If you lead a learning team, you’ve probably had the feeling that everyone else has AI figured out. Between conference sessions and the headlines, it’s easy to come away thinking your team has some catching up to do. That pressure is starting to show in how organizations spend, too. Keeping pace with technology, and with AI in particular, has become the most common reason organizations buy learning technology, moving ahead of replacing tools that no longer deliver (Source: Training Industry).
The teams we work with paint a different picture. For most of them, the conversation has only recently moved on from whether to use AI to how to make it work in their own setting. Fewer are treating it as a side project, and more are working it into the tools and workflows they already rely on. So if your team is somewhere in the middle of that shift, you have plenty of company.
We spent time on this in our recent webinar with Training Industry, where learning leaders from Cisco and Anaplan talked through a year of building AI into how their teams work.
Why it feels like everyone else is ahead
Part of the problem is how AI stories get told. They tend to jump from before to after, which makes adoption sound like someone flipped the main breaker, with nothing running on Thursday and everything running by Friday morning.
In practice, it’s closer to walking through a house and switching on the lights one room at a time. Teams move at different speeds, and so do the people in them. Some dive straight in, while others want time and some hands-on practice before they trust what the tools produce.
Mazie Baskin, VP of Anaplan Academy, took that gradual route with her team. Getting a new tool through company review takes time, so they leaned on the ones already approved and saved their requests for the few that mattered most to them, including AI video and the AI features in their authoring tool.
Picking the right problem matters more than picking the tool
Choosing a tool is often the quickest AI decision a learning team makes. Working out what you want one to change or impact takes more thought, and that decision shapes everything that comes after it.
For Anaplan, for example, the starting point was a business problem. Over roughly a year, the company went from two main products to around 40, and partners needed to implement all of them well. Mazie’s team used Gemini to turn practices that had worked in the field into readiness kits partners could put straight to use. Lucidchart’s AI summaries helped them capture what stakeholders needed during requirements sessions.
Many of our clients come to us with a similar starting point. They ask us to look across their learning workflows and show them where AI could speed things up, bring costs down or make the output more consistent.
Where to switch on the first light
Start with one bottleneck. For a lot of learning teams, that’s documentation, graphics, approvals or keeping existing content up to date. If you can tackle it with tools you already have access to, you can start testing ideas right away rather than waiting on a new approval. Make sure your subject matter experts still review what comes out, though. The people who know your products and learners still know things no model does, and their judgment is what protects quality. Once that first area is working, you have a model to repeat in the next one.
In short:
- Pick one bottleneck to start with.
- Use the tools you already have.
- Keep your experts reviewing the output.
- Build on what works.
Closer to the business
In our conversations with learning leaders lately, we’re hearing more enthusiasm and curiosity about AI. Many see it as a chance for L&D to work closer to the business and help solve its problems faster. For most teams, getting there starts with choosing one problem worth solving and building from it.
Watch the full webinar recording to hear how the learning teams at Cisco and Anaplan chose where to start with AI and what they learned along the way.
If you’d like help deciding where your team should start, talk to our team.





























