Blog · Strategy

Nobody at your firm decided to adopt AI

In most firms, nobody sat down and decided to adopt AI. People just started using it. Someone tidied up a proposal with a chatbot, a developer added a coding assistant, and the sales team found a tool that writes follow-up emails.

That's mostly good news. It also means a few questions nobody has answered yet. Which tools see client data? Where do the keys live? What does it all cost, and who is paying?

What it looks like day to day

  • A client contract gets pasted into a free chatbot to tidy up the wording.
  • An API key sits in a script on a laptop whose owner left months ago.
  • The AI bill doubles, and nobody can say which team did it.
  • An AI agent can send email on someone's behalf, and nobody remembers turning that on.

No villains here. Just useful tools with no owner.

Three ways firms respond

A. Ban it

Block the tools and write a strict policy. It's easy to explain and feels safe on paper. In practice people often reach for personal phones and accounts, the useful work gets lost, and AI use doesn't stop. You just stop seeing it.

B. Let everyone pick

Each team chooses its own tools. It's fast and the tools fit the work. But three teams end up paying for the same thing, everyone follows different rules for client data, and one day a client asks what AI touches their data and nobody has an answer.

C. Give AI a front door

One approved way in, clear rules for each team, and a regular look at what has changed. It takes some setup and someone has to own it. In return you can see usage and cost per team, your people keep their speed, and you have an answer when clients ask.

Where to start

  • No written policy yet? Write a one-page policy and list the tools people actually use. (Here's a template.)
  • Adding AI features for customers? Treat whatever the model says as untrusted input.
  • AI agents wired into email, CRM or code? Give each one only the access it needs, and have a person approve sending, deleting and paying.
  • AI bill climbing? Find out spend per team before you cut anything.
  • Clients asking about your AI? Write down what you use, where data goes, and who checks it.

Eight questions for your next leadership meeting

  1. Is there one list of every AI tool we use?
  2. Do we know which of them see client data?
  3. Are our AI keys in a vault, or in code? (Five-step check.)
  4. Can we see AI spend by team?
  5. Does a person approve what AI agents do?
  6. Are we on business terms, so our data isn't used for training?
  7. Who reviews code that AI wrote?
  8. When did anyone last check all this?

A few "not sure" answers are normal. They show you where to start.


Want a second pair of eyes on how your firm uses AI? Book a free 20-minute call.

More from the blog

Let's talk about what you're building.

A free 20-minute call. No sales pitch.

Book a free call