This week was one of those weeks where you walk away with more questions than answers — but in the best possible way.
I attended the ATLCS Summit here in Atlanta and an invite-only AI deployment session hosted by Zapier featuring AI leaders from Netflix, Indeed, and ServiceTitan. Two very different rooms. Two very different conversations. And one through-line that kept showing up in both.
AI adoption is not a technology problem. It’s a people problem. And most companies are still treating it like the former.
Here’s what I took away.
One of the sessions I was most looking forward to at the Summit was Katie Christian’s talk on how AI helps surface customer signals earlier by connecting siloed data sources.. product usage, support interactions, customer engagement.
The premise is something I’ve been thinking about a lot with clients: post-sale teams are sitting on more signal than they realize. The data exists. It’s just scattered across systems that don’t talk to each other, which means the signal arrives late.. after the customer has already gone quiet, already disengaged, already started evaluating alternatives.
AI changes the timing. When you can connect those sources and identify meaningful patterns sooner, you can act at the right moment instead of the reactive one. That’s the difference between preventing churn and managing it.
But here’s the part that landed hardest.. and the most practical thing I heard all week.
Someone in the room asked the question that always comes up when you talk about AI and data: “What if our data isn’t good enough?”
Katie’s answer was simple and I’m going to be repeating it for a long time:
“No one will ever say they have good data. Start small with the data you do trust and go from there.”
That’s it. That’s the unlock for every team that’s been waiting for perfect data before they start. The data will never be perfect. The question is which data you trust enough to act on today.. and whether you’re willing to start there.
I had heard a lot about Kristi Faltorusso work before this week but had never seen her present in person. She did not disappoint.
Her session.. going from overwhelmed to operational, the human side of AI.. was the kind of talk that makes a room go quiet. Not because it was heavy, but because it was honest.
Throughout the session, Kristi said what so many don’t say outloud… we are all starting at the same place again.
Think about what that actually means. Every expert in your organization, every high performer, every person who has spent 15 years becoming excellent at their job.. they are navigating this shift from the same starting line as everyone else. The playing field has genuinely leveled in a way it almost never does.
And her point was that the people who will still be standing when this settles are the ones who invest the time right now to learn and uplevel themselves with AI. Not the ones who wait until it’s mandatory. Not the ones who watch from a distance. The ones who get in and build the muscle while there’s still room to learn slowly.
That hit differently in a room full of CS and post-sales professionals. Because this is a community that prides itself on expertise. On knowing the customer, the product, the industry. And suddenly that expertise has a new requirement attached to it.
The people who will carry their expertise forward are the ones who figure out how to pair it with AI. The ones who don’t.. risk having their expertise replaced by someone with less experience and better tools.
The same day, I was in an invite-only session hosted by Zapier with AI leaders from Netflix, Indeed, and ServiceTitan. The framing was AI deployment reality.. what’s actually working at companies that have moved beyond pilots.
A few things stood out.
Resistance is almost never ideological. Tracy St, Zapier’s Global Head of Talent, made the point that what gets labeled as “resistance to AI” is almost always something else: no time to experiment, no relevant examples to follow, no management support, no clear ask. Ask a supposedly resistant employee where AI could help their specific work and they almost always have an answer. They’re overwhelmed, not opposed.
Your team is probably further along than they think. Netflix surveyed employees on their AI fluency self-assessment and compared it to actual usage data. About 40% of employees were in the top half of actual usage but described themselves as early beginners. Almost nobody overestimated. A lot of “I need more training” requests are people not realizing how much they’ve already learned. That’s a recognition problem, not a skills gap.
Governance is the accelerant, not the brake. The organizations moving fastest with AI aren’t the ones who gave everyone access and hoped for the best. They’re the ones who built the governance infrastructure early enough that people could experiment freely without creating risk. When the guardrails are solid, you can move fast with confidence. When they’re not, you end up with shadow IT at scale.. except now it has agentic access to your CRM.
Two events. Multiple speakers. A lot of ideas.
But the thing I keep coming back to is this: every conversation this week was ultimately about the same challenge. Not the technology. The humans navigating it.
Whether it’s a CSM who doesn’t know where to start, a CS leader trying to get team-wide adoption, or an enterprise company trying to govern AI at scale.. the constraint is almost always people, culture, and change management. Not the tools.
The tools are ready. The question is whether we are.
I’m going deep on the AI leaders lab session next week.
The first covers the culture and change management side of the Zapier session.. the resistance myth, the Netflix finding, and the function-first principle from Indeed that I think every CS leader should steal.
The second covers governance and tools.. why governance done right is what makes fast experimentation possible, the tiered model Indeed has been developing, and the three things that determine whether an AI tool actually gets adopted or quietly abandoned.
Both are free. Both go deeper than what I covered here.
If someone forwarded this to you and you want both articles in your inbox next week.. subscribe below.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.