Insight

AI Readiness Assessment for Manufacturers: A 20-Question Scorecard for Sales, Finance and Planning

By: Michael Hamilton

Share at:
Summarize with AI:
New Media

An AI readiness assessment grades how prepared a company is to get results from AI in four areas: people, process, data and systems. For a manufacturer, the useful version looks at specific workflows in sales, finance and planning, then ranks projects by their value and by how ready the business is to adopt them. The output is a short, ordered list of what to do first and what to fix before you start.

Larger companies are moving quickly. The US Census Bureau reported in May 2026 that 37% of firms with at least 250 employees use AI, compared with 19.8% of all businesses. Using AI and being ready to get value from it are two separate things, and the gap between them is where budgets get lost.

Below you will find the four areas to grade, a 20-question scorecard you can complete in under an hour, and guidance on what to do with your score. Most readiness checklists for manufacturers focus on sensors and plant equipment. This one covers the office functions where quotes, prices, invoices and forecasts are produced.

What is an AI readiness assessment?

An AI readiness assessment is a structured review of whether your company can turn an idea into a working, adopted solution. It looks at the people who will use it, the process it will change, the data it will depend on and the systems it must connect to. A good assessment is specific to a workflow. "Are we ready for AI?" has no useful answer, while "Are we ready to automate deduction research?" does.

Readiness also changes by department. A finance team with clean invoice data and a stable close process may be ready today, while the sales team in the same company records pricing in personal spreadsheets. Grading each area separately shows you where to start and where to wait. For a wider view of where companies sit, see the AI maturity ladder.

The 4 areas to grade: people, process, data and systems

Every AI project depends on the same four areas. A weakness in any one of them can stall a project that looks strong on paper. The scorecard below asks five questions in each area.

People: is there an owner and a team willing to change?

People readiness is about ownership and adoption. Someone in the business, outside IT, has to own the outcome and have the authority to change how work gets done. The people who do the work today need time to help design the new process and a reason to trust it. Projects without a named owner tend to drift, which we cover in AI only works when someone owns it.

Process: is the work stable enough to automate?

Process readiness asks whether the work follows rules that can be written down. If three people handle the same task three different ways, automation will copy the confusion. You should know the volume, the common exceptions and how long the task takes today. That baseline is also what you will measure results against later.

Data: can the project reach the records it needs?

Data readiness is narrower than most people assume. Clean data everywhere is unnecessary. What matters is that the specific records one project depends on are accessible, reasonably accurate and owned by someone. Our article on using AI when your data and systems are a mess explains how to start without a full cleanup.

Systems: can your ERP and CRM connect securely?

Systems readiness covers access and security. The solution must be able to read from and write to your ERP, CRM and document stores, through an API, an export or a controlled interface. IT needs to agree on who can see what, where data is processed and how changes are logged. Settling this early prevents the most common late-stage delay.

The 20-question AI readiness scorecard

Pick one workflow you are considering for AI, such as quoting, order entry, invoicing or demand planning, and answer the questions for that workflow. Score each question 0 for no, 1 for partly and 2 for yes. The maximum score is 40. Complete it with the people who do the work, because their answers are usually more accurate than a leadership estimate.

  1. People: Does a business leader outside IT own the outcome of this project?

  2. People: Do the people who do this work today have time to help design the new process?

  3. People: Has leadership explained why the change matters and what happens to the roles involved?

  4. People: Do the team's managers already use AI tools in their own work?

  5. People: Is there a plan for training and for collecting feedback after launch?

  6. Process: Is the process documented, or could it be written down in a day?

  7. Process: Do different people perform the task the same way?

  8. Process: Do you know the monthly volume and the time each transaction takes?

  9. Process: Can you list the common exceptions and how each is handled?

  10. Process: Is there a measurable outcome, such as cost, cash, cycle time or error rate?

  11. Data: Is the data this process needs stored in a system and not only in personal files?

  12. Data: Are customer, item and price records consistent across systems?

  13. Data: Does someone own the accuracy of each data set?

  14. Data: Is there at least a year of history for the transactions involved?

  15. Data: Can documents such as purchase orders, contracts and remittances be accessed digitally?

  16. Systems: Can your ERP and CRM share data through an API, export or integration tool?

  17. Systems: Has IT defined who may access the data and where it can be processed?

  18. Systems: Is there a test environment where a new solution can be tried safely?

  19. Systems: Are system upgrades or migrations in the next 12 months known and scheduled?

  20. Systems: Is it clear who will support the solution after it goes live?

How to read your score

Your total gives a general picture, and the lowest-scoring area tells you where to act. A workflow can score well overall and still fail because of one weak area, most often data or ownership. Use the table below to decide what to do next.

