Every recruiting agency claims to use AI now. It is on every website, in every pitch deck, at every conference. But there is a massive difference between a founder experimenting with ChatGPT at midnight and a whole team actually integrating AI into daily workflows. That gap is where most teams get stuck.
I know because we got stuck there too.
At EvoTalents, we have been a remote team for over ten years. We recruit IT specialists across the UK, US, Europe. When AI tools started popping up everywhere, I was all in. I tested everything: prompts, automation, assistants, you name it.
And naturally, I wanted my team of 10 to be right there with me. I imagined that by next week we’d all be running on AI, fully automated.
Reality hit differently. Even in a small team, adoption was much slower than I expected. And when I finally stepped back to figure out why, I realized the problem wasn’t the tools. And it wasn’t the people either.
It was my approach from the very beginning.
My first instinct was to show the team cool tools. “Look at this! It writes Boolean strings! It drafts outreach emails!” I was met with polite nods and zero behavior change.
So I tried something different. Instead of pushing tools, I sat down with my team and shared my vision for their specific role. Not a generic “AI is the future” speech. A concrete picture of what their job could look like when AI becomes a real partner in it.
For my assistant, who now essentially manages our AI assistants, I described a role that did not exist one year ago: someone who maintains our AI projects, feeds them updated context, quality-checks their outputs, and keeps the whole system running smoothly.
For our marketing person, I outlined how AI could draft first versions of content, research trending topics, and handle translations, freeing them to focus on strategy, voice, and the creative decisions that actually differentiate a brand.
For our sourcers, I described a model where AI handles the mechanical parts of sourcing: building initial candidate lists, parsing job descriptions, generating search strings, while the human focuses on what machines still cannot do: reading between the lines of a candidate’s career trajectory, sensing cultural fit, building trust in the first conversation.
This was not about replacing anyone. It was about showing each person where their role was heading and making that direction feel exciting rather than threatening. People do not resist change when they can see themselves in the future you are describing.
Vision alone is not enough. You need to create the space for people to learn, experiment, and talk about what they are discovering.
We did three things that made a real difference.
First, we started holding demo sessions every two weeks. Sometimes I show something new I have been testing. Sometimes a team member demonstrates a workflow they figured out. We talk about what worked, what flopped, and what surprised us. These sessions are not polished presentations. They are honest, messy conversations where someone might say, “I spent three hours trying to make this tool do X and it was terrible” — and that is just as valuable as a success story.
Second, we created a dedicated Slack channel “AIBuzz” for AI tools and experiments. It is not a knowledge base or a formal wiki. It is a living conversation. Someone drops in a prompt that worked well. Someone else asks if anyone has tried a particular tool for a specific task. Someone shares a screenshot of a hilariously wrong AI output. The informality matters. It lowers the barrier to participation.
Third, we allocate a budget for external AI training that we attend as a team. Not one person goes to a webinar and reports back. We all go together, and then we discuss what we learned in the context of our own work. This shared experience creates a common vocabulary and a sense that we are all in this together.
I am a big believer in being transparent about exactly what tools you use and how. Vague statements like “we leverage AI” help no one.
Here is what our stack actually looks like right now. We migrated our entire team to Claude Teams, which gave everyone access to the same AI capabilities with shared context about our company. We use Projects within Claude to maintain persistent knowledge bases for different clients and workflows, so the AI actually understands our business context, not just generic recruiting knowledge. We use Claude Cowork for collaborative tasks where the AI assists with complex deliverables like client reports and market analyses. And we use Claude Code for internal automations and workflow scripting.
Also our sourcing and recruiting stack runs on AI at every stage:
But the stack is not the point. The point is that every tool earned its place by solving a real problem someone on the team was experiencing. We did not adopt Claude Teams because it sounded impressive. We adopted it because our sourcers were wasting time re-explaining company context in every new AI conversation. The tool solved a friction point that the team identified.
This is important: let adoption be pulled by need. Your team will tell you what they need if you create the space for them to talk about it.
Let me tell you what we got wrong, because I think this matters more than what we got right.
The first mistake was not addressing the team’s fears head-on. When you start implementing AI, people get scared. They do not say it in team meetings, but they think it: “Is this thing going to replace me?” And if you do not address that fear directly, it sits there quietly and sabotages every adoption effort you make. People will not invest time learning a tool they believe is being built to eliminate their job.
What I learned is that you have to be upfront about it. Not with empty reassurances like “AI will never replace you” — people see through that. You need to honestly share your vision for how the team evolves. In our case, that means moving toward a model where AI and humans work side by side, each doing what they do best. But you can only say this if it is genuinely true. If your actual plan is to reduce headcount, your team will figure that out fast and you will lose their trust entirely.
The second mistake was the biggest one, and the hardest to admit: at the start, I was not personally involved in the process. I would buy a training course and tell the team, “Implement this.” Or I would say, “Figure out how to use AI in sourcing” — and step back. The team would test something, it would not work perfectly on the first try, and they would quietly go back to their usual workflow.
Everything changed when I started being in the room. When I began sharing what I was personally using — even when my ideas seemed strange or half-baked at first — the dynamic shifted. When I showed up to demo sessions with my own experiments, when I started asking in our daily check-ins what people were trying, when I made it clear that I was figuring this out alongside them and not delegating it from above, that is when the team actually started engaging.
If the founder or team leader is not visibly, actively using AI themselves, the team will not either.
The third mistake was trying to implement too many tools at once. Three new platforms in one month. The team was overwhelmed. People reverted to their old workflows because the cognitive load of switching was too high. We learned to introduce one tool at a time and give people at least two to three weeks to integrate it before adding the next thing.
And we learned that some experiments just fail. A tool that works beautifully in theory turns out to be clunky in practice. Someone spends hours configuring something that produces mediocre results. That is fine. The point is not that every tool works. The point is that you have a culture where people feel safe saying, “This is not working” — and where that honesty is treated as progress, not failure.
What I have realized through all of this is that AI adoption in a small team looks a lot like community building. You are not just changing processes. You are changing how people think about their work, their skills, and their future.
The demo sessions, the Slack channel, the shared training, the transparent conversations about what is working and what is not — these are not just change management tactics. They are the foundations of a team culture that can absorb continuous change. Because AI is not going to stop evolving. The tools we use today will be outdated in eighteen months. What will not be outdated is a team that knows how to learn together.
We are not done. We are nowhere close to done. But we have built something I am genuinely proud of: a team that sees AI not as a threat or a silver bullet, but as a craft that we are learning together. And that, I think, is where every recruiting team needs to start.
