How Consultants Assess AI Readiness in Organizations

Consultants assess AI readiness in an organization by looking past the work and the data to the people a change will run through: whether the leadership team agrees on what the change is for, whether the managers who'll carry it will back it, whether the people who'll use it think it's worth doing and believe they can do it, and whether anyone has the time. That's a change readiness assessment. The good ones assess readiness for one specific change among the specific people it touches, not a general mood about AI across the company.

That last point is where most published advice goes wrong. We read the 13 readable pages ranking for this question and close variants in September 2026. Most listed "culture," "leadership buy-in," or "change management" as one area among six or eight, and the firms that described a method mostly measured it with a self-scored questionnaire, one of them 15 questions long. Only two mentioned managers at all, each as a single line in a list. None said that readiness belongs to a particular change rather than to the company as a whole, and none mentioned the design choice with some of the best evidence behind it, which is letting people correct what the tool produces.

This page covers the people side of AI readiness for an organization of roughly 50 to 200 people, where there's at least one layer of management between the owner and the person doing the work. It's the leadership and team companion to our guide to AI readiness assessment services, which covers all six areas an assessment examines. The interview questions and the rubric for rating the work itself are in our sibling page on how consultants assess AI readiness in businesses.

Key Takeaways

• Change readiness is specific to one change. The same organization can be ready for one change and not another, so a consultant should assess the change you're actually planning, among the people it touches (Weiner, 2009).
• Managers carry adoption. In Gallup's 2026 data, 33% of employees whose manager actively supported AI use said it improved how work gets done, against 9% of those without that support.
• Plans lag rollouts. In 2025, 44% of US employees said their organization had begun integrating AI, but only 22% said it had communicated a clear plan (Gallup).
• Worry is not the same as fear of replacement. 52% of US workers said they were worried about AI's future impact at work (Pew Research Center, 2025), but only 15% thought it likely that AI would eliminate their job within five years (Gallup, 2025).
• Give people control. In experiments, 47% chose an imperfect algorithm when they couldn't change its forecasts, against 68% to 71% when they could adjust them even slightly (Dietvorst and others, 2018).

Table of Contents

• What does a consultant assess in an organization that it doesn't in a small business?
• Readiness belongs to a change, not to the company
• The four things consultants assess in an organization's change readiness
• Why managers decide whether AI gets used
• Worry, trust, and why control matters
• Time is the readiness factor nobody budgets
• How consulting firms help with AI change readiness
• A worked example: a 130-person millwork manufacturer
• What a change readiness assessment is not
• When you need one, and when you don't
• Where Avolis fits
• Frequently Asked Questions
• Continue Learning

What Does a Consultant Assess in an Organization That It Doesn't in a Small Business?

In a 15-person business, the owner often knows the person who'll use a new system well enough to predict how it'll land. In an organization of 100 people, the change passes through a leadership team, one or more layers of managers or supervisors, and then the people who do the work, often across shifts or sites. Each layer can stall it, and nobody at the top sees all of them.

So on top of the work itself, a consultant assessing an organization adds the people the change runs through. The pillar guide lists this as the "people" area of an assessment, covering skills, workload, and who would own the system day to day. In a small firm those come down to one or two conversations, as the business-level page explains. In an organization, they become their own piece of work, usually called a change readiness assessment, and it's the part most likely to be skipped because it doesn't show up in a systems inventory.

The work still comes first. If the workflow isn't frequent, repeatable, and owned, no amount of enthusiasm makes it a good AI project, and our methodology page treats the work as a gate for exactly that reason. Change readiness answers the next question: once you've found work worth changing, will the organization actually change it?

Readiness Belongs to a Change, Not to the Company

The ranking pages leave out the most useful idea in the research. Readiness for change isn't a general trait of a company, like its size or its age, that you can measure once and file away. It belongs to a particular change and to the particular people who'll have to carry it out.

