Inspired by the webinar: AI Changed the Buying Journey. Most Websites Aren't Prepared.
I sat on an RTM Nexus webinar panel with Tim Zawislack moderating, Breanna Fowler from Dell Technologies, and Girish Joshi from LG Electronics. The official topic was how AI is changing the buying journey. The unofficial topic, the one that kept surfacing whether we steered toward it or not, was how we’re actually talking about AI inside enterprise commerce, and whether the talk is pointed at the right outcome.
I keep noticing the same thing: most enterprise AI conversations are happening on the wrong side of the P&L.
When the CFO is the one asking the AI questions, the answers come back about cost. When the COO is asking, the answers come back about efficiency. When the CTO is asking, the answers come back about consolidation. All defensible. All interesting. None of them, by themselves, are why this technology actually matters.
The brands I’m watching pull ahead are doing something different. They’re running AI on the other side of the page. The revenue side.
The dollars going into enterprise AI right now are enormous, and most of the discussion still revolves around cost reduction. Headcount. Tech consolidation. SaaS bloat. All defensible. There’s a real argument that a lot of software got bought over the last decade and a half, and some rationalization is overdue.
But cost reduction is a one-time outcome. You consolidate the stack once. You renegotiate the contracts once. You absorb the reorg once. After that, the math goes flat.
Revenue is a compounding outcome. You build a faster experimentation loop, and every cycle gets cheaper relative to what it produces. You add an AI layer that catches issues before they hit the customer, and you don’t just stop the leak. You keep the segment that was about to churn. You find revenue that didn’t exist before because nobody was looking in that direction.
If you’re using AI only as a hatchet to your cost base, you’re playing defense in a game that’s only won on offense.
I’ll admit it: the public narrative hasn’t helped. The dominant AI story in the press for the last 18 months has been about replacement. AI is going to take jobs. AI is going to flatten orgs. AI is going to shrink your team.
Some of that is real. A lot of it is overstated. Girish made the point on the panel that resonated with me: jobs aren’t being lost to AI in commerce so much as roles are getting rerouted. AI is absorbing the work humans were never going to get to in the first place. The broken-link audits. The malformed checkout flows nobody’s funnel ever flagged. The QA pass that used to happen at four in the morning, out of one bloodshot eye. That work was always there. It was always undone. AI just has the bandwidth a human team never had.
The brands that lead with “we’re going to use AI to reduce headcount” are sending a signal to their best people that AI is a threat. Then their best people leave. Then the AI initiative doesn’t ship because the institutional knowledge walked out the door. Self-inflicted wound.
The brands that lead with “we’re going to use AI to give you back the parts of the job you actually went to school for” are building something different. Their AI initiatives have allies, not survivors.
Inside the brands I watch up close, it plays out the same way.
AI handles the mundane. The reporting. The data reconciliation. The tagging. The audit. The cleanup. The brain-numbing work that used to chew up a senior merchandiser’s Tuesday and produced nothing the business could feel.
Humans handle what humans are actually good at: strategy, prioritization, judgment. They decide which insight matters and which is noise. They pick the experiment worth running. They write the brief for the campaign. They sit across from a customer and figure out what the data missed.
The pattern is consistent: the orgs that get AI right don’t end up with fewer people. They end up with people doing more interesting work. The job descriptions shift. The mundane goes to the machine. The judgment stays with the human.
That’s not a layoff. That’s a promotion.
The most interesting AI use cases I’m seeing in enterprise commerce aren’t faster versions of work humans were already doing. They’re work humans were never going to do.
Run a hypothesis test on every PDP in the catalog, not just the top 50. Spin up a personalized landing variant for every paid keyword you’re bidding on, not just the top 20. Audit product copy on 8,000 SKUs against the campaign that drove the click and flag every contradiction in an afternoon, not over six months of QA. Watch how ChatGPT, Perplexity, and Gemini describe your brand and adjust your content until they describe it the way you want.
None of that was on the roadmap last year. Not because the strategy people lacked imagination. Because no human team in the world had the hours to actually do it.
That’s the revenue side of AI. Not “do the same work cheaper.” Do the work that was never economically possible. That’s where the lift lives. That’s where the multiplier compounds.
One more thing from the panel I want to underline, because almost nobody outside the trenches is saying it: AI is only as good as the context you give it.
A brilliant consultant who only gets to read your annual report is going to give you a generic strategy. Put that same consultant inside your offices for a month, with access to every doc, every meeting note, every customer transcript, and every dashboard, and you get something you can actually use.
The same is true for AI.
Breanna made this point sharply on the panel: AI is an accelerant. It doesn’t surface anything new. It shines a very bright light on what was already broken, and the root is almost always data and content. One pair of headphones in two colors, pulling from two wildly different copy banks. The AI didn’t create that problem. It just made it impossible to ignore.
The brands getting weak AI results aren’t necessarily using weak AI. They’re using AI without context. Product data sits in one system. Campaign data sits in another. Site behavior sits in a third. The CRM lives somewhere else entirely. The AI is asked to reason across all of it with access to none of it. That’s not an AI failure. That’s an architecture failure dressed up as an AI problem.
I watch this die the same way every week. A team spots conversion slipping on a PDP segment. They know exactly what they want to test. Then they log a ticket, the sprint is full, and the fix waits in a dev backlog behind four to six handoffs across two or three systems. That’s not a failure of ambition or strategy. It’s a structural execution problem. And it’s the first one AI can actually collapse, if you let it.
The brands getting compounding AI results have done the unglamorous work of unifying that context. Not by replatforming, but by putting analytics, content, deployment, and learning into a single workspace where the AI can see what’s happening end-to-end. That’s the Insight Gap and the Activation Gap closing at the same time. It’s the reason we built Fastr Workspace the way we did.
Without context, AI is a generalist. With context, it’s a specialist who already knows your business.
One last thing, and it should govern everything else.
We spent a chunk of the panel on org structures, centers of excellence, ownership models, who reports to whom. Necessary conversations. But here’s the line I keep coming back to: your customer doesn’t know which department owns your checkout. They don’t know if your PDP team sits under merchandising or product. They don’t know whether your AI initiative reports to the CIO or the CMO. And if you told them, they wouldn’t care.
What they care about is whether the page loads. Whether the variant matches the ad that drove the click. Whether the price is right. Whether checkout works on their phone. Whether the experience feels like a brand that knows what it’s doing.
Every AI conversation that doesn’t end at that customer experience is a conversation about internal mechanics. Necessary, but not sufficient. The leaders who win this cycle are the ones who can hold both: fix the internal mechanics, and keep the customer experience as the only outcome that ultimately matters.
The reframe I’d offer every commerce leader staring at an AI budget right now:
You can use AI to shrink the org chart. Or you can use AI to grow the business.
Both are valid. Only one compounds.
The brands stuck on the cost side are going to spend the next 18 months optimizing inputs. The brands working the revenue side are going to spend the same 18 months building experimentation velocity, customer experience advantage, and content authority across AI search engines that none of their competitors will catch in 2027.
Pick the side of the P&L you want to play. But pick.