Andrea Lechner-Becker, Chief Strategy Officer at GNW Consulting, explains why AI search is forcing B2B marketers to give up the idea that they control the buyer journey, and what to measure instead.
If an LLM can’t find your pricing on your website, it doesn’t give up. It goes to Reddit. Then G2. And if it still comes up empty, it makes something up. That one idea from this episode of Deconstructing Data has stuck with me all week.
I wasn’t on this one, but our CTO DW (David Wellborn) sat in with Jessie Lisak, and they had a great conversation with Andrea Lechner-Becker, Chief Strategy Officer at GNW Consulting. Andrea calls herself a data nerd. She has been working in martech since 2011, and most of her consulting work now is generative engine optimization (GEO), also called answer engine optimization (AEO). That means getting brands mentioned, and ideally recommended, inside the LLMs people now use to research what to buy.
Here’s what stood out.
How is AI search changing the B2B buyer journey?
Andrea’s main argument is that marketers never really controlled the buyer journey. They just thought they did.
She pointed to research from HockeyStack showing that B2B purchases involve more than 700 touch points between a buyer first hearing about you and actually paying you. Most of those happen in the dark funnel: Slack communities, happy hour conversations, a vendor name that keeps coming up at conferences. Marketers tried to make up for that by pulling people onto their websites and gating everything behind a form.
“I would contend that we’ve never really been in control of the journey.”
— Andrea Lechner-Becker, Chief Strategy Officer, GNW Consulting
LLMs get around all of that. When a buyer asks ChatGPT what your product costs, the model checks your site first because it treats you as the authority. If the answer isn’t there, it goes to forums, which leaves you exposed to whatever people (and maybe your competitors) are saying about you. And if it still can’t find anything, it guesses.
“The third thing that it will do is hallucinate because it will give an answer, and it will give a best guess whether you want it to or not.”
— Andrea Lechner-Becker, Chief Strategy Officer, GNW Consulting
Her conclusion is simple. Stop hiding your pricing and your data sheets. Buyers never wanted to talk to a sales rep just to get a number, and the LLMs are now giving them that information whether you like it or not.
“I would give LLMs all the information. I want to be in the answer.”
— Andrea Lechner-Becker, Chief Strategy Officer, GNW Consulting
DW mentioned something we’ve seen at BDEX too: lots of sites behind Cloudflare block AI crawlers by default. Andrea’s tip was to go check that setting, because you might be invisible to the LLMs without knowing it.
Where should marketers start with GEO and AEO?
Andrea gave a practical two-step starting point.
First, fix your Google Analytics 4 setup. GA4 already tags traffic coming from AI assistants with a source like “ai-assistant” plus the specific engine, such as Perplexity or ChatGPT. Make sure your goals are configured, then compare the conversion rate of AI-referred traffic against everything else.
Second, find out how the LLMs actually see you. She recommends the free trial at AirOps (she made a point of saying she isn’t paid by them) because it lets you write your own prompts and runs them across four engines. Running prompts by hand isn’t worth it, in her view. It takes about ten minutes per prompt per engine.
She is skeptical of the “paste your URL into our AEO scanner” tools that seem to pop up every week. Her reason is that everything depends on the prompt set.
“The prompt set that you create is really where all of the gold is. And so all of these metrics are quite easy to manipulate based on your prompt set that you use to begin with.”
— Andrea Lechner-Becker, Chief Strategy Officer, GNW Consulting
That’s a good point. If someone else writes the prompts, you’re measuring their idea of your market, not yours.
Can you track which LLM conversations lead to a sale?
DW asked the obvious question. If someone researches you in a desktop LLM app and shows up on your site three days later, can you connect the two?
Andrea’s answer was short: “The fast answer is you don’t.” What you can get is aggregate visibility, and she went through the options.
Bing Webmaster Tools has an AI performance dashboard covering Copilot, Bing, and partner properties. It shows “grounding queries” (Bing’s way of normalizing all the different ways people phrase the same question), how many times you were cited, and your citation share. She considers it the most accurate source because it comes straight from the platform. The catch is that it only reflects the Bing and Microsoft audience, which leans toward large enterprises.
Semrush uses clickstream data collected from browser extensions, software installs, and ad networks. Profound says it has millions of users who agreed to be monitored. Andrea suggests taking both with a big grain of salt.
To fill the gap, she recommends self-reported attribution. Ask “how did you hear about us?” on your forms and on the first sales call. If the answer is an LLM, ask what they were searching for.
Why identity resolution and data enrichment are getting harder
Andrea was honest that AI search makes identity resolution harder. The dark funnel keeps getting wider. A couple of years ago ChatGPT was the only engine anyone tracked. Now it’s Claude, Google’s AI Overviews, and more coming.
Her advice is to stop chasing perfect attribution and focus on what you can control, which is your message. When you look at LLM responses about your brand, she says, the biggest problem usually stands out quickly.
“Most organizations have a massive messaging and positioning problem.”
— Andrea Lechner-Becker, Chief Strategy Officer, GNW Consulting
On enrichment, she described what she hopes comes next. B2B email segmentation still mostly runs on job title, industry, and company size. She’d rather segment on things like sense of humor or communication style, built by enriching profiles from social data and having an LLM summarize who a person really is. She doesn’t think the tools are quite there yet.
What makes data “quality” data?
This part was close to home for us. DW explained that BDEX has always chosen quality over volume, and asked Andrea how she defines quality data.
“Accuracy equals quality, almost equal.”
— Andrea Lechner-Becker, Chief Strategy Officer, GNW Consulting
DW summed it up as “accurate and actionable,” and she agreed. She also stressed speed. Data has a half-life, especially in marketing, so if you pay to enrich a list you need to use it quickly. Her approach is to get things into the market, see how the market reacts, do more of what works, and drop what doesn’t.
For her own stack, Andrea mentioned Marketo, Adobe, Salesforce, Semrush, Profound, AirOps, and Claude. Her team keeps its GEO context files in GitHub, which she says works better than a messy shared drive, even if some marketers think it feels too technical.
The bigger picture
What I like about Andrea’s view is that it isn’t really about a new channel. It’s about accepting reality. B2B buyers were always researching in places we couldn’t see. LLMs just make that obvious, and they penalize companies that hide basic information.
That’s why her definition of quality matters so much to me. If the machines are going to describe your company to buyers, the inputs have to be accurate. That applies to the narrative you publish, and it applies to the identity data behind your targeting. Accurate and actionable is a good standard, and it’s the one we try to hold ourselves to at BDEX.
Connect with Andrea: Find Andrea Lechner-Becker on LinkedIn (she asks that you send a connection request, not an InMail) and learn more about GNW Consulting at gnwconsulting.com.
Work with BDEX: Want identity data that’s accurate and actionable? Visit bdex.com and click “Talk to an Expert.”
Watch the episode: Marketing Measurement in the AI Era: Identity Resolution & Data Quality
This article was adapted from an episode of Deconstructing Data, BDEX’s weekly podcast on data-driven marketing. Tune in live every Thursday at 4:15 PM Eastern on LinkedIn.
About BDEX: For companies that need clean identity data to power their products, BDEX offers unmatched quality and execution. Visit bdex.com and click “Talk to an Expert” to get started.
