This year, one of our biggest goals as a company is what we call ‘AI + Me = Greater Productivity.’ It’s the humans behind the work, with AI as a booster. Our people are the drivers of innovation, and AI just helps the process move faster, leaving more room for our people to ideate and execute creative ideas.
So when a few of our teammates went to an AI-focused industry conference recently, we wanted to know what they brought back: which ideas are worth acting on, how perspectives may have shifted, and how it connects to the work we’re already doing here.
The Noise vs. The Signal
If there was one thing everyone agreed on, it was that “agentic AI” was inescapable – in session titles, on slides, in every hallway conversation. Underneath the buzzword, though, a few real, useful ideas kept surfacing.
One is a shift in how to think about automation itself. As one teammate explained, the instinct to automate the easiest parts of a workflow is often backward: it’s frequently the steps that most need human judgment that get automated first, because they’re automatable. The better question isn’t “can this be automated?” It’s “where does a person’s judgment actually need to stay in the loop?”
That connects to another idea that came up from a different angle: agents aren’t just models, they’re a model plus a harness. That harness is the layer that decides what an agent can touch on its own, and how you’d know if something went wrong. Build those checks into the workflow itself, rather than counting on someone to remember to run them. However, that’s not a groundbreaking idea, as we already knew we needed operational frameworks in place, long before AI came into the picture.
Practical Over Impressive
One teammate spent day one in high-level sessions, then pivoted on day two to small-business-focused, hands-on content and found more traction there. Her takeaway wasn’t that the conference was weak, but that there’s a real gap between talking about AI and integrating it. That gap is exactly where we’ve already been doing the work as a company, so what this trip confirmed is that our approach to invest, apply, and iterate is the right one. It’s also a reminder to keep pushing rather than assume we’ve arrived.
Another idea worth mentioning: create a genuine low-stakes sandbox where people can try things without fear of breaking something real. Two completely different sessions landed on the same point from different directions –one was about agent safety, one about organizational change – and the plain-language version was “permission looks like play.” If you want people to experiment, you have to make it safe to learn, fail, and iterate. first.
Where We Go From Here
This is about turning what our people heard into a direction we can use: prioritize by value, not by novelty. Instead of chasing whatever’s exciting that week, look at where AI genuinely overlaps with high-value work, and focus there first.
Build guardrails into the process, not just the policy – ownership, budget, and basic checks belong inside the workflow. And protect space to experiment: growth doesn’t happen in a high-stakes environment. It happens where people feel safe trying. Beyond the sessions, several people mentioned that getting real time to spend with each other outside the day-to-day was valuable and appreciated. That matters too. Growth isn’t only technical.
That’s the real point of sending people off-site to learn in the first place: we invest in development because we take growth seriously, and because the best ideas in this industry usually come back through our own people, not around them. ‘AI + Me’ is a constantly evolving idea of what our workflows could look like, and this conference was one way to spark creativity among our people about what is possible and how to apply AI in ways we haven’t tried yet.