She is a seasoned GTM and marketing leader with deep expertise in demand generation, marketing operations, revenue strategy, and business transformation. She has spent over a decade helping organizations improve marketing operations through technology, data, and processes. She specializes in building scalable go-to-market programs that align marketing, sales operations, and other teams to create integrated revenue engines. She has conducted extensive research on GEO (Generative Engine Optimization) and AEO (AI Engine Optimization), tracking market trends and job opportunities in this emerging discipline.
Andrea Lechner-Becker, Chief Strategy Officer at GNW Consulting, has worked in revenue operations since 2011, starting in B2C before moving into B2B GTM and marketing leadership. Her team runs its own ongoing research tracking something most companies haven’t started measuring yet: how GEO and AEO jobs, the roles built around visibility in AI-generated answers, are actually being staffed and structured across the industry, published weekly alongside a broader survey of 225 B2B marketing and revenue leaders that GNW Consulting fielded with Demand Metric.
Generative engine optimization, or GEO, is how B2B companies influence the way AI tools like ChatGPT, Perplexity, and Gemini represent and cite their brand.
Unlike traditional SEO, there’s no transparent algorithm to optimize against, which makes GEO an iterative discipline of hypothesizing, testing, and measuring rather than a fixed set of best practices. On this episode of Ops in Motion, Andrea, whose team surveyed 225 B2B marketing and revenue leaders as part of its research, joined Kawal to explain why GEO requires both on-site and off-site work, how to measure AI-driven traffic when referral data is incomplete, and why change management, training, and strong data governance remain essential to a successful RevOps transformation.
Who Should Read This
- RevOps and marketing leaders trying to figure out where AI genuinely changes their workflow
- Teams starting to invest in visibility within AI-generated search results.
- Product Marketing professionals
- Marketing leaders building repeatable GTM motions
- B2B SaaS executives evaluating growth strategy
Key Takeaways
1. Pattern recognition, spotting that a specific step in a process consistently causes a problem three steps later, remains one of the most valuable RevOps skills, and it’s something AI still can’t reliably replicate in human behavior.
2. Change management and training are consistently the most underfunded parts of any rollout, not because they’re hard to understand, but because they don’t have a clean finish line the way a tool implementation does.
3. 73% of GEO and AEO job postings in the research covered on-site optimization, even though off-site presence, reviews, YouTube, Reddit, and LinkedIn also influence AI visibility and are harder to control.
4. Attribution from AI referral traffic is unreliable right now because most LLMs don’t pass referral parameters. Self-reported data (asking how someone found you) is still the most dependable stopgap.
5. Before AI can meaningfully improve attribution or reporting, the underlying data capture and governance has to be solid. Visitor data you don’t capture today is gone permanently.
AI still can’t out-pattern a human, and that’s not changing soon
Andrea’s answer to what’s remained constant in RevOps since 2011 comes down to one skill: pattern recognition. Noticing that a specific step in a process consistently causes a problem three steps downstream, and knowing which part of that process actually needs to change, is a discipline that AI still can’t reliably replicate, particularly when it comes to human behavior rather than structured data.
Data architecture is the other constant. The way a team structures relationships between objects and concepts in its data was foundational back when Andrea was running a B2C data warehouse in 2009, and it’s still foundational now. What’s changing is how much of that architecture AI can help build directly, particularly through newer headless data approaches like the one Salesforce is developing, which Andrea expects will meaningfully simplify a lot of the sloppy, manual data work RevOps teams have historically done. She estimates most companies are still three to five years away from being able to actually adopt that shift at scale.
The sharper point is about who’s actually using AI day to day right now. Andrea sees a real split forming between RevOps professionals genuinely changing how they work with AI folded into their workflows, and those still operating the same way they did two years ago. Her read on that gap is blunt: if your day-to-day process hasn’t changed, that’s worth treating as a warning sign, not a stable position.
“If you’re in RevOps and doing things the same way that you did them two years ago, you probably are missing something.”
— Andrea Lechner-Becker
Change management and training are chronically underfunded because neither has a clean finish line
Asked what creates the biggest bottleneck across organizations trying to align technology, process, and people, Andrea didn’t point to a tool gap. She pointed to change management specifically, the human side of getting someone used to a new workflow, which she considers massively underinvested even at large enterprises with entire teams dedicated to it.
Her example is a familiar one: a typical B2B webinar rollout, where the presenter hasn’t actually prepared, the marketing team writes the title and description based on a vague planning call, and the actual deck gets built the day before the event, if the team is lucky. That pattern repeats across dozens of workflows inside most companies, not because people don’t care, but because falling back into the old way of doing something is simply the path of least resistance.
