Three Lessons From the First Month of Running an AI Focus Group
Early lessons on communication, adoption, and learning from an AI focus group in finance & data
As I mentioned in my last Substack post, the AI Focus Group in our Finance and Data organization officially launched. For the first batch, we have now completed the two discussion-focused sessions (#1 and #2) and the first experimentation session.
It’s still early, but even within the first month a few patterns have already stood out. Interestingly, none of them had much to do with the technology itself. They were about how finance teams learn, communicate, and experiment with new tools.
Below are three lessons that stood out the most so far, particularly from the perspective of working at the intersection of finance, data, and a rapidly evolving field like AI.
1. People are often more willing than it looks on the surface
One of the clearest patterns that emerged early on was how willing people were to engage once the environment was right.
As the facilitator for all of the focus groups, I committed to being present for every group and every session during the pilot. I wanted to ensure everyone had the same experience, but I also wanted to observe how people approached the discussions firsthand.
Before the first two sessions, I shared the discussion questions in advance. Almost everyone came prepared with thoughtful responses. Several participants mentioned they had written notes beforehand so they would have something meaningful to contribute.
Sometimes what they shared was simply that they had not tried AI much yet. That in itself was useful context. It highlights an important reality many organizations are discovering right now: giving people access to new AI tools does not automatically translate into meaningful adoption.
But what stood out to me was the effort people made regardless of their starting point. They signed up for a pilot that I openly described as experimental, found time in already busy schedules, and still prepared ahead of the discussions.
There will always be people who jump on new technologies immediately. Just as there will always be people who prefer to observe, understand the structure, and build confidence before they start experimenting.
Most people fall somewhere in between. The real work of enablement is creating an environment that supports both.
2. Important messages rarely land the first time you say them
Another lesson that surfaced through the discussions had nothing to do with technology and everything to do with communication.
When our company rolled out ChatGPT Enterprise months ago, we shared specific guidelines about what data could be uploaded to enterprise-approved tools. For example, financial data is acceptable if it is uploaded into an approved tool.
On paper, the message seemed clear.
But during the focus group discussions, more than one participant expressed confusion about this guideline. One person even mentioned they had assumed that financial data could not be uploaded at all.
That moment was a useful reminder that enablement is rarely about saying something once and expecting it to stick. Even when communication is clear, important messages often need to be repeated in different formats and over time before they actually land.
In most organizations, people are constantly joining teams, changing roles, or simply missing an announcement in the flow of daily work. Linking to an AI policy once last December does not mean someone who joined in January is aware of it.
For something as important as AI usage guidelines, especially when it touches data sensitivity and risk, the responsibility sits with the organization to reinforce the message repeatedly. If people are confused, it is rarely because they are careless. More often, it is because the message has not been communicated enough times to become shared understanding.
3. Inclusive environments don’t happen by accident
A final lesson that emerged had to do with the environment we create for people to learn together.
After the first session with my very first group, one participant sent me a message that stuck with me. She said she appreciated that I made sure everyone answered the discussion questions because otherwise the quieter participants might not have had a chance to speak.
Her comment meant a lot to me because it confirmed something I had been thinking about while designing the sessions.
I have always been someone who is comfortable speaking up in meetings. Even when I was a junior accountant, I was usually the one asking questions or sharing an opinion. Because of that, I have become increasingly aware of how easily a few voices can dominate a conversation without anyone intending it to happen.
That awareness influenced the structure of the focus group discussions. Instead of an open conversation where the same few people might naturally carry the discussion, we go around and give each person the opportunity to respond to the same set of questions.
On one level, this format helps participants get to know each other across different functions and levels. Just as importantly, it ensures that everyone in the group has space to contribute their perspective.
In a learning environment, especially one centered around experimentation and uncertainty, hearing from everyone matters.
Her message reassured me that the structure was doing exactly what it was supposed to do.
Closing thought
AI technology is evolving quickly. But the first month of running this focus group has reinforced something that feels just as important.
The biggest challenges around AI adoption rarely come from the tools. They come from communication, clarity, and the environments we create for people to learn safely.
Those are human systems problems. And solving them may ultimately matter more than the technology itself.