The Four Legs of AI Readiness
Clean data, documented processes, clear priorities and a team with room to change: what each one means, a quick test for yours, and where to start.
Before AI can work across a team, a few things have to support it, like the legs of a table. From the teams I've worked with, I group them into four: clean data, documented processes, clear priorities and decisions, and a team with the time to take on something new. Others would split it differently, but these four are where I keep seeing teams get stuck. There's no set order to work through them, and you don't have to fix all four at once.
I'm Malcolm, and I run Optimi. Each leg is an area of how your firm runs: two cover your systems, and two cover your people. If one area is weak, the whole table wobbles.
The good news is that AI makes a lot of this easier. Tools like ChatGPT or Claude can turn a messy explanation into a clear checklist, or help tidy a spreadsheet, in minutes. And most of this work makes your week easier anyway, whether or not you go on to use AI across the team.
Just finished the assessment? Jump to your weakest area below. Your results page gives you a simple first move, which may differ from the ones here.
Clean Data · Documented Processes · Priorities & Decisions · Team Capacity
Clean Data
What it is: Whether the information your work runs on is accurate, organised and easy to find.
The test: Can anyone see the current state of a client, project, task or job without asking someone, or opening multiple tools?
When it's weak: The same data gets keyed into two or three places, and nobody is sure which version is right. Even inside one system, the same item might turn up multiple times, or with details that are out of date. New tools get bolted on top of the mess, and each one adds another place to check.
I worked with a foundation tracking fundraising across over 1,200 offices in one spreadsheet. Lots of different people needed to see different parts of it, and nothing off the shelf could sit on top of that spreadsheet and give them what they needed. In the end it had to be rebuilt as something custom-made. The lesson stuck with me: when the state of things isn't easy to see, you can't bolt tools on top.
A first move: List every tool you paid for last month and cancel any nobody has opened in the past three months. Decide the one place your client and job information will live from now on. Don't move anything yet.
Documented Processes
What it is: Whether the steps for doing your work are written down where people can find them.
The test: Could a new hire run your most common process from the page, without asking anyone?
It's a fair test for AI too. If a new hire couldn't follow your process from the page, an AI agent couldn't either.
When it's weak: The work stops when one person is away. Fixes get layered on fixes, because nobody wrote down how it was meant to work.
I often see teams running dozens of automations built up over the years, and nobody can say which ones the work depends on. When one breaks, the founder has to be there to fix it. Until the work is written down, there's no way to tell which to keep, or which ones double up.
A first move: Pick the one process that matters most and would stop if the person who knows it were away for a week. Write down how it's done today as a one-page checklist.
You can use an AI tool to help write it: talk the process through out loud, then have the AI turn what you said into steps.
Priorities & Decisions
What it is: Whether priorities are clear and it's obvious who decides what, rather than everything feeling urgent at once.
The test: Ask three people what matters most this week. Do you get the same answer, and do they know who made that call?
When it's weak: There's a lot of chaos, a lot of reacting and a lot of last-minute work. Everything seems urgent all the time.
I've seen the same pattern in businesses and in community groups. When everyone owns something, no one owns it, and it soon falls over. When one person owns a call, everyone else can relax on that thing because they know someone's got it, so they can get on with their own work.
Think of a busy kitchen, where each cook owns a station. Nobody grabs the same pan, nobody leaves the sauce to burn, and nobody has to watch the whole room, so the food goes out on time.
A first move: Every Monday, one named person writes down the three things that matter this week. Nothing new gets added mid-week unless they agree.
Team Capacity
What it is: Whether your team has the time and skills to take on something new and look after it.
The test: If a new way of working landed on Monday, who has the hours to learn it and look after it?
When it's weak: Questions go unanswered because nobody has the time. Past changes were started and never stuck.
This is one of my worst worries. It's why Optimi doesn't take on a client unless someone on the client's own team owns the systems. Without that person, it doesn't last. As I put it to clients: "It doesn't matter how fancy the thing is we build, no one's going to maintain it."
An outside partner can't be the first line of support, because some of the context only lives inside your team. What we're aiming for is tech a team can run themselves, so it keeps working for them long after we've stepped back. That only happens when someone inside the team owns it.
It turns into a loop. You can't improve what a team does if they don't have time to improve it, and I see it most in lean teams where everyone is already stretched.
A first move: Protect 90 minutes a week for four weeks: one person, same slot, spent only on writing down the process that broke most recently.
You don't have to fix all four at once
There's no fixed order. Start with your weakest area, because it's usually where a small move makes the most difference, not because it's the cause of everything else.
Sometimes two areas come out even. To break that tie, start with the people: Team Capacity comes first, then Priorities & Decisions, then Documented Processes, then Clean Data. That order is only a tie-breaker, not the only path.
When all four areas score the same, there's no weakest one to name. If all four score low, start with your team's time, because it's the cheapest place to begin. If all four sit in the middle, pick the move you're most likely to finish. Being steady like that is a fine place to be for a while, because nothing there needs fixing urgently.
All four solid means there's nothing to fix! Yah. Protect what's working by keeping the docs reviewed, the owners named and the hours protected, and take the assessment again if you hire quickly or a key person leaves.
There's a lot of work here in total, but nobody does it all at once.
Your next step
Haven't taken it yet? Take the free AI Readiness Assessment. It's 17 questions and takes about 5 minutes. You'll see your results straight away, and for most people that means a weakest area and a first move to start on.
Already have your results? They're on Optimi Insights, in the tab you came from. If you'd like me to read your answers and provide insight, use the "Want Malcolm's personal take?" form at the bottom of your results page.
If one area is weaker than the rest, one small move this week is enough to start. Pick yours and see what it does.