Score
What it means
Suggested next step
30 to 40Ready to buildScope the project, set a baseline and start
16 to 29Ready with gapsFix the lowest-scoring area first, then begin with a narrow scope
0 to 15Not ready yetChoose a different workflow, or prepare the people, process and data before building


Why you should score more than one workflow

Scoring three or four workflows side by side is more useful than scoring one. The comparison shows which project is ready now and which needs preparation, and it often changes the order leaders had in mind. Companies frequently find that their most exciting idea is their least ready one. A smaller project with a high score makes a better first step, because it builds skills and trust for the larger one.

How do you prioritize what the assessment finds?

A readiness score tells you whether a project can succeed, and it says nothing about whether the project is worth doing. Combine it with an estimate of value. The common approach is a two-by-two grid of impact and effort, which is a reasonable start and leaves out the factor that decides most outcomes: whether people will adopt the result.

We rank opportunities on three dimensions, which are impact, effort and adoption. A high-impact project that the team resists will return less than a modest project people want. We explain the method in why the impact and effort matrix falls short for AI.

Why adoption deserves its own score

Most manufacturers that try AI see some benefit, and few carry it across the business. In a Deloitte survey of more than 140 manufacturing organizations published in June 2026, 84% reported measurable value from AI, while only about one in five use cases had been scaled. Technology is rarely what holds the rest back. Our article on adoption gaps describes the usual causes.

What should a paid AI readiness assessment deliver?

A self-scored checklist is a good first look, and it has limits. People inside a process tend to rate it more favorably than an outsider would, and they cannot easily compare their data and systems with what a build requires. A paid assessment adds an independent review and turns the findings into a plan. Before you hire a firm, confirm that the engagement includes these deliverables:

  • Process review: interviews and observation of each function in scope, with volumes and pain points documented.

  • Data and systems review: a check of the actual records and access methods each candidate project would need.

  • Ranked opportunity list: every candidate scored on value, effort and adoption, with the reasoning shown.

  • Roadmap: a sequence of projects with owners, dependencies and rough timing.

  • Cost and benefit estimates: enough detail for a budget request and for comparing implementation proposals.

  • Independent recommendations: advice that does not depend on buying a particular product.

How long does an AI readiness assessment take?

The self-assessment above takes less than an hour per workflow. A professional assessment across several departments usually takes a number of weeks, depending on how many processes are in scope and your staff's availability for interviews. Engagements that promise a full roadmap in a few days are generally working from assumptions and have not reviewed your data.

What to do with your readiness score

The scorecard shows where you stand today. The next step is deciding what to do first, in what order and at what cost. If your scores are high, you are ready to scope a first project. If they are mixed, an outside review can separate the gaps that matter from the ones that can wait.

Ignite is OnTrac AI's 60-day discovery and roadmap program. It uncovers your highest-impact opportunities and delivers a plan your leadership team can act on. You can see examples of the results in how we integrated AI in manufacturing, or contact our team to discuss your situation.

Frequently asked questions

How do I know if my company is ready for AI?

You are ready when a specific workflow has a business owner, a stable process, reachable data and systems that can connect securely. Readiness differs by department, so assess each workflow on its own. The 20-question scorecard above gives you a quick reading.

What should an AI readiness checklist include?

It should cover people, process, data and systems. Look for questions about ownership, training, process documentation, exception handling, data accuracy, system access and support after launch. A checklist that only covers technology misses the most common causes of failure.

What is the difference between AI readiness and AI maturity?

Readiness describes whether you can succeed with a particular project now. Maturity describes how far the whole organization has progressed in using AI, from early experiments to solutions in production. You can find your current stage in which AI stage your company is actually at.

Do we need clean data before starting an AI project?

You need clean data for the project you are starting, and the rest can wait. Cleaning everything first delays results and spends effort on records no project will use soon. Fixing data one workflow at a time is faster and builds good habits along the way.