AIuthority

Harvard Business Review: AI Upending Marketing on Two Fronts, Emphasizing GEO Strategies

By Charles Ryder

Since late 2022—when ChatGPT shoved generative AI into the mainstream—I’ve watched the marketing playbook change week by week. What began as casual experimentation has turned into a structural shift: customers increasingly discover brands through AI-generated answers, and they’re starting to buy through AI-assisted decisions, too.

Harvard Business Review’s framing lands cleanly: AI is upending marketing on two fronts. The first is already here (conversational discovery). The second is accelerating quickly (agentic purchasing). At the center of the first front sits a capability many teams still haven’t operationalized: Generative Engine Optimization (GEO).

Diagram showing the shift from traditional blue link search to AI-generated answers for brand discovery, key to GEO strategies. The first front: Discovery has moved from blue links to generated answers

For years, “search” meant ranking in Google and earning clicks. Now the customer journey often starts in ChatGPT, Perplexity, Grok, or whatever assistant is baked into a phone, browser, or workplace tool. The behavior looks different:

  • Customers ask longer, more specific questions.
  • They want a synthesized recommendation, not a menu of links.
  • They often never click through.

That last point changes the scoreboard. When an AI response is “good enough,” classic SEO outcomes—traffic, sessions, pageviews—matter less in the moment. Visibility becomes about whether the model mentions you, summarizes you accurately, and includes you in its recommended set.

This is where GEO comes in. Put simply, GEO is the set of strategies that make your content more citable and usable inside AI-generated answers.

What GEO actually optimizes for (and why it’s not just SEO renamed)

GEO makes more sense once you look at what the machine is doing. A generative engine isn’t just “ranking pages.” It’s assembling an answer from patterns, sources, and retrievable documents—then presenting a single narrative.

So the optimization target shifts:

  • From links and rankings
  • To inclusion, citations, accurate summaries, and category association

In practice, that forces a different publishing standard. Content that tends to perform well in AI outputs is:

  • Explicit about what it claims (clear definitions, minimal fluff)
  • Structured for easy parsing (headings, lists, tables, FAQs)
  • Specific (numbers, comparisons, constraints, scenarios)
  • Consistent across channels (site, docs, listings, marketplaces, press)

There’s also a real first-mover advantage. HBR’s point that best practices are still forming matters: when norms aren’t set, disciplined experimentation compounds. I’ve seen teams treat GEO like a side project, then act surprised when competitors show up in AI answers while their brand doesn’t