Adoption is the work: enablement for AI initiatives

The tool launched. The work did not change.

Quarterly operations review, third agenda item. The AI assistant went live three months ago: four hundred licenses, a launch webinar, a page of tips on the intranet.

The COO has the usage report open. About a third of license holders signed in last month. Most of what they did falls into two categories: rewording emails and summarizing meeting notes.

The head of customer operations says her team is too busy to learn a new tool. The finance director says his analysts tried it, got one wrong answer, and went back to their spreadsheets. IT points out, fairly, that the system has had no downtime.

Someone proposes a second webinar. Nobody objects, and nobody expects it to help.

The technology works. What never got designed was the part where four hundred people change how they do their jobs.

A license is not a new way of working

The room is looking at a common result. Gallup’s data for the second quarter of 2026 show 51% of leaders using AI frequently, against 36% of managers and 26% of individual contributors (Gallup). The people who approve the purchase are the people most likely to use it.

Shallow use is the other half of the pattern. Among U.S. employees who use AI at work, Gallup finds that 45% of those using it for one or two purposes say it improves their productivity. Among those using it for seven or more, the figure is 90%. Rewording emails is one purpose.

The cause is rarely the tool. The tool was added on top of work that stayed exactly as it was. McKinsey’s 2026 global survey of 1,719 participants found that nearly three-quarters of AI high performers had fundamentally redesigned workflows because of AI. High performers are the 6 percent of respondents who attribute an earnings impact of 5 percent or more to AI and say their organizations have seen significant value from it. Among other respondents, just one-quarter had redesigned workflows (McKinsey).

Training shows the same gap. In BCG’s 2026 survey of close to 12,000 employees, managers and leaders, 72% said expectations for the skills they need have shifted, and only 36% felt they had received adequate upskilling (BCG). A launch webinar explains what the buttons do. It does not show a claims processor which step of her Tuesday changes.

So the holdouts are behaving reasonably. Gallup surveyed 23,717 U.S. employees in February 2026. In organizations where AI is available, 46% of non-users said they prefer to keep doing their work the way they do it now, and 39% said they do not believe AI can assist with the work they do (Gallup). Nobody has shown them otherwise.

Then there is the part people say less readily in a meeting. Pew Research Center found in late 2024 that 52% of U.S. workers were worried about the future impact of AI use in the workplace (Pew Research Center). An employee asked to find her own uses for a tool that might shrink her team has little reason to be inventive in public.

Design the adoption, not just the rollout

Adoption can be designed the way a rollout is: on purpose, with owners. Five moves carry most of it.

Redesign the workflow before you train anyone. Pick two or three workflows where the value is clear and map them step by step: which steps AI takes, where a person checks the output, and how the handoff changes. Where AI is available, Gallup found 88% of employees who strongly agree that AI integrates well with the systems and processes they use are frequent users, against 55% of those who do not (Gallup). That is the difference between embedded and bolted on.

Train by role, on real tasks. Generic prompt training produces generic use. Build each session around the team’s own documents and its redesigned workflow. BCG’s 2025 survey found regular usage “sharply higher for employees that receive at least five hours of training and have access to in-person training and coaching” (BCG). One webinar is neither.

Make managers the channel. People take their cue from whoever assigns their work. In AI-adopting organizations, Gallup finds 78% of employees who strongly agree that their manager actively supports AI use are frequent users, against 44% of those who do not. Only 36% strongly agree (Gallup). Microsoft, which sells AI tools, reports from a survey of 20,000 knowledge workers that organizational factors such as culture and manager support account for more than twice the reported AI impact of individual factors (Microsoft). Train managers first, and ask them to use the tool in front of their teams.

Keep a support and feedback loop running. Give every team a named person to ask and a place to report what does not work. Feed those reports into a change backlog, a running list of fixes to the workflow, the training and the tool setup, and review it on a schedule. BCG’s 2026 survey found 66% of frontline employees who use AI regularly get limited or no guidance on what to do with the time they save. That guidance belongs in the loop.