The clearest statement of this comes from a widely cited 2009 paper by Bryan Weiner. He defines organizational readiness for change as members' "shared resolve to implement a change (change commitment) and shared belief in their collective capability to do so (change efficacy)" (Weiner, "A theory of organizational readiness for change", Implementation Science, 2009). In other words, people have to want to make this change and believe they can pull it off together, and both halves matter. He also calls implementation a "team sport," because the change stalls when some people are committed and others aren't.

Then he makes the point that matters most for an AI project. A healthcare organization with good relationships between managers and clinicians and a history of successful change, he writes, could still show high readiness to implement electronic medical records but low readiness for a new patient scheduling system. Commitment is partly specific to the change, and so is the belief that you can do it.

Here's what that means in practice. An organization-wide survey asking how people feel about AI tells you something, but not whether your dispatchers will trust an AI-drafted schedule or whether your estimators will stop rebuilding quotes by hand. A consultant who assesses change readiness properly starts from the change you've decided is worth making, names the people it touches, and assesses them. A general AI sentiment score for the whole company is a different and much less useful measurement.

The Four Things Consultants Assess in an Organization's Change Readiness

A change readiness assessment for AI looks at four things, and each one maps to a different group of people. Weiner's two halves, commitment and capability, run through all four. What changes from one to the next is whose commitment and whose capability is being assessed.

Consultants gather this from conversations with each group, from what happened in the organization's last significant rollout, and from the plain facts of the calendar and the org chart. The questions and interview technique are the same ones used for the work itself, covered on the business-level page. What differs is what the consultant is listening for, which the next three sections cover.

The last rollout deserves special attention, because it's the best evidence available before the change starts. If the organization moved to a new scheduling system or ERP two years ago, the people who lived through it will tell you how it was announced, who was trained and when, what broke, and whether the old spreadsheet is still running in parallel. That story predicts the next rollout better than any rating of how people feel about change in general.

Why Managers Decide Whether AI Gets Used

Managers decide whether AI gets used because they set what people are expected to do each day. The owner can announce a change, but the supervisor decides whether the new system is how the work gets done this week or something to try when things calm down. In organizations of 50 to 200 people, that's the shop foreman, the office manager, the service manager, and the shift lead.

The survey evidence here is large and consistent. In Gallup's first-half 2026 workplace data, a sample of 43,262 US employees, those whose manager actively supported their team's use of AI were far more likely to say AI improves how work gets done, 33% against 9%. Their engagement rate was 48%, against 30% for those without that support. Among the AI conditions it measured, Gallup called manager support the factor "more strongly associated with employee engagement than any other" (Gallup, "Employee Engagement Remains Flat as AI Adoption Accelerates", July 2026). These figures describe associations, not proof that one causes the other. Managers who back a tool may simply work in places where the tool is better. But the pattern is large and it points one way.

The same research points to where the gap usually opens. In Gallup's 2025 data, 44% of US employees said their organization had begun integrating AI, but only 22% said it had communicated a clear plan for doing so. Frequent AI use among leaders, defined as managers of managers, stood at 33%, twice the 16% among individual contributors. Among production and front-line workers it stayed essentially flat, at 11% in 2023 and 9% in 2025 (Gallup, "AI Use at Work Has Nearly Doubled in Two Years", June 2025). The people deciding on AI are using it far more than the people whose work it's meant to change, and the plan often never makes it down the org chart.

So a consultant assessing change readiness spends real time on the manager layer, trying to learn whether the supervisors helped choose what to change or will hear about it in the same meeting as their crews, whether they think it addresses a real problem in their area, and whether they'd hold their team to using it. A supervisor who's quietly against a change can stop it without ever saying no, simply by letting the old way keep working.

Worry, Trust, and Why Control Matters

Most people aren't afraid AI will take their job, but many are uneasy about it, and the unease shows up as distrust of the output. That distinction matters, because the fix for fear of replacement is reassurance, while the fix for distrust is design. A consultant who hears worry and prescribes a town hall has misread it.

