My AI Lesson From A Yogini

A Note From our Head of Advisory Services — Written by James Scott

Developing AI Proficiency Through Continuous Practice

My wife has practiced yoga for more than 25 years, and I have always found the terminology slightly amusing. After that long, surely at some point you stop *practicing* and can simply say you know how to do yoga?

Of course, that misses the point. Practice does not mean endlessly repeating the basics until you finally become competent. It is the ongoing process of deepening what you know, refining what you can already do, challenging yourself with things that are uncomfortable or unfamiliar, and continuing to learn from the people around you.

Lately, I have been thinking that this is a useful way to approach AI too.

There is certainly plenty to learn. The capabilities are evolving quickly, and keeping up with what different tools can do matters. But AI proficiency is not just about accumulating knowledge. It comes from repeatedly applying that knowledge to real work and getting better at using it.

In yoga, someone may return to the same pose hundreds of times. They are not doing exactly the same thing each time. They notice something different, make a small adjustment, develop greater control, or find a little more range than they had before.

We can do the same with AI.

Maybe you already use AI to prepare for customer meetings. The next step is not necessarily finding a completely new use case. It might simply be getting better at the one you already have: providing richer context, asking better questions, connecting more information, or producing an output that is genuinely more useful.

At the same time, practice also requires stretching yourself. In yoga, eventually you attempt the pose you cannot yet do. With AI, that may mean moving beyond the safe and familiar uses and experimenting with something more ambitious: analyzing a portfolio of customers, creating a workflow, interrogating a dataset, or building an agent to handle part of a recurring process.

Some of those attempts will work. Some will be clumsy. That is rather the point.

And while practice can be solitary, there is also real value in practicing alongside others. Watching someone take a different approach, borrowing a technique, asking a question, or simply seeing what is possible can accelerate your own growth. The same is true with AI. Some of the most useful learning comes from seeing how colleagues and peers are applying it to real work.

With enough repetition, something else starts to happen: you find a flow. You stop having to consciously ask yourself whether AI could help and start spotting opportunities more instinctively. It becomes less of a separate tool you visit and more part of how you naturally approach the work.

For Customer Success professionals, that may be the more useful definition of AI proficiency. Not knowing every feature or mastering the perfect prompt, but developing a regular practice: deepening your understanding, refining what already works, stretching into things you cannot yet do, learning from others, and using AI often enough that it begins to flow naturally into the way you work.

So, with that in mind, I'm heading back to my 'AI mat'. I still have plenty to practice.

Roles on the Radar

• Director, CX Business Systems at Braze
• Director of Customer Success (in New York, Amsterdam, & Mexico City) at DataSnipper
• Customer Success Manager at QBench
• Customer Success Manager - Mid - Market / Enterprise at PandaDoc
• Head of Customer Success at Phocas Software

Notable News

ICYMI: What should a Customer Success (CS) leader actually be earning in 2026?

As CS leaders take on greater responsibility for retention, expansion, NRR, AI transformation, and overall business growth, compensation needs to evolve with the role. CS Recruiting Firm, Melo Associates, broke down 2026 CS leadership compensation benchmarks by title and company stage. See how your role, your team, or your next hire compares.