A comms officer uses an AI tool to draft newsletters. It came bundled with her software, it's useful, and nobody told her not to. Down the corridor, a finance manager is doing something similar with a different tool, switched on by default in an upgrade nobody thought to question. Neither mentions it to the other, because neither thinks of it as a decision. It was already there.
If that sounds familiar, it's because almost every organisation has a version of it running quietly in the background. Nobody sat down and chose these tools. They arrived, they were useful, and staying on top of everything else meant nobody circled back to ask whether they were the right ones.
Choosing the free tool, or the one bundled into an existing licence, is often the sensible answer for a stretched organisation. The problem isn't the choice, it's that almost nobody can say whether it was a choice at all. The Charity Digital Skills Report 2026 found 79% of charities now use AI in some form, but only 4% describe that use as embedded in strategy. Only 17% have agreed, as an organisation, which tools are actually approved. Where that agreement is missing, the decision doesn't disappear, it just defaults to whoever's interest is loudest.
The fix isn't a procurement process. It's making an existing choice visible. Convenience can be chosen deliberately, rather than happening by default, once someone can say: this is the tool we use, this is what it's for, these are the boundaries, and this is who to ask about it.
Naming the tool in use turns an invisible risk into a manageable one. A named owner means questions have somewhere to go, rather than nowhere. Checking existing licences first can save money before a new contract is signed. The people the organisation serves stay part of the decision, not an afterthought to it.
A decision that's never made doesn't stay neutral. It just goes to whoever's closest to the licence.
Click 'Read More' to continue reading...