The numbers support the distinction. In a Pew Research Center survey of US workers, 52% said they were worried about the future impact of AI in the workplace, 36% said they felt hopeful, and 33% said they felt overwhelmed (Pew Research Center, "U.S. Workers Are More Worried Than Hopeful About Future AI Use in the Workplace", February 2025). Yet in Gallup's 2025 data (Gallup, June 2025), only 15% of employees thought it very or somewhat likely that automation, robots, or AI would eliminate their job within five years, a figure unchanged since 2023. The two surveys asked different questions of different samples, so they can't be subtracted from each other, but together they suggest that most of the worry is about something other than being replaced.

One plausible version of that something, and one worth listening for, is accountability. Picture a dispatcher who has built schedules for fifteen years. She isn't mainly worried about being let go. She's worried about being blamed for a schedule she didn't build, and about losing the judgment that makes her good at the job. So she checks every line by hand, which means the tool saves nothing, and within a month the old way is back.

Research on this reaction goes back more than a decade. In a set of experiments published in 2015, people who watched an algorithm make forecasts became less likely to rely on it than on a human, even when they'd seen it outperform the human. They lost confidence in the algorithm faster after seeing it make the same mistake a person made (Dietvorst, Simmons, and Massey, "Algorithm Aversion", Journal of Experimental Psychology: General, 2015). The researchers called this algorithm aversion.

Their follow-up study is the more useful one for anyone planning a rollout. When people could adjust the algorithm's forecasts, even slightly, they became far more willing to use it. In one experiment, 47% chose the algorithm when they couldn't change its forecasts, against 71% when they could adjust each forecast by up to 10 percentiles and 68% when they could adjust it by only 2 (Dietvorst, Simmons, and Massey, "Overcoming Algorithm Aversion", Management Science, 2018). Shrinking the allowed change by 80% barely moved the number. People didn't need much control, just some.

Those were forecasting experiments, not workplaces, so the percentages shouldn't be read as adoption rates. The direction, though, is a reasonable guide for real work. A schedule the dispatcher can edit gives her the control the experiments point to, while one she can only accept or reject invites the line-by-line checking that erases the savings. A change readiness assessment should therefore find out what each group of users would need to control, and the answer should shape what gets built. That's why the people side belongs in the assessment before anything is designed, not in a training session afterward.

Time Is the Readiness Factor Nobody Budgets

The most practical question in a change readiness assessment is also the easiest one to skip: who has the hours to learn the new way while the old work keeps coming in? Weiner lists three things people weigh when they judge whether they can pull off a change. They ask what it will take, whether they have the resources, and whether the timing works given everything else going on. At 50 to 200 people, the scarce resource is almost always time, not money.

This is where the busy-season problem shows up. A new estimating workflow launched in a roofing company's April, or a scheduling tool launched in a manufacturer's peak quarter, asks people to learn something new at exactly the moment they have no slack. They'll fall back on the method they know because the work has to go out, and that fallback often becomes permanent. Pew's finding that a third of workers feel overwhelmed about AI in the workplace may partly reflect this.

A consultant assessing capacity looks at the calendar, the org chart, and the last rollout: when the slow season falls, who would cover for the planner during the hours she spends testing the new system, and whether anyone was given time off their normal work during the last rollout or simply expected to absorb it. A plan that doesn't answer these questions has budgeted for the software and not for the people learning to use it.

How Consulting Firms Help With AI Change Readiness

Consulting firms help with AI change readiness in two ways: by assessing it, which is the diagnosis described above, and by turning what they find into a plan. The assessment is worth little if it ends in a score. Each finding should change something about how the AI system is built, timed, or introduced, and it should come with a way to tell whether that change worked.

