sales department becomes the biggest bottleneck?

2 total