Adaptability

What this dimension measures

This dimension looks at what happens to the surrounding work when something new becomes possible. When a tool makes a step unnecessary, does the step disappear, or does it stay because the procedure says so? When a task can be done by a different role, does the responsibility move, or does the original role keep doing it alongside the new tool? When an assumption built into a process turns out to be outdated, does anyone revise the process?

The assessment asks respondents to recall recent changes to the normal workflow: something that was removed, reordered, or reassigned because a new tool made another way possible, and which procedure or role had to change as a result. Evidence of this kind is more telling than general openness to change, because it shows the organization actually redrawing its work rather than talking about doing so.

An organization with high Adaptability tends to revise procedures after a trial shows a better way and to move responsibilities between roles without a crisis. An organization with lower Adaptability may run the same trials and reach the same conclusions, but the production process stays as it was. The tool gets added; the work around it does not change.

This is distinct from the ability to run trials in the first place. That belongs to Experimentation & Learning.

Why this matters for AI

The common hope for AI is that it will do existing work faster. Sometimes that is true. More often, the value arrives only when the process changes around the capability. A tool that drafts a document in seconds adds little if the document still passes through the same five reviewers it did before. A tool that answers routine questions does not reduce anyone's load if the people who used to answer them are still required to check every reply.

This is why Adaptability matters for AI in particular. AI tends to make particular steps cheap, which shifts where the real work is. Review becomes more important than production; judgment becomes more important than execution; the hand-offs between roles change. An organization that can redraw its workflows can capture that shift. An organization that cannot ends up layering AI onto an unchanged process, where the gains are absorbed by the steps around it.

There is a second reason. AI capabilities keep moving. What a tool can do this year is different from last year, and the process that made sense around the earlier version will not fit the later one. Adoption is therefore not a single change but a series of them, and each one asks the organization to revisit how work gets done.

Finally, changes to workflow are changes to people's jobs. Responsibilities move, some tasks shrink, others appear. An organization that has already shown it can move responsibilities without damage has a path through this. One that has never done so will meet it for the first time under pressure.

How this can work in your favor

When your responses describe steps removed and responsibilities moved as new tools changed what was possible, AI adoption has a clearer route from trial to normal work.

Successful trials become real changes. The hardest moment in many AI efforts is not the pilot but what comes after it, when a promising result has to be turned into a changed process. An organization with a habit of redrawing its work can take that step; the pilot does not stay a pilot forever.

Value shows up where it was expected. Because the surrounding steps change, the time or effort an AI tool saves is actually released, rather than being quietly consumed by the unchanged procedure around it.

Roles can absorb new responsibilities. When review and judgment become more important than production, people's jobs shift. An organization used to moving responsibilities can make that shift deliberately instead of leaving people to work it out alone.

Each round of change costs less. AI will keep changing what is possible, and an organization that has reshaped its work before can do it again with less disruption. Adaptability is itself a capability that compounds.

How this can make AI adoption harder

When your responses suggest that procedures and role boundaries stay put even when a new tool makes them awkward, a specific kind of friction tends to appear.

The tool arrives; the work stays the same. AI gets layered onto an unchanged process. The step it was meant to replace continues, now with a tool in front of it, and the organization pays for both. People experience AI as an extra thing to do rather than a different way of working.

Pilots stay pilots. A trial can succeed on its own terms and still fail to change anything, because changing the production workflow is a separate, harder act. Successful experiments accumulate alongside normal work rather than replacing any of it.

Gains leak away. Any time an AI tool saves is absorbed by the surrounding procedure. Reports that the tool "did not really help" are common here, even when it worked exactly as intended.

Responsibilities get stuck. When AI makes a task cheap, the role that used to own it keeps it, and the judgment work that AI creates has no owner. Both sides of the change are left unfinished.

This friction does not mean AI cannot be adopted. It means that, in this environment, the hard part is likely to come after the pilot rather than before it, and that the real question each time will be whether the process is allowed to change.

What this dimension is not

Adaptability is not a willingness to try things. An organization can run frequent pilots and still be unable to change the production workflow once a pilot succeeds; that distinction is exactly why the assessment keeps Adaptability separate from Experimentation & Learning.

It is not instability or constant reorganization. An organization that reshuffles every quarter is not necessarily adaptable in this sense; the question is whether work changes when a new capability makes a better way possible, not whether change happens for its own sake.

It is also not a judgment on stable, well-optimized processes. Many organizations are successful precisely because their processes are hard to change. That becomes a readiness consideration only when AI adoption requires those processes to move.

What we look for

The assessment favors recalled changes to actual work over statements about flexibility. Signals that tend to be informative include:

• A specific step in the normal workflow that disappeared because a new tool made it unnecessary.
• A responsibility that moved from one role to another when a tool changed who could do the task.
• A procedure that was rewritten after a trial showed a better way, rather than left as the official version while people worked around it.
• Whether, when a tool made a procedure awkward, the procedure changed or the tool was abandoned.
• Whether role descriptions and approval chains have been revised in recent months, or have stayed as they were through several new tools.

"We are agile" and "we embrace change" are not evidence. One concrete example of a role or procedure that was redrawn around a new capability tells the assessment more than any amount of self-description.