The AI adoption gap: half the team is flying, half opted back out
You budgeted for the agents. Whether that budget pays off is decided at the team layer, where adoption is spreading unevenly and nobody is measuring it. Here is how teams close the gap without mandates or surveillance; the pulse and working agreement are free further down.
The tools arrived; the culture is on its own
Fiona Fung manages the Claude Code and Cowork teams at Anthropic. When she listed what remains unsolved in the most AI-leveraged engineering org anywhere, scaling culture was on the list: the tools multiply individuals, and nothing in the box multiplies the team. Inside an ordinary team that lands as a widening spread. One engineer runs three agents in parallel; the next tried an assistant in March, got burned, and went back to typing everything.
Left alone, the gap compounds. The quality bar forks into two bars depending on who wrote the code, which is the opening move of the verification gap. Review burden lands unevenly and resentment runs both directions: the fast half waits on the slow half, the slow half feels judged by a standard nobody agreed to. And the two default management responses both backfire: mandates produce usage without trust, and per-engineer dashboards kill the honesty that would have located the real blockers.
How Aurora Coach addresses it
The end state is one team again: agents multiplying a shared way of working instead of fragmenting it. Aurora Coach runs that as a continuous improvement loop, Sense, Analyze, Recommend, Refine, Commit, Execute, Re-evaluate. The adoption gap lives mainly in the Engineering domain (AI-assisted development), with Alignment (tooling effectiveness, removing impediments) alongside, and it is the team-layer half of the AI budget: the agents were the easy line item, the way of working is the part that decides the return. That timing argument is the whole of why now.
- Sense + Analyze The pulse above is the sensing: where agents help, where trust broke, whether the bar is drifting, asked of everyone each period, with delivery signal from Aurora Coach for GitHub alongside. The analysis maps the spread, who is blocked by access, who by skill, who by doubt, with no vendor adoption playbook underneath.
- Recommend + Refine + Commit Proposals come back sized to that map: a show-and-tell cadence, an access fix, a shared bar for agent-written code, each with rationale and success criteria. The team votes, the lead refines, and no mandate enters through the back door.
- Execute + Re-evaluate Commitments run inside normal delivery, and the next pulse shows whether the gap narrowed: abandonment points cleared, skeptics answered, one bar holding. Adoption becomes a trend the team owns instead of a mandate it endures.
The gap's downstream effects have their own pages: forked quality bars feed the verification gap, and the opted-out half skipping the learning curve is tomorrow's skill atrophy in the other direction.
What is the AI adoption gap?
The spread inside one team between engineers working at agent speed and engineers who tried the tools once and opted back out. It shows up as forked quality bars, uneven review burden, and resentment running both directions. Fiona Fung, who manages Anthropic’s Claude Code teams, names scaling culture among the problems AI has not solved; the adoption gap is that problem at team size.
Should we mandate AI tool usage?
Mandates produce compliance theater: tokens burned, nothing trusted, and skeptics who stop objecting without being convinced. What closes the gap is conditions, not orders: equal access and budget, workflows demonstrated in the open, one quality bar regardless of authorship, and skeptics treated as a source of risk information rather than resistance.
How do you close the adoption gap without surveilling engineers?
Measure the team, not the person. A per-engineer usage dashboard kills the honest answers that make the gap fixable, and usage numbers cannot distinguish healthy skepticism from being stuck anyway. A team-level pulse, asked well and safely each period, shows where adoption actually stands and whether it is moving, which is the only thing leadership needs to know.
The six-question adoption pulse, free
Ask these as a team each period, out loud or in check-ins. No individual metrics, no scoreboard; the trend is the point.
- Where did an agent save you real time this period? Concrete wins, named. A team that cannot answer this is not adopting; a team where only two people can is adopting unevenly.
- Where did you try one and abandon it? Abandonment points are the map of where trust broke. They are also the cheapest thing to fix, because someone else on the team usually got past the same point.
- What do you not trust an agent with, and why? Distrust is information, not resistance. Some of it is correct and belongs in the team’s working agreement; some of it is three months out of date.
- What is blocking more use: access, skill, or doubt? Three different blockers, three different fixes. Access is a budget email, skill is a show-and-tell, doubt is a conversation the team has not had yet.
- Whose workflow would you like to see up close? Adoption spreads by demonstration, not decree. This question tells you who should run the next show-and-tell.
- Has the team’s bar for merged code moved? Up, down, or forked into two bars depending on who wrote it. The honest answer here is the state of your engineering culture in one line.
The adoption working agreement
Five commitments that close the gap without a single mandate:
- One bar, regardless of authorship Standards do not fork by tool use. Code is held to the team’s bar whether a human, an agent, or both wrote it, and review effort follows risk, not origin.
- Show, don’t mandate A short weekly slot where someone demos how they actually work: prompts, failures included. Mandates produce compliance theater; demonstration produces adoption.
- Equal access, equal budget Everyone gets the same seats, models, and limits. An adoption gap that is really an access gap is leadership’s to close, not the engineer’s.
- No individual scoreboard Adoption is measured at the team level, through outcomes and the pulse, never as a per-person usage metric. Surveillance kills the honest answers everything else depends on.
- Doubt is data Skeptics name their risks out loud and the team decides together what agents touch. A skeptic argued with is a future power user; a skeptic dismissed just stops telling you why.
Not ready to change anything today? You already have the adoption pulse and working agreement above, copy buttons and all. If you want one improvement loop like these in your inbox each month, leave your email.
The pulse and the agreement work as a standalone exercise. The trial runs them every period so adoption becomes a trend instead of a one-off reading. Map your ROI below asks a few short questions about your AI spend, then gives you a quick read on the return, about five minutes, no signup.