TLDR; Run a 30-minute conversation across five areas: data, process, people, tech, and budget. Most small teams are more ready than they think; the real blockers are usually where to start, who owns it, and scattered data, rarely budget. No score, no framework: a prioritised list of what to do first.
Every client I work with asks the same question, usually in the first conversation: “Are we ready for AI?”
It’s the right question. Most AI readiness assessments answer it badly.
Why most assessments are useless
If you search “AI readiness assessment” right now, the top results are Microsoft, Cisco, UNESCO, and a handful of consulting firms. They all offer the same thing: a sprawling framework designed for enterprises with dedicated data teams, IT governance committees, and multi-year digital transformation budgets.
I’ve seen these frameworks. Some run to 75 questions. They produce a score, a percentage or a maturity level, and that score doesn’t tell you anything useful. It doesn’t tell you which process to automate first. It doesn’t tell you why your team stopped using the AI tools you bought six months ago. It doesn’t tell you what to do on Monday.
The bigger problem for my clients is fit. Enterprise frameworks are often designed to cover governance, data, security, workforce, infrastructure, and portfolio management across a large organisation. That breadth can be appropriate for an enterprise and disproportionate for a 30-person firm looking for its first useful workflow. I cannot infer the authors’ commercial motive from the length of a framework; my criticism is that the output often does not answer the smaller client’s immediate prioritisation question.
Most of the businesses I work with are 10 to 200 people. Startups, agencies, professional services firms, regional businesses. They don’t need an enterprise framework. They need a conversation.
So that’s what I run instead.
The 5 areas I assess:
- Data readiness: what data exists and can your team actually access it?
- Process readiness: which tasks are strong AI candidates?
- People readiness: who’s on board, who’s resistant, who’d own it day-to-day?
- Tech readiness: what can you plug AI into without engineering work?
- Budget readiness: off-the-shelf tools or something custom?
The 5-area framework I use
Over the past few years, I’ve settled on five areas that tell me almost everything I need to know about where a team stands with AI, and where they’ll get stuck if they try to move too fast.
These are the five things that, in my experience, determine whether an AI initiative succeeds or quietly dies three months in.
1. Data readiness
The first thing I want to understand is what data this organisation has, and whether they can get to it.
The questions I ask:
- Where does your business data live? Spreadsheets? A CRM? A custom database? In someone’s head?
- Can your team access that data without asking IT or another department?
- How current is it? How often is it updated?
- Is it structured, clean rows and columns, or a mix of PDFs, emails, and notes?
I’m not looking for perfection. Most small businesses have messy data and that’s fine. What I’m trying to understand is the gap between what they have and what an AI tool would actually need to be useful.
The pattern I see most often: people assume their data problem is enormous, but when we dig into the specific use case they have in mind, the actual data requirements are much simpler. A team that wants to use AI to answer customer questions from their documentation doesn’t need a data warehouse. They need to get their docs out of a shared drive and into a consistent format.
2. Process readiness
Not every task is a good candidate for AI. Some processes are strong candidates; others are better left alone. Part of my job in this conversation is helping people distinguish between the two. Elsewhere I’ve mapped out which work fits AI and which doesn’t in more detail.
The questions I ask:
- Walk me through a task you do every week that feels repetitive or time-consuming. What does it actually involve?
- Where do you spend time doing things that feel like “moving information from one place to another”?
- Which decisions in your business rely on pattern recognition or classification? Spotting good leads, triaging support requests, categorising invoices.
- Where do you produce a lot of written output? Reports, emails, proposals, documentation.
I’m listening for two types of signal. The first is high-volume, low-variation work: things done the same way dozens of times a week. The second is content-heavy work: anything involving reading, writing, summarising, drafting.
Both are strong AI candidates. What I steer people away from: anything requiring nuanced human judgement, deep relationship context, or real-time situational awareness. AI makes a poor replacement for a senior sales conversation or a difficult client negotiation.
3. People readiness
This is the area most assessments skip entirely. It’s a mistake. People readiness is often the actual constraint, not data, not technology.
The questions I ask:
- Is anyone on your team already experimenting with AI tools on their own?
- What’s the general attitude in the team, excited, sceptical, anxious?
- Have you tried to introduce AI tools before? What happened?
- Who’d be the natural internal champion for this?
What I’m listening for is the mood. Teams with one or two genuine enthusiasts are much easier to work with than teams where the whole culture is “we tried it and it didn’t work.” Teams where the resistance is concentrated in one or two specific people (often, honestly, people worried about their jobs) need a different approach than teams where everyone is curious but nobody knows where to start.
I’m also trying to identify who I’d want to work with closely during any implementation. Almost every successful AI rollout I’ve seen has had a motivated internal person who owns it day-to-day. Without that, tools get set up, used for two weeks, and then abandoned.
4. Tech readiness
I look at the technology stack with two questions in mind: what’s easy to plug AI into, and what would require real engineering work?
The questions I ask:
- What software does your team use every day? Communication, CRM, project management, document storage.
- Are any of those tools already building in AI features you haven’t explored?
- Do you have developers on staff, or would any technical work need to come from outside?
- Do you have any existing APIs or integrations you rely on?
Most of the time, the tools people already pay for, Microsoft 365, Google Workspace, HubSpot, Notion, now have AI features baked in. Much of what people want to achieve with AI can be done without writing a line of code, just by learning what’s already in their stack.