Say what this means for jobs. Redesign raises the stakes. BCG’s 2025 survey found employees at organizations undergoing comprehensive AI-driven redesign more worried about job security (46%) than those at less-advanced companies (34%). Tell people which roles will change, what the saved time is for, and what has not been decided. Only a third of frontline employees in BCG’s 2026 survey say leadership’s communications about AI are clear. Silence gets read as the worst answer.

The usual objection is that the best people will work it out for themselves. Some will, and they are already in the usage report. A handful of enthusiasts is not adoption. It is a pilot nobody is running. The licenses are already paid for, and two workflows and their managers are a smaller commitment than another year of the same report.

The same report, read differently

Back in the review, the second webinar is still on the table. The more useful move is smaller. Pick the two workflows where the usage report looks weakest, sit down with the managers who run them, and redesign the work with the people who do it.

The report was never measuring whether the tool works. It was measuring whether anyone had been shown what to do differently on Monday. That part was always the work.

The opening scenario is a composite of common patterns, not a client engagement. SopraLabs builds the enablement plan, training assets and support model with leadership teams as the Enablement & Adoption module of our services.

Frequently asked questions

Why aren’t employees using the AI tools we bought?

Usually because the tool was added to existing work without changing the workflow, the training or what managers expect. In Gallup’s February 2026 survey, 46% of non-users in organizations where AI is available said they prefer to keep doing their work the way they do it now, and 39% said they do not believe AI can assist with their work (Gallup). Both are reasonable responses to a workflow that never changed.

How much AI training do employees need?

More than a launch webinar, and tied to their role. BCG’s 2025 AI at Work survey found regular usage “sharply higher for employees that receive at least five hours of training and have access to in-person training and coaching” (BCG). Treat five hours as a reference point from one survey, not a standard.

Who should own AI adoption?

The leaders and managers who own the workflows being changed should own adoption. IT keeps the tool running, and training and communications staff support the change. Frequent use tracks manager support closely, so accountability belongs with the people who assign the work.

How should leaders talk to employees about AI and their jobs?

Directly and early. Say which roles and tasks will change, what the time saved is for, and what has not been decided yet. In late 2024, Pew Research Center found 52% of U.S. workers worried about the future impact of AI use in the workplace (Pew Research Center), and people fill silence with their own answer.

Sources

• McKinsey, The state of AI in 2026: On the road to ROI. Published August 25, 2026; 1,719 participants in 97 nations, fielded May 4 to June 8, 2026. Supports the workflow redesign comparison and the definition and size of the high-performer group.
• Gallup, AI in the Workplace: What Separates Adopters and Holdouts. Published April 12, 2026; 23,717 U.S. employed adults, Gallup Panel web survey, February 4 to 19, 2026. Supports manager support and frequent use (78% and 44%), systems integration (88% and 55%), and non-users’ stated reasons (46% and 39%).
• Gallup, AI and Workplace Productivity: What Leaders Need to Know. Published July 27, 2026, updated September 30, 2026. Supports frequent use by job level in Q2 2026 (51%, 36%, 26%), productivity by number of work purposes (45% and 90%), and the 36% who strongly agree their manager actively supports AI use.
• BCG, AI at Work: Why Strategy Matters More Than Tools. Published June 3, 2026; close to 12,000 frontline employees, managers and leaders in more than a dozen markets. Supports the upskilling gap (72% and 36%), guidance on time saved (66%), and clarity of leadership communications (a third).
• BCG, AI at Work 2025: Momentum Builds, but Gaps Remain. Published June 23, 2025; more than 10,600 leaders, managers and frontline white-collar employees in 11 countries and regions. Supports the five-hours training finding and job security worry under redesign (46% and 34%).
• Microsoft, 2026 Work Trend Index: Agents, human agency, and the opportunity for every organization. Published May 5, 2026; 20,000 knowledge workers who use AI at work across 10 markets, surveyed by Edelman Data x Intelligence, February 18 to April 7, 2026. Supports organizational versus individual factors (67% vs. 32%).
• Pew Research Center, U.S. Workers Are More Worried Than Hopeful About Future AI Use in the Workplace. Published February 25, 2025; 5,273 employed U.S. adults, surveyed October 7 to 13, 2024. Supports the 52% of workers who are worried.