The AI Readiness Assessment

You can buy the tools and train the team without knowing whether the firm is working differently yet. This assessment asks what changed in the work and the business.

Sixteen questions place your firm on The Delivery Model Ladder. They cover four business numbers, where any free capacity went, two results, and eight parts of daily operations. The result gives you a stage and up to four next moves.

The assessment is for established businesses with teams and repeatable processes. Its delivery measures fit customer work; use it for the part of your business those measures describe. It takes less than three minutes. It runs in your browser, stores nothing, and asks for no email. If the numbers are not being tracked, it says that instead of guessing.

Why using AI does not mean the firm is ready

Tool use and business change are different things. A Goldman Sachs survey of 1,256 small-business owners in early 2026 found that 76% used AI, while 14% had put it into core operations.

Another sign comes from the April 2026 Census Bureau working paper. Among firms using AI, 66% only added it to existing tasks and 64% made no organizational changes.

That is why this assessment does not count tools, pilots, or logins. A person can use AI every day while the firm still sells, staffs, and delivers work the same way.

And you need a way to check what happened. Cisco reported that only 32% of organizations had a process for measuring the effect of their AI work. If the firm cannot see what changed, it cannot tell whether a rollout worked.

What should an AI readiness assessment measure?

It should measure two things: what changed in the business and whether the firm can keep running the new process.

This assessment begins with The Four Numbers: cost to acquire, cost to deliver, retention, and price. Margin and revenue per person show whether those changes reached the firm. One more question asks where any free capacity went.

Then eight questions look at how you run the work. They cover ownership, a written boundary, learning time, rebuilt workflows, shared knowledge, written decisions, changes customers noticed, and review before work ships.

The first question asks you to guess your stage. The guess does not change the result. It only lets you compare how the firm feels with what its numbers show.

How the business questions place your firm

These seven questions cover the four numbers, the use of free capacity, and the two results.

Question	Why it matters
Has the total cost of winning one new customer dropped in the last two quarters?	Cost to acquire shows what the firm spends to win one customer. Measure customers won, not leads or content produced.
Has your cost to deliver a typical engagement dropped in the last two quarters?	Cost to deliver shows whether one unit of service became cheaper to produce. Hours saved do not count unless the unit cost changes.
Are customers staying longer, or buying again more, than a year ago?	Retention shows whether customers noticed enough improvement to stay longer or buy again. Sales quality can affect it too.
Have you changed what you charge for, not just the rate?	Price shows whether the firm changed what it sells. In an hourly firm, faster work may lower the bill unless the firm changes the unit.
Of the hours AI freed last quarter, where did they go?	This question checks whether the free capacity has a job. It gives an earlier answer than financial results, which can take several quarters.
Has margin improved in a way you can tie to how work gets delivered?	Margin shows whether changes in the four numbers reached the business. Read it after both acquisition and delivery costs.
Is revenue per person growing while headcount stays flat?	Revenue per person shows whether growth still requires hiring at the same rate. It helps confirm Stage 3 but does not place any earlier stage.

Each financial question lets you say that the number moved, stayed flat, or is not tracked. If two or more of the six financial measures are not tracked, the assessment does not assign a stage.

That does not mean your firm has failed, it means the first job is to measure cost to acquire, cost to deliver, retention, price, margin, and revenue per person. Without those numbers, I would just be guessing at your stage.

Why the assessment asks how the firm runs

The financial questions place your firm, and the eight questions about daily work help explain what may be keeping it there. Those answers shape the plan.

Question	Why it matters
Does one named person own AI at your firm?	One named owner keeps useful experiments from ending as private habits or abandoned pilots.
Is there a short written AI policy people can actually find?	A short, findable rule tells people which tools and data are allowed. Without one, people make their own rules.
Is time to learn and rebuild workflows scheduled on the calendar?	People need protected time to learn the tool and change the work. Otherwise the old process keeps winning.
Has any workflow been rebuilt around AI, with steps removed or changed?	Stage 2 begins when the workflow changes, not when AI is added on top of the old steps.
Is your firm’s knowledge in one place an AI tool can reach?	Shared knowledge is required for Stage 3. Agents cannot use context that remains in individual inboxes and heads.
Could someone follow last month’s key decisions from what’s written down?	Written decisions and processes give people and agents something to follow and something reviewers can check.
Have customers noticed a change in your work?	Customer comments, reviews, and renewals show whether better internal work became better delivery.
Who checks AI output before it ships?	Checking AI output takes practice, and reviewers may face more work and different mistakes than before. The review process needs to keep up.

