How to choose the right AI projects for your business
TL;DR
Most businesses start AI in the wrong place, chasing flashy tools before fixing real problems. The right first project solves something that already frustrates your team, can be measured within 60–90 days, and carries low risk. A UK construction company cut weekly reporting from four hours to ten minutes with a simple summarisation tool. That is the template: start small, prove value, then build.
Why does choosing the right AI project matter so much?
Your first AI project sets the tone for everything that follows. When it succeeds, it builds belief across the business, your team sees what is possible, your leadership backs the next investment, and momentum compounds. When it fails, hesitation sets in and AI gets written off as 'not for us.' The difference between those two outcomes is almost always project selection, not technology. That is why it pays to start with a clear framework for deciding where to focus.
What makes a great first AI project?
A great first project has three essential qualities: it solves a real business problem, it is easy to measure, and it carries low risk with high potential reward.
Solves a real problem. Start with something that already frustrates your team or your customers, too many repetitive emails, manual reporting, or customer delays. When AI fixes something people genuinely care about, adoption happens without a change management campaign.
Easy to measure. You need to show success quickly. Choose a project where you can track time saved, accuracy improved, or satisfaction increased. Visible progress keeps everyone motivated and builds the business case for the next project.
Low risk, high reward. Avoid areas involving sensitive data or complex, interconnected systems in your first attempt. Pick projects that can be tested safely before scaling, that way you learn without creating disruption.
What does a successful first AI project actually look like?
A construction company in the UK began its AI journey by automating weekly progress reports. Before AI, this process took four hours per manager every single week. With a simple data connection and summarisation tool, the same reports now take ten minutes.
It was not a flashy project. It created instant relief for the team and proved that AI could save real time and money, and that success opened the door to more ambitious ideas. That is the kind of first win every business should be looking for.
See where AI fits in your business. Free.
A 45-minute audit. We map the highest-value automations and what they're worth in time and money. No pitch, no pressure.
How do the three lenses of AI project selection work?
Before committing to any project, look through three lenses.
The Business Lens. Does this project align with your goals for revenue, service quality, or efficiency? If it does not move a business metric you care about, it is a side project, not a priority.
The Data Lens. Do you already have the data needed to make it work? If your data is messy or incomplete, start smaller or clean it first. Garbage in still means garbage out, even with AI.
The People Lens. Will this make life easier for your team or customers? The best projects help people, they do not replace them. If the people affected cannot see the benefit, you will fight adoption at every step.
How quickly should a first AI project show results?
Aim for results within 60 to 90 days. You do not need to automate everything at once, you just need to show that AI can work in your specific environment and that your people can benefit from it.
Once that belief takes hold, you can move to bigger, more strategic ideas with the confidence of a proven win behind you. Small wins build momentum, and momentum is what turns an AI experiment into a genuine business transformation.
What is the biggest trap businesses fall into with AI?
Trying to design the perfect AI system before taking a single step. AI is not a one-time project, it is a journey of learning and continuous improvement.
Many companies get stuck because they try to plan everything before beginning. The perfect plan will never be as powerful as a small action that actually works. Start simple. Prove value. Then build gradually.
What to do this week
- Write down three things that currently frustrate your team or slow your customers down, repetitive tasks, manual reporting, data entry bottlenecks.
- Run each one through the three lenses: does it align with a business goal, do you have clean enough data, and will it directly help your people?
- Pick the option that scores highest across all three and sketch a rough 60-day proof of concept, what would success look like, and how would you measure it?
- Identify the one person in your business who would benefit most from solving this problem and bring them into the conversation now, not after you have built something.
- Do not wait for the perfect plan. A small, working AI project delivered this quarter is worth more than a comprehensive roadmap that never launches.
Where to from here
Book a free 60-minute AI audit, we'll explore exactly what workflows are worth augmenting with AI.
Live with passion & AI,
Brett
Running an event? Put practical AI on your stage.
Keynotes and workshops that send business owners home with a plan they can use Monday morning. No hype.
Frequently asked questions
What makes a good first AI project for a business?
+
A great first AI project solves a real frustration your team or customers already have, can be measured clearly within 60–90 days, and carries low risk by avoiding sensitive data or complex interconnected systems. The UK construction company that cut weekly reporting from four hours to ten minutes with a simple summarisation tool is a textbook example.
How do I know if my business data is ready for an AI project?
+
Run the project through the Data Lens before you commit: do you already have the data needed to make it work? If your data is messy or incomplete, either start with a lower-data project or clean your data first. Garbage in still means garbage out, even with AI.
How long should it take to see results from a first AI project?
+
Aim for visible results within 60 to 90 days. A tight timeframe forces you to scope the project properly, keeps the team motivated, and creates a clear business case for the next initiative. If a project cannot show meaningful progress inside three months, it is probably too large for a first attempt.
What is the three-lens framework for choosing an AI project?
+
Look through three lenses before committing. The Business Lens asks whether the project aligns with goals for revenue, service quality, or efficiency. The Data Lens asks whether you already have clean, usable data. The People Lens asks whether the project will make life easier for your team or customers, not just automate something for its own sake.
What is the biggest mistake businesses make when starting with AI?
+
Trying to design the perfect AI system before taking a single step. AI is a journey of learning and improvement, not a one-time implementation. Many businesses stall in planning because they want everything figured out in advance. A small working project this quarter beats a comprehensive roadmap that never launches.
Does a first AI project need to be complex or expensive?
+
No. The UK construction company example involved a simple data connection and summarisation tool, nothing exotic or expensive. The best first projects are often the least glamorous: automating a repetitive report, summarising emails, or flagging anomalies in a spreadsheet. Simplicity breeds fast adoption and fast results.
Brett is a four-time founder (Darra Tyres, Gladfish, EzyTrac, Anaboo) and the operator behind AIOS, Anaboo's AI Operating System. He writes from inside the build, installing AI in his own businesses first and reporting back what actually moves the numbers. Based between Singapore, the UK and Australia.
