Simple Assistant vs. Expert Mentor

If you let AI do all the thinking, you end up with systems your team can't support. If you never let it think at all, you miss most of the value. We've spent the last few months working out how to get both - here's the framework we landed on.

Back in late November we started playing around with Claude Code to see how we'd use it in our environment and what rolling it out with our team would look like. Probably a bit late to the party - but our business is about solving business problems using tech, and with that hat on we don't just care about pure tech, we care about how that tech lands in business environments in ways that are sustainable, supportable, value-creating and risk-managed.

After our December break, we IMMEDIATELY started driving adoption of Claude across the team. We engage with Claude Code and CoWork in one of two roles:

• Simple Assistant - executes instructions and delivers on tasks assigned by the team member, without too much "thinking" and with lightning fast speed.
• Expert Mentor - advises, informs, enlarges and trains the team member by asking and answering questions, making suggestions, introducing new possibilities, and "thinking" on their behalf.

This framework has been helpful for us. It keeps the team member as the responsible party. They still make the decisions, set the direction, evaluate options and call the plays. The assistant does multiple things at once, delivers detailed feedback, and frees them up to focus on what matters. Gold.

They ALSO have access to the wealth of knowledge the AI provides through the role of expert mentor. They can ask questions, consider ideas, invite suggestions, and be introduced to possibilities they'd never have found on their own. They grow in knowledge and understanding much faster than before. Also gold.

Here's the key: the mentor's job is not to execute tasks - it's to grow and develop the team member. The assistant does the execution. If the expert does the execution and bypasses upskilling our team, we have a problem; we end up with systems we don't understand and we lose the ability to support and maintain them. If we don't engage the expert at all, we don't grow. We need both - but they need to stay in their lane.

Later this week we'll share some of the numbers showing the productivity improvements we've seen over the last 3 months. Follow the page to see the results.

And if you're also trying to work out how to roll AI into your environment in a risk-managed, value-creating way - get in touch. Turning tech into business value is where we live.