A few things separate real help from a change management slide deck. The plan should name people, not "stakeholders." Training should run on the organization's own work, such as last week's orders or this month's quotes, rather than a generic demo. And the firm should track adoption for a period after launch, because a readiness assessment predicts adoption but only usage data confirms it. Gallup's own recommendation from its 2026 findings points the same way, advising leaders to define AI expectations by role rather than announce a broad policy.

If you're weighing firms, what these engagements cost and the free and subsidized options are covered in AI readiness assessment consulting firms, and how to tell an assessment from a survey is in the business-level page's buyer checks.

A Worked Example: A 130-Person Millwork Manufacturer

Here's how the four areas come together in one assessment. This is an illustrative composite, not a client, and its figures are illustrative.

A 130-person architectural millwork manufacturer builds custom cabinetry and casework for commercial projects, running two shifts. The work assessment had already found its most promising workflow: turning approved shop drawings into cut lists and a production schedule. One production planner with 22 years at the company did it by hand, and it took her about two days per large project, which set the pace for the whole plant. The data was in the drawing files and the order system, and the task was frequent and repeatable. On the work alone, it was a go.

The change readiness assessment found four problems, one in each area.

Leadership. The owner described the problem as the time from signed drawings to the first cut, and wanted it shorter to win more fast-track jobs. The plant manager described it as the CNC router, which he thought was the real bottleneck. Both were partly right, but a system built to fix the owner's problem would be judged by the plant manager's measure, and it would look like a failure.

Managers. The first-shift supervisor was keen. The second-shift supervisor had run the floor through the company's ERP rollout three years earlier, which went live in the spring rush with no extra hours for anyone. The paper travelers his crew still used had come back during that rollout and never left.

Users. The planner wasn't worried about her job, since she was the only one who understood the edge cases. She was worried about being blamed for a schedule she hadn't built, especially when material arrived late and the plan had to be reshuffled. Her plan, she said, was to check every line the system produced.

Capacity. The proposed launch date fell in March, the start of the plant's busiest quarter, with no cover for the planner's time.

The plan. Leadership agreed on one measure, the days from approved drawings to a released cut list, with the plant manager as owner, and the CNC question became a separate project. The system was designed to produce a draft schedule the planner edits line by line, with her changes logged so the rules could improve, rather than a finished schedule she could only approve or reject. The second-shift supervisor was asked to test the draft schedules against his crew's actual sequence, and the paper travelers were given a retirement date. Launch moved to January, with four hours a week of the planner's regular work covered by a scheduler from the office for the first six weeks.

The assessment added no new technology to the plan. It changed the success measure, the design, the timing, and who was involved, and each of those changes came straight from something a person had said about the last rollout or the next one.

Citation-ready summary: Consultants assess AI readiness in an organization by examining the people a specific change will run through: whether the leadership team agrees on what the change is for and who owns it, whether the managers who set daily expectations will back it, whether the people who'll use it think it's worth doing and trust the output, and whether anyone has time to learn it. Readiness is specific to one change, so a consultant should assess the change being planned, among the people it touches, not general AI sentiment across the company.

What a Change Readiness Assessment Is Not

A change readiness assessment is not a culture survey. A questionnaire asking employees to rate their organization's openness to innovation measures a general mood. It can't tell you whether this group of people will adopt this change on this timeline, which is the only question that matters before you build something. Several of the firms we reviewed offer short self-scored surveys, and those can be a reasonable starting point for a conversation, but they aren't an assessment.

It's also not a justification for treating most change as doomed. The claim that 70% of organizational change initiatives fail appears constantly in change management material, and it has no solid basis. A review of five published sources for the figure concluded that "there is no valid and reliable empirical evidence" to support it (Hughes, "Do 70 per cent of all organizational change initiatives really fail?", Journal of Change Management, 2011). If a proposal leans on that number to sell you change management services, it's worth asking what else it hasn't checked. The honest version is that some changes fail, and the reasons are usually specific and knowable in advance, which is the case for assessing them at all. Our page on why AI projects fail covers the failure patterns that are actually documented.