Three questions matter early: who owns the work, whether the boundary is written down, and whether people have scheduled time to learn. None requires a new purchase.

Review matters too. Checking AI output is its own skill. If output rises but review does not change, the reviewer becomes the bottleneck or mistakes reach clients. The assessment asks who checks the work and whether that process can handle the volume, because you can afford more agents, but you can't afford to watch them.

How the placement works

You do not earn points for buying a tool or get a percentage to compare with another firm. The rules run in this order:

1. Not enough measurement. If two or more financial measures are not tracked, the firm stays unplaced. The plan starts with the numbers.
2. Stage 3: AI-native. Revenue per person grows faster than headcount, the firm's knowledge is available to AI, and the Stage 2 pattern is already present.
3. Stage 2: Augmented. Retention improves or the firm changes what it sells, and margin follows.
4. Stage 1: Enhanced. Cost to acquire or cost to deliver falls, while retention, price, and margin stay flat.
5. Stage 0: Assisted. None of the business numbers has moved yet.

One Stage 2 number can move without the full pattern. For example, retention may improve while margin stays flat. The assessment keeps the firm at Stage 1 and says that it is close to Stage 2. The single number may have changed for a reason unrelated to delivery.

The free-capacity question does not change the stage. It gives an earlier warning. Financial results can take quarters, but the firm can decide this week whether saved time has a job.

The assessment never asks how often people use AI. Usage alone does not prove that delivery changed.

How the next-step plan is built

The result includes no more than four moves. It uses the answers marked “no” or “partly,” then keeps the moves that matter for the firm's current stage.

• Toward Stage 2: name an owner, write the boundary, schedule learning time, build review into the work, rebuild one workflow, decide where free capacity goes, and make the change visible to customers.
• Toward Stage 3: write down decisions and processes, put company knowledge where AI can reach it, define work that can be checked, and measure quality beside output.

I am not going to tell you to buy something, because no Ladder stage depends on a purchase. And if you have already made most of these changes, the plan stays short.

How is this different from an enterprise assessment?

Enterprise assessments answer a different question. Microsoft's assessment takes about 45 minutes and scores seven areas. Cisco compares an organization with thousands of large companies. These tools cover data systems, governance, talent, and infrastructure.

Consultant-led assessments can go much deeper. One published price guide puts them at $15,000 to $175,000. That work may be useful when a firm is planning a large program.

This assessment has a smaller job, helping you see where delivery stands now and what to work on next.

Assessment	Time	Cost	What it covers
Microsoft AI readiness assessment	About 45 minutes	Free	Seven enterprise IT areas
Consultant-led assessment	Several weeks	$15,000 to $175,000 in the cited guide	A detailed roadmap
This assessment	Under three minutes	Free, no email	A Ladder stage and up to four next moves

How long might the next stage take?

The result shows a rough range based on three conditions: a named owner, scheduled learning time, and a written policy. The current bands are 6–9 months with all three, 9–12 months with two, 12–18 months with one, and 18+ months, if at all with none.

These ranges are a judgment based on what I have observed across a franchise network. They are not a forecast or a promise. A stage change also depends on the starting point, the workflow, the people involved, and whether the firm keeps making the required decisions.

What this assessment cannot tell you

The answers are self-reported. For a stricter result, use the last two quarters of financials instead of memory.

The assessment cannot verify quality. A Stage 2 or Stage 3 result means the pattern is present. It does not prove that output increased while rework, complaints, or errors stayed flat.

It also cannot make the operating decision. Moving from Stage 1 to Stage 2 requires someone with authority to change the workflow. The result can show what is missing, but the firm still has to do the work.

---

The AI Readiness Assessment uses the three Proofwork frameworks: The Delivery Model Ladder for the stages, The Four Numbers for measurement, and The Production Gap for the ways firms get stuck. Nothing you enter leaves your browser. A shared link contains only your answers. I write about these operating patterns in the newsletter.