Tommy Findlay calls for a practical approach that helps leaders decide where Ai belongs and when it is ready to use
Good Ai governance should give businesses the confidence to move forward with useful technology, says Tommy Findlay, founder of iS3. Speaking ahead of Web Summit in Lisbon this week, Findlay, who specialises in Ai governance, is calling for businesses to bring governance into their Ai investment decisions from the outset. For a leadership team considering an Ai tool, the useful question is what would make it ready to use. Findlay argues that governance should help answer that question, setting clear conditions for approval and giving someone responsibility for deciding whether those conditions have been met. “Business owners want to understand where Ai can make a worthwhile difference and whether a particular proposal is ready to use,” said Findlay. “Governance should help them reach an answer they can act on. If something needs to change before a project can go ahead, the business should know exactly what that is. Otherwise, leaders are left with a promising demonstration and no clear route to putting it into practice.”
From ambition to approval
Findlay argues that the starting point should be the business task, rather than the technology. Before choosing a tool, leaders should establish what they want to improve and how they will judge whether the change has worked. The next step is to define the responsibility being given to the system. An Ai tool that drafts a response for an employee to check presents a different decision from an agent authorised to send that response directly to a customer. “‘We want to use Ai in customer service’ is too broad to approve meaningfully,” says Findlay. “Does that mean helping an employee find an answer, or allowing a system to resolve a complaint without anyone reviewing it? Once you make the proposal specific, you can have a useful conversation about what is acceptable and what needs testing. You can also decide where a person must remain involved.” That approach gives leaders a basis for approving a defined use, asking for changes or deciding that a proposal does not justify further investment.
Showing what progress looks like
Surprise Shirts, a UK subscription business that sends surprise football shirts to its members, wanted to find a quicker, more streamlined way to select the next shirt for each customer.
Before Ai was introduced, choosing the next shirt for every order meant a person had to read the customer’s notes and order history, before working through the stock from memory. On average this took around 10 minutes per customer, of which they have thousands of orders per month.
The current process was mapped first, and the data was all transferred into one place, then Ai was layered on top with clear boundaries. It now narrows thousands of shirts down to five that fit that customer’s individual rules, a member of the team then selects one and every stock movement is logged. The 10-minute task is now one click, and the owner puts the productivity gain at over 200% giving them a much more streamlined process to get through the busier festive periods and resulting in them no longer needing the temporary hire they had planned to help support during the World Cup season.
For Findlay, the value of governance lies in helping a business reach a decision it understands and can explain.
Sometimes that’s as simple as answering the following questions.
Is the process mapped, and do we know what good looks like today?
Is the data it needs in one place, and can we trust it?
Where can the Ai act on its own, and where does a person approve?
What happens when it gets it wrong, and is every action logged so we can see it?
How will we measure the result against today’s baseline?
Findlay added: “The business should leave that process knowing what will be approved and why. If a proposal can’t answer those yet, the answer isn’t necessarily no, it’s not yet. Most Ai projects fail on process and data, not on the model. So, if you fix the process first and implement decision boundaries, approval points and an audit trail around anything that Ai touches, you can build a governed, human-centred system that makes organisations more productive, accountable and audit-ready.”
iS3 works with business owners and senior leadership teams to implement ISO-aligned Ai governance and automation inside day-to-day workflows, helping leaders make informed decisions about how Ai is introduced and used.
For more information visit https://www.is3.co.uk/.