Finally, it isn't a replacement for assessing the work. A team that's eager to adopt AI on a workflow that isn't worth automating will adopt it quickly and get little from it. Change readiness decides whether a good project lands, not whether a project is good.

When You Need One, and When You Don't

You need a separate change readiness assessment when a change crosses a layer of management, spans shifts or sites, or follows a rollout that went badly. In those cases, the people who decide on the change and the people who'll live with it are far enough apart that nobody in the room sees the whole picture.

You probably don't need one as a separate exercise at 10 to 30 people, when the owner works alongside the people whose work will change. At that size, the people questions come down to whether the person doing the work will use what gets built and who'll own it when it breaks, and they fit inside the ordinary assessment described in how consultants assess AI readiness in businesses. If you'd rather start on your own, how can I assess my organization's AI readiness? walks through a do-it-yourself version, and its closing section covers where a self-assessment runs out.

Either way, one question belongs alongside the buyer checks on the business-level page: which people did you talk to about this change, and what will you do differently because of what they said? If the answer is a company-wide score, the assessment measured the wrong thing.

Where Avolis Fits

Our diagnostic covers the people side as part of the same piece of work, not as a separate engagement. We work through each workflow in scope with the people who run it, and for an organization that includes the managers who set expectations for that work and the people who'll use what gets built. We ask about the last rollout, find out who would own the system and what users would need to control, and note what the calendar allows. Those findings shape the ranked result as much as the counts and timings do.

Every engagement starts with that diagnostic, and the ranked result is yours whatever you decide next. If you're comparing partners, which AI consulting company to choose sets out what to look for, including in us.

Frequently Asked Questions

How do consultants assess AI readiness in organizations?

They assess the work first, then the people a specific change will run through: whether leaders agree on what it's for and who owns it, whether managers will back it, whether users think it's worth doing and trust the output, and whether anyone has time to learn it. Findings should change the design, timing, and rollout.

What is an AI change readiness assessment?

It's an assessment of whether the people a planned AI change will affect are ready to carry it out. It looks at leadership alignment, manager support, users' commitment and trust, and capacity. Good ones assess one specific change among the people it touches, because the same organization can be ready for one change and not another.

How do consulting firms help organizations with AI change readiness?

They assess readiness, then turn each finding into a change: one agreed success measure and owner, supervisors involved in testing, a workflow where users edit the AI's draft rather than accept or reject it, a launch timed for the slow season with backfill hours, and adoption tracked after launch.

Why do managers matter so much for AI adoption?

Managers set daily expectations, so they decide whether a new system is how work gets done or something to try later. In Gallup's 2026 data, 33% of employees with active manager support for AI said it improved how work gets done, against 9% of those without it. These are associations, but the gap is large.

Do employees resist AI because they fear losing their jobs?

Mostly not. In 2025, 52% of US workers told Pew they were worried about AI at work, but only 15% told Gallup their job was likely to be eliminated within five years. The two surveys suggest the worry is mostly about something else, and one plausible version, fear of being blamed for output they didn't produce, is a design problem.

Continue Learning

When consultants assess AI readiness in organizations well, they find that an organization can be ready for one change and not the next, and the difference usually comes down to a handful of people: a leadership team that agrees, a supervisor who backs it, a user who can correct the output, and a calendar with room in it.

Before you hire anyone, ask which of your people they plan to talk to about the change.

How assessments work:

• AI readiness assessment services
• How consultants assess AI readiness in businesses
• AI readiness assessment methodology
• AI readiness assessment checklist and PDF

Assess it yourself:

• How can I assess my organization's AI readiness?
• How businesses assess readiness for AI adoption
• AI readiness assessment for SMBs

Choosing who does it:

• AI readiness assessment consulting firms
• Which AI consulting company should I choose?
• Why do AI projects fail?

---

Sources

All sources retrieved 2026-09-24.

