Insights from Wpromote’s Webinar with OpenAI and Kargo
“We are at an accountability moment in marketing,” Christine Schrader (Wpromote) said partway through our webinar, “Conversational AI: What Early Advertisers Are Learning,” right before she posed the question on every marketer’s mind: “How do we know if it’s working?”
Six months ago, that was a hard question to answer. “We’re so used to it, the way measurement works in paid search,” Schrader continued, pointing to the practice of a consumer clicking and a marketer tracking what happens next.
But conversational AI doesn’t work like search, which is something else the panel explained, and something we covered in our first post.
So what does measurement actually look like in conversational AI right now, and what should marketers expect from it?
WHAT’S ALREADY LIVE
“This is the top conversation happening inside OpenAI right now,” Whitney Muldowney, on their global ad solutions team, said.
Back in February, phase one was about getting delivery right and keeping things safe for advertisers and users. Phase two is measurement, and according to her, it’s moving fast.
The week of the webinar, OpenAI turned on configurable attribution windows. Click-through and view-through are both trackable now. They are also in talks with third-party MTA providers to track ChatGPT’s role across the full customer journey, with incrementality and marketing mix modeling further down the roadmap.
Multi-touch attribution vendors, tracking platforms, and answer-engine-optimization tools are all racing toward the same gap right now: visibility into what’s happening upstream of the click.
And it’s not just AI, either. According to the IAB, 75% percent of marketers say their overall measurement isn’t delivering the speed, accuracy, or trust they need.
Which is exactly why brands can’t afford to sit around waiting for the tools to catch up.
DON’T SACRIFICE OPPORTUNITY FOR PERFECT
Every channel goes through this. “Measurement takes time to develop, and it’s really important to keep a true testing mindset,” Casey Geiger (Wpromote) said.
“Just because it’s not all the way to the most perfect, ideal thing, do not sacrifice opportunity for that,” Schrader added. “You have to understand what success looks like for now, and understand where you’re trying to get to.”
Geiger called it the 800-pound gorilla in the room. Want the edge that comes from getting in early, from getting good at this before competitors do? You have to start. Test. Experiment. There’s no version of this where waiting for certainty gets you there faster.
KNOW YOUR OWN APPETITE FOR PROOF
So if AI measurement is still maturing, what should a brand hold itself to in the meantime? Geiger’s answer cuts through a lot of the noise: know your own appetite.
“Know what your specific brand’s appetite for measurement is,” she said. “Every brand is going to be a little bit different in what they need to determine success and what they need to measure. Get really clear on what your brand’s appetite for measurement is, and then what specific tagging or tracking that you’ll need in place to achieve that measurement.”
Think about how differently that plays out across two brands. A direct-response ecommerce account might need same-day ROAS before it’ll call a test a success. A brand with a six-month B2B sales cycle might care more about whether a lead even mentions ChatGPT in that first conversation with sales. Same platform, same measurement maturity, but a completely different bar for proof.
Nobody hands you that bar. You set it, then build backward from there.
BUILD YOUR OWN SOURCE OF TRUTH
Platform-side measurement is only half of it. Brands need to take responsibility into their own hands too.
“If you’re not maintaining your source of truth with your agency, with your own technology, that’s absolutely critical,” Schrader said. “That is a necessary component to modern marketing, especially accountable marketing.”
Geiger described Wpromote’s practice of mapping conversational AI campaigns back to one centralized intelligence platform, in our case, PolarisIQ, so advertisers get one source of truth across everything they run.
WHAT TO DO THIS WEEK
Asked for the one tactical move brands should make right now, Muldowney was very clear: Get conversion tracking in place. “Getting it in place now means you’re getting learnings as conversion bidding comes out, and as we move into 2027. I’d say act with urgency here, just on a technical standpoint.”
That’s the technical half of the same idea. Know what you need to prove, then make sure the plumbing exists to prove it.
THE TAKEAWAY
Geiger’s big-picture advice is also worth keeping close. Get clear on what your brand actually needs to call something a success. Build your tracking around that answer.
The platform side will keep maturing. On its own timeline. Attribution windows, MTA partnerships, incrementality models. All of it.
Your internal standard for proof doesn’t have to wait on any of it.
If you want to know more about advertising in Conversational AI, you can watch the complete webinar on demand here.