For teams that do want to build custom workflows or integrations, I’m assessing whether they have the technical capacity in-house or whether that becomes a dependency on me or another external resource. That shapes the realistic scope of what we plan.
5. Budget readiness
AI tools are cheap. Custom AI development is not. I want to understand early where on that spectrum we’re likely to land.
The questions I ask:
- What are you already paying for software, per person per month?
- Have you budgeted anything specific for AI, or is this still exploratory?
- Are you looking for quick wins in weeks, or are you open to a longer build?
- What would “worth it” look like to you? Time saved? Revenue generated? Something else?
I’m not trying to upsell anything here. I’m trying to calibrate the plan to the reality. A team with no budget and no appetite for a long project needs a different starting point than a business committed to a six-month transformation.
The most common finding: people overestimate what a useful AI implementation costs. The highest-value first move for most teams is a $20/month tool subscription and two hours of setup time.
How I actually run this
It’s a 30-minute conversation. Not a form, not a survey, not a prepared presentation.
I ask questions and I listen. The specific questions shift based on what I hear. If someone mentions they’ve tried AI before and it didn’t stick, I want to understand that failure before we talk about anything new. If someone’s data is clearly the problem, I spend more time there and less on the tech stack.
I take notes, actual notes, not a scoring matrix. I’m not trying to produce a maturity level. I’m trying to understand a specific organisation well enough to give them useful recommendations.
By the end, I almost always have a clear picture. Not a complete one, there are always things you only discover once you start, but clear enough to prioritise.
One thing I’ve found consistently: these conversations go better when the person I’m speaking to has operational responsibility, not just decision-making authority. A CEO who knows the strategy but not the workflows can’t answer the questions I need to ask. Ideally I’m talking to the person who does the work, or who manages the people who do.
What I typically find
After running this with enough teams, certain patterns repeat.
Most teams are more ready than they think. The anxiety about AI readiness is usually worse than the actual situation. They have data, processes, at least one enthusiastic person. What they’re missing is a clear starting point and someone to help them not waste the first three months on the wrong thing.
Data is the most common blocker, but rarely in the way people expect. It’s not usually that the data doesn’t exist. It’s that it’s scattered across multiple tools, in PDFs, locked in someone’s head. The fix is structural, not technical: getting documents into a consistent format, consolidating contacts into one CRM, building a simple knowledge base. Mundane tasks that open up a surprising amount of AI capability.
The biggest single barrier is not knowing where to start. Teams get paralysed by the breadth of what AI can theoretically do, and they either try everything at once (spreading too thin) or wait for a clearer sign that AI is worth their time (waiting indefinitely). A prioritised starting point breaks both patterns.
Skills gaps are almost never the blocker people think they are. People worry that their team isn’t technical enough, isn’t young enough, isn’t digitally savvy enough. In practice, I’ve helped teams of 50-year-old tradies adopt AI writing tools, and teams of digital natives fail to adopt anything because the culture wasn’t right. Technical skill is far less predictive of success than attitude and motivation.
The output: a list, not a score
At the end of the assessment, I give the client a prioritised list of opportunities. Usually three to five specific recommendations, ordered by a combination of expected impact and implementation ease.
It looks something like this:
- If your existing Microsoft 365 licence includes the required Copilot features, pilot it on weekly report drafts before buying another tool. No additional software cost in that case; allow time for permissions, templates and review rules rather than promising a fixed setup duration.
- Build a simple internal knowledge base from your SOPs so your team can query it directly instead of searching through shared drives. Medium effort, high value for onboarding and support.
- Explore AI-assisted lead scoring once your CRM data is more consistent. Medium-term, worth revisiting in three months once you’ve cleaned up the data.
No scores, no maturity levels, no radar charts. A clear answer to: “What should we actually do, and in what order?”
If they want to go further, I can help implement any of those through a short Applied AI engagement or as part of ongoing fractional CTO support. The assessment itself stands alone as a useful thing.
Who this is for
The conversation I’ve described works best for teams of 10 to 200 people trying to figure out their first serious move with AI. It’s for businesses that have heard enough to know AI is relevant but haven’t found a practical entry point.
It’s not the right fit if you’re a 2,000-person enterprise with an existing AI governance structure and a dedicated data team. For that, the Microsoft and Deloitte frameworks probably make sense. They were built for you.
If you’re a mid-size professional services firm, a growing agency, a regional business, or a startup that wants to move fast without wasting time on the wrong thing, this is the conversation I’d want to have with you.
If you want to run this assessment for your team, it’s now part of my Applied AI work; get in touch to start.
If you want a broader sense of how I think about helping teams adopt AI before you get in touch, this piece on using AI at work without the hype covers a lot of the same territory from a different angle.
Common questions
What is an AI readiness assessment?
A short, practical check of whether a team can actually adopt AI usefully, not an enterprise framework. I use five areas and about thirty minutes to find where AI will help and where it will not, then hand back a prioritised list of next steps rather than a score.
How do you assess whether a team is ready for AI?
I look at five areas in a thirty-minute conversation: data, process, people, tech and budget. The point is honest signal about where to start, so small and mid-size teams get a plan they can act on rather than a maturity grade.
What do you usually find in these assessments?
That the barrier is rarely the technology. It is more often unclear workflows, messy or unowned data, or no one accountable for adoption. The output is a short list of the highest-value, lowest-risk places to begin, not a readiness number.