• Bryan J. Weiner, "A theory of organizational readiness for change," Implementation Science 4, 67, 2009, retrieved 2026-09-24 — https://pmc.ncbi.nlm.nih.gov/articles/PMC2770024/
• Gallup, "Employee Engagement Remains Flat as AI Adoption Accelerates," July 21, 2026, retrieved 2026-09-24 — https://www.gallup.com/workplace/712433/employee-engagement-remains-flat-adoption-accelerates.aspx
• Gallup, "AI Use at Work Has Nearly Doubled in Two Years," June 15, 2025, retrieved 2026-09-24 — https://www.gallup.com/workplace/691643/work-nearly-doubled-two-years.aspx
• Pew Research Center, "U.S. Workers Are More Worried Than Hopeful About Future AI Use in the Workplace," February 25, 2025, retrieved 2026-09-24 — https://www.pewresearch.org/social-trends/2025/02/25/u-s-workers-are-more-worried-than-hopeful-about-future-ai-use-in-the-workplace/
• Berkeley J. Dietvorst, Joseph P. Simmons, and Cade Massey, "Algorithm Aversion: People Erroneously Avoid Algorithms After Seeing Them Err," Journal of Experimental Psychology: General 144(1), 114–126, 2015, retrieved 2026-09-24 — https://marketing.wharton.upenn.edu/wp-content/uploads/2016/10/Dietvorst-Simmons-Massey-2014.pdf
• Berkeley J. Dietvorst, Joseph P. Simmons, and Cade Massey, "Overcoming Algorithm Aversion: People Will Use Imperfect Algorithms If They Can (Even Slightly) Modify Them," Management Science 64(3), 1155–1170, 2018, retrieved 2026-09-24 — https://faculty.wharton.upenn.edu/wp-content/uploads/2016/08/Dietvorst-Simmons-Massey-2018.pdf
• Mark Hughes, "Do 70 per cent of all organizational change initiatives really fail?," Journal of Change Management 11(4), 451–464, 2011, retrieved 2026-09-24 — https://research.brighton.ac.uk/en/publications/do-70-per-cent-of-all-organizational-change-initiatives-really-fa/

On the page review. "The 13 readable pages" are the pages describing how consultants or firms assess AI or change readiness among the results for "how consultants assess ai readiness in organizations" and "ai change readiness assessment consulting firm" on 2026-09-24: UMU, Panorama Consulting, Centric Consulting, RSM US, Purple Shield Security, Shispare, Harvard Business School Online, CBIZ, BDO, Airiodion Group, IMA Worldwide, D2D, and Bosio Digital. Airiodion and IMA Worldwide were the two that mentioned managers. IMA Worldwide's pulse check is the 15-question self-scored survey. SkillPanel's page was blocked to automated retrieval and isn't counted. It's a snapshot of one day's results, not a market survey, and the commercial pages are cited only for what they say about method.

On the Gallup and Pew figures. The Gallup 2026 manager-support figures come from quarterly Gallup Panel web surveys in the first half of 2026 (43,262 responses) and are associations, not causal estimates. The Gallup 2025 and Pew figures come from separate surveys with different questions and samples. They're shown together to illustrate the difference between general worry and expected job loss, not combined into a single figure.

On the algorithm-aversion experiments. The Dietvorst percentages come from incentivized forecasting experiments with online and lab participants, not from workplace AI rollouts. We cite them for what they show about how people respond to having some control over an algorithm's output, not as adoption rates.

On first-party claims. Statements such as "our diagnostic" describe Avolis's own engagements. They are not independent research and are not offered as benchmarks. The millwork manufacturer is an illustrative composite, not a client, and its figures are illustrative.

---

About Avolis Research Group

Avolis Research Group is Avolis's in-house research practice, focused on how operations-heavy small and mid-sized businesses actually adopt AI. It synthesizes primary economic research, government survey data, and results from real implementations into practical, vendor-neutral guidance.

More about Avolis and how we work · Get in touch