Training runs into the same problem for a different reason. It’s easy to treat implementation as the finish line, new tech deployed, new Salesforce object built, task complete, but the change management and the training required to make that change actually stick don’t have a natural endpoint. They require ongoing investment, monthly or quarterly check-ins, that’s easy to deprioritize once the initial rollout feels done.
“It’s always been important, it’s always been underfunded, it’s always been under-executed.”
— Andrea Lechner-Becker
Most GEO strategies only cover half the job
Andrea’s research surfaced a specific gap: roughly 73% of open GEO and AEO job postings in a recent quarter covered on-site work only, your website, blog, product pages, despite off-site presence, what she calls “rented land,” mattering just as much for how AI systems generate and cite answers. That includes review sites like G2, YouTube, Reddit, forums, and LinkedIn, which AI Overviews and tools like Perplexity draw on heavily.
Her point about on-site optimization specifically is that it’s necessary but no longer differentiating. Clean page structure, meta descriptions, proper heading hierarchy, and schema markup matter, but once your competitors are doing the same things, on-site work alone won’t create a lasting advantage. The real work, and the harder work, is managing off-site presence deliberately, closer to how a company would manage a PR situation than how it would manage a website.
Publishing speed matters more than most teams expect. LLMs reward timeliness and frequent updates, which means a slow, heavily staged publishing process, however sound it looks on paper, can lose ground to competitors publishing multiple pieces a day. Andrea’s team leans on primary research specifically because it creates an ongoing signal of relevance without requiring a full site refresh every few months, aggregated findings and trends that are genuinely new information, not just republished content with a new date.
“Nothing about on-site optimization is going to create a moat for you to have ongoing visibility wins over your competition. That’s really the job of GEO: how do we as quickly as possible give ourselves a unique advantage and maintain it.”
— Andrea Lechner-Becker
AI attribution only works if your data capture was solid before you started
Referenced research Andrea’s team fielded found that 73% of organizations are tracking AI referral traffic, but only 34% actually trust that data. The core issue is technical: most LLMs, unlike ChatGPT, don’t pass a referral parameter when sending someone to your site, which makes it genuinely difficult to know whether a visitor came from Claude, Gemini, or Perplexity. Cookie consent requirements compound the problem further, echoing the same dark-social measurement challenges marketers have dealt with for years.
Andrea’s practical fix in the meantime is self-reported data: asking directly on forms and sales calls how someone found you, and if they mention AI, following up on what they were actually asking. Beyond that, tracking straightforward GA4 referral data alongside conversion rate and deal size over time is enough to tell you whether GEO investment is working, revenue and pipeline movement is still the metric that matters most.
The bigger warning is about timing. Visitor data you’re not capturing today from AI-referred traffic is gone permanently once that visit happens, there’s no way to retroactively recover it. Before layering AI onto attribution or reporting, the fundamentals have to be in place: solid governance around how a multi-touch journey gets recorded (especially when a single sales cycle involves 20 or more touches), clear rules for when new visit data should override existing records, and enough underlying diligence, consistent UTM tagging across every promoted link, to make that data trustworthy in the first place.
“If you aren’t today tracking who’s coming to your website through ChatGPT, that is a problem you need to solve yesterday. You’re never going to get that data back after they hit your site the first time.”
— Andrea Lechner-Becker
What You Can Do Next
Audit whether your team’s day-to-day workflows have actually changed with AI in the past year. If they haven’t, that’s a gap worth addressing directly, not a stable position.
Budget for change management and training as ongoing line items, not one-time costs tied to a launch date.
Expand your GEO strategy beyond on-site optimization. Review sites, YouTube, Reddit, and LinkedIn all influence how AI systems generate and cite answers.
Start capturing self-reported source data now, both on forms and in sales conversations, since most AI platforms don’t pass reliable referral data.
Confirm your UTM tagging and multi-touch attribution governance are solid before expecting AI to improve your reporting or measurement.
Frequently Asked Questions
Can AI replace the pattern-recognition skills RevOps professionals rely on?
Not reliably, at least not yet. AI still struggles to recognize patterns in human behavior as effectively as an experienced RevOps professional, which keeps that skill valuable even as AI gets integrated into more workflows and systems.
Why do change management and training efforts tend to fail even at well-resourced companies?
Because unlike a tool implementation, neither has a natural finish line. It’s easy to treat a new system as “done” once it’s deployed, while the ongoing work of getting people to actually adopt and sustain new behavior gets deprioritized.
Why is GEO harder to measure than traditional SEO?
Most AI platforms don’t pass referral parameters the way ChatGPT does, making it difficult to know which AI tool sent a given visitor. Cookie consent requirements add another layer of difficulty. Self-reported data and tracking conversion and deal size over time are the most reliable current workarounds.
If your team is trying to figure out where GEO fits into your GTM strategy, or your attribution data isn’t solid enough to trust yet, talk to our team about your setup.
