Dan Guido Maps 4-Step AI Adoption Ladder for 130-Person Trail of Bits
Updated
Updated · O'Reilly Media · Jul 31
Dan Guido Maps 4-Step AI Adoption Ladder for 130-Person Trail of Bits
1 articles · Updated · O'Reilly Media · Jul 31
Summary
Guido said most enterprises stall because they stop at AI-assisted use—handing out ChatGPT or Claude licenses—rather than redesigning workflows so AI becomes a core participant in how work gets done.
Nearly 90% of about 6,000 executives in an NBER-cited study reported no measurable productivity or employment change from AI over three years, a gap Guido ties to poor deployment rather than weak models.
Trail of Bits answered internal resistance—95% initially pushed back, including 20% actively—with a four-level maturity matrix, an AI handbook, hackathons every two months, and visible CEO-led adoption.
The 130-person firm also built public and internal skills repositories, hardened defaults, sandboxed agent workflows, and a seven-day package-install delay to make experimentation reusable and safer.
Guido argues AI-native companies will shift jobs toward higher-level work such as skills product management and agent evaluation, with productivity gains arriving only after organizations restructure around the tools.
If AI tools are everywhere, why do most companies still miss the productivity payoff until they redesign workflows, permissions, and roles around them?
What happens when a security firm treats every AI failure as “scar tissue” to harden into standard infrastructure and shared skills?
Could the real barrier to AI adoption be organizational resistance, not the technology itself?