Skip to content
IronbridgeAI
All insights

Small Business AI

AI Readiness Assessment for Small Business: 9 Questions to Answer First

By Jordan SolenderUpdated 9 min read

An AI readiness assessment is a short, structured check of where a business loses time to manual work and whether its data and habits can support automation. For a small business it does not need to be a consulting project. Nine questions across four areas will tell you where AI would pay back first and what has to be fixed before it can.

This guide gives you the nine questions, a way to score them yourself and what each low score means. They are the same nine scored questions in our free Automation Scorecard, which does the math for you in about three minutes.

What an AI readiness assessment should tell you

A useful assessment answers three things: where the hours go, which job to automate first and what would stop it from working. If it ends with a maturity label and no next step, it has not helped you.

Many published readiness frameworks were written for large enterprises. They ask about data governance boards, model risk committees and centers of excellence. A company with 20 or 200 people has none of those and does not need them to start. What it needs to know is simpler:

  • Which repeat tasks consume the most skilled hours each week
  • Whether the information those tasks rely on lives somewhere a system can reach
  • Whether anyone on the team has run an AI system beyond a chat window

The nine questions below cover exactly that.

The four areas to assess

Small businesses lose most of their recoverable time in four places. Assess each one separately, because a strong score in one area often hides a weak one next to it.

Area What it covers Questions
Revenue follow-up Speed to lead, missed calls, quote follow-up 1 to 3
Customer support Repeat questions your team has answered before 4
Back office Re-typing, billing and reconciliation, weekly reporting 5 to 7
AI readiness Where your data lives and what you have tried so far 8 and 9

The nine questions

Answer each question with the option that describes a normal week, not your best week. The first option listed scores 100, the second 67, the third 33 and the last 0. A higher score means that part of the business is already automated or well prepared.

Revenue follow-up

1. When a new lead comes in, how fast does someone respond? Under 5 minutes, within the hour, same day, or next day or later.

2. What happens to calls after hours or when your team is busy? Answered live or by AI, voicemail returned the same day, voicemail returned the next day, or often missed.

3. How are quotes and proposals followed up? An automated sequence, tracked in a CRM, when someone remembers, or rarely.

Customer support

4. How many customer questions are ones you have answered before? Few, some, most, or nearly all. This one runs the other way: the more repeat questions you get, the lower the score, because more of the work could be handled by a system.

Back office

5. How often does your team re-type the same information into more than one system? Never, weekly, daily, or constantly.

6. How are invoicing, billing and reconciliation handled? Fully automated, mostly automated, mostly manual, or fully manual.

7. How are your weekly numbers and KPIs produced? A live dashboard, partly automated, built by hand each week, or you do not have reliable reports.

AI readiness

8. Where does your core business data live? One main system, two or three connected systems, many disconnected tools, or spreadsheets and inboxes.

9. What has your company done with AI so far? Systems running in production, pilots or experiments, some ChatGPT use, or nothing yet.

How to score it

Average the scores inside each area first, then average the four area scores. That keeps the three revenue questions from outweighing the single support question.

  1. Revenue follow-up: average of questions 1, 2 and 3.
  2. Customer support: the score for question 4.
  3. Back office: average of questions 5, 6 and 7.
  4. AI readiness: average of questions 8 and 9.
  5. Overall score: average of those four numbers.

Read the overall score this way:

Overall score Band What it usually means
70 to 100 Ahead of most The basics are automated. The gains left are in specific, higher-value workflows.
40 to 69 Room to grow Some systems are in place, with clear manual gaps between them.
0 to 39 Big opportunity Most repeat work is still done by hand. The first project should be small and easy to measure.

The overall number matters less than the three lowest individual answers. Those are your starting list.

What each low score points to

Each question maps to one job that AI or automation can take on. If a question scored 33 or 0, this is the fix it points to.

Low score on The job to automate
1. Lead response time An agent that replies to every new lead in seconds, day or night, and books them
2. Missed and after-hours calls A voice agent that answers, captures the details and routes the call
3. Quote follow-up A sequence that follows up on every open quote until it is won or closed
4. Repeat support questions A support agent that answers known questions and escalates when unsure
5. Re-typing between systems A sync that takes information entered once and pushes it everywhere it is needed
6. Billing and reconciliation Automation that invoices, matches payments and flags exceptions
7. Weekly reporting A report that assembles your numbers and delivers them on a schedule
8. Scattered data Consolidation into one source of truth that people and AI both work from
9. No AI in production One well-chosen workflow taken from idea to live, to build the habit

If you want more detail on any of these, see our guides on how to build an AI sales agent, AI voice agents for customer support, how AI agents deflect support tickets and how to automate data entry with AI.

