In a survey of 800 business leaders conducted by Coastal with Oxford Economics, 77% said their employees are eager to work with AI. While encouraging, that enthusiasm doesn’t predict AI adoption. Organizations whose leaders say their people are eager to work with AI experience adoption challenges at nearly the same rate as everyone else.
What Decides AI Adoption: Fit and Trust
When asked where their AI initiatives most commonly stall or fail, 57% of leaders pointed to user adoption after launch — the challenge of getting people to use what’s been built. Two things decide whether that happens: whether the tool fits the work, and whether people trust it.
Start with fit. Nearly half (46%) said the outputs were accurate but still didn’t fit the way work gets done. When the solution doesn’t fit, people work around it.
Then comes trust. Half (50%) said users didn’t trust the outputs enough to act on them, and that caution is rational: people won’t stake a decision on an output they can’t check, and models are wrong often enough to justify the doubt.
Users don’t automatically trust answers from a black box. They need to understand why the AI made a recommendation, what guardrails are in place, and where human oversight fits into the process. When those elements are missing, the technology may be running and the dashboards may be updating, but user behavior and business outcomes remain unchanged.
Why Employee Enthusiasm Doesn’t Drive AI Adoption
What fit and trust have in common is that neither is about whether employees want AI. They’re about choices the organization makes: how the tool is built, where it’s deployed, and whether people have reason to trust it. The workforce’s enthusiasm was never the variable.
That’s how the eagerness number misleads. When leaders are confident their people are on board, they stop asking the questions that decide adoption. Only half (50%) said their AI strategy is clearly understood and supported across their organization. Leadership can agree on a strategy, but that doesn’t mean it will reach the people expected to act on it.
The commitment is real: 84% of leaders surveyed say AI is making them more competitive. The results are what’s lagging. Only 20% strongly agree that their initiatives have delivered measurable value. Closing that gap depends less on employee willingness and more on how organizations design, deploy, and communicate the AI they put into production.
Design for AI Adoption Before Launching Agents
Adoption is mostly decided before the solution launches. It comes down to understanding what you’re trying to solve (and whether it needs to be solved) before you start building.
Whether an output fits how people work — and whether they can trust it — depends on knowing the workflow in concrete detail. Skip that, and you get AI that’s technically impressive but operationally out of place. It surfaces the wrong information, arrives at the wrong moment, or interrupts a process that was working.
Understanding the work also means knowing how it performs without AI, so you can later tell whether AI improved things or the workflow was the real problem.
None of this ends at launch. Each new release is another adoption event, and the earliest signs that fit or trust is slipping are behavioral: usage tapering off, people overriding the recommendations, unprompted complaints. The organizations that catch those signals early treat user feedback as operational data.
See it in practice: How Big Brothers Big Sisters of America designed for trust and adoption from the outset.
Embed AI Into the Workflows People Already Use
Then there’s where AI lives. It gets adopted when people don’t have to go anywhere or do anything outside of their normal routine to use it. The deployment patterns that work put AI inside the systems people already use every day — a CRM record, a service console — or run it in the background with no interaction required.
It’s the standalone tool that struggles: a separate chatbot or dashboard people have to remember, switch into, and carry the answer back from. Each of those steps is a point where adoption breaks down.
The numbers track with it. Among organizations that deeply integrate AI into the systems people already use, 94% are confident they can scale it. Among those that barely integrate, that number drops to 48%. After all, you don’t bother scaling what people don’t use.
4 Questions to Ask When AI Adoption Stalls
So before asking whether your people are ready, ask the questions that decide adoption:
- Does the solution support how people work?
- Do people have reason to trust its outputs?
- Has the organization agreed on what problem it’s solving?
- Is there clear training and change management in place?
Employee enthusiasm is a real advantage. It just can’t answer any of those, and the answers are what adoption depends on.
What’s Next for Your AI Roadmap
Knowing which AI use cases are worth backing (and whether your Salesforce foundation can support them) is critical. Coastal’s AI Pathfinder gives you a clear read on your IT architecture, data quality, and technical debt — and a prioritized AI roadmap of which use cases to build first, how to sequence the work, and where the biggest payoff is.
The figures above are drawn from the 2026 AI Operations Report, a survey of 800 business leaders conducted by Coastal with Oxford Economics.
