How to run your first AI audit in an afternoon
Somewhere in your business there is work nobody enjoys, work that takes too long and holds your team back from what really matters. But if you want to use AI, where do you even begin? And how do you avoid months of analysis before anything actually happens?
The answer is simpler than you might think. You don’t need an elaborate consulting engagement to get a first picture. With a free afternoon, a spreadsheet and the four steps below, you can map out your biggest opportunities. This is the same approach I use when I run a full AI audit for a client, just the lighter version you can do yourself today.
Step 1: list the tasks per role
Grab your org chart, or simply the list of roles on your team. Walk through each role and write down every recurring task that person performs. Don’t think in projects, think in actions. Not “customer service” but “read incoming emails, categorise them and assign them to the right colleague.” The more specific you are, the more useful the list becomes.
Rather take the first step together? The free AI scan shows you where to start.
Book a free intro call →Ask your team the same question: “Which task would you love to never do again?” You will get answers you hadn’t thought of yourself. Write everything down, including the small stuff. Especially the five-minute tasks that happen twenty times a day add up fast.
Aim for at least five to ten tasks per role. With a team of twenty people that gives you roughly a hundred lines. It feels like a lot, but that is exactly what you want. You will filter it down in the next step.
Step 2: score each task on four criteria
Add four columns next to your task list and rate each task from 1 to 5 on these criteria:
- Repetitive: how often is this task repeated? Daily and in the same way scores a 5.
- Time-consuming: how many hours per week does your team spend on it in total?
- Error-prone: how often does it go wrong? Think typos, wrong data, missed deadlines.
- Not scalable: if your company doubles in size, does the headcount for this task double too?
Add up the four scores. The maximum is 20. Sort your list from high to low. The tasks at the top are your biggest opportunities. Those are the spots where an AI agent can make the most difference.
Don’t overthink this. A rough estimate is good enough. You are looking for the ranking, not decimal-point precision.
Step 3: calculate the hours and the cost
Take your top ten and estimate how many hours per week your team spends on each task. Multiply by the average fully loaded hourly rate. That gives you a weekly figure. Multiply by 48 working weeks and you have an annual number.
A quick example: suppose three employees together spend eight hours a week manually transferring data between two systems. At a loaded hourly rate of 45 euros, that is 360 euros per week, or over 17,000 euros per year. Suddenly it becomes very concrete whether a solution is worth the investment.
You don’t need to do this for all hundred tasks. Just the ones that scored highest in step 2. That top ten gives you a clear view of what is on the table.
Step 4: pick your top 3 and set the order
You now have a sorted list with scores and a euro value. Choose three tasks to start with. As you pick, consider two extra questions:
- How complex is the task to automate? A task with clear rules and digital input is easier than something that requires you to digitise paper-based processes first.
- How big is the impact on your team? Start with something that gives people immediate breathing room. That builds buy-in for the next step.
Rank your top 3 from easiest to hardest. Your first project should be a quick win: something that is running within weeks and delivers visible results. That builds confidence, both for you and for your team.
What does this give you?
After one afternoon you have three things most companies lack: a complete overview of repetitive work, a scored and ranked shortlist, and a starting point with a clear business case. You don’t have to tackle everything at once. You just need to know where to begin.
In a full AI audit I go deeper: I also look at the connections between tasks, at which agents can work together and where a human should step in to make the call. But this light version already gets you surprisingly far.
Want to walk through it together?
If you have filled in your spreadsheet and want to talk through the results, or if you would rather have me map out the full picture: book a free intro call. We will look at your top 3 together and figure out what it takes to get your first agent up and running.
Ready to handle more?
Book a free call. Together we look at where AI makes the difference for your people, no strings attached.