Estimate what the manual work costs

A readiness score tells you where to look. A cost estimate tells you whether it is worth acting on. You need three numbers:

  1. How many people spend real time on repetitive work
  2. About how many hours each of them spends on it per week
  3. What an hour of their time costs you, fully loaded

Multiply people by hours by 52 to get the yearly hours. Then apply a realistic share, because no system removes all of the work. Our ROI calculator starts at 40 percent and $55 an hour, and you can change both. Those are starting assumptions for a rough estimate, not a forecast for your business.

A worked example: 5 people at 10 hours a week is 2,600 hours a year. At a 40 percent share, that is 1,040 hours. At $55 an hour, that is $57,200 a year of work a system could take on. Use your own wages and be conservative with the share. Our guide on how to measure the ROI of AI automation covers how to check the estimate against real results after launch.

What to fix before you automate

Two findings should change your order of work. Fix them first, or at least alongside the first build.

Scattered data (question 8). An AI system can only act on information it can reach. If customer records live in inboxes and personal spreadsheets, start by choosing one system of record for the workflow you want to automate. You do not need to clean everything. You need one reliable place for that one process.

No one owns it (question 9). A business that has only used chat tools has not yet had to decide who reviews AI output, who fixes it when it is wrong and what data it may touch. Decide those three things in writing before the first system goes live. A one-page AI use policy is enough to begin, and our AI use policy template gives you one to copy.

Everything else on the list can usually be built in the order of payback: largest cost of manual hours first, easiest to measure first when two are close.

These two gaps are also behind most stalled projects. Our guide to why AI projects fail in small business lists the other mistakes to avoid on a first build.

When a self-assessment is enough, and when it is not

A self-assessment is enough to pick your first project and to have a sharper conversation with any vendor. It is not enough when the work touches regulated data, when several systems must be connected, or when you cannot tell which of two processes is the real bottleneck.

In those cases a paid assessment from a consultant can be worth it. Published prices for those start around $1,500 and rise with scope, as covered in our guide to what AI consulting costs. Before you pay for one, ask what you will receive at the end. A ranked list of projects with reasoning is useful. A maturity score alone is not.

Take the assessment

The Ironbridge Automation Scorecard asks these nine questions plus three more about team size, hours and revenue. It returns your overall score, a score for each of the four areas, the three jobs worth automating first and an estimate of the hours and dollars involved. It is free and takes about three minutes. We email your results, and our assistant may follow up by text or phone to walk you through them.

Frequently asked questions

What is an AI readiness assessment?

An AI readiness assessment is a structured review of how prepared a business is to use AI. For a small business it should cover where manual work is concentrated, whether the data those tasks need is reachable, and what experience the team already has. The output should be a short, ranked list of what to automate first.

How long does an AI readiness assessment take?

A self-assessment takes a few minutes if you know how your business runs week to week. Our Automation Scorecard takes about three minutes. A paid assessment from a consultant typically runs from one to several weeks, depending on how many systems and teams are reviewed.

Do I need clean data before using AI?

No, not across the whole business. You need reliable data for the specific workflow you are automating. Pick one process, decide which system holds the truth for it and start there. Waiting for perfect data everywhere mostly delays the first project.

What is a good AI readiness score?

On the nine questions in this guide, 70 or above means the basics are already automated. Between 40 and 69 means there are clear manual gaps. Below 40 means most repeat work is still done by hand. A low score is not bad news. It means the first project has more room to pay back.

What should a small business automate first?

Start with the lowest-scoring question that also costs the most hours. For many businesses that is lead response or re-typing data between systems, because both happen every day and are easy to measure before and after. Our small business AI adoption playbook walks through choosing a first project.

Is an AI readiness assessment the same as an AI audit?

The terms overlap. A readiness assessment usually asks whether you are prepared to start and where. An audit usually reviews AI or automation you already run, including accuracy, cost and risk. A company with nothing in production needs the first one.

If you would rather talk it through, book a strategy call. We will tell you what we would build first and why. AI consulting for small business explains how we work with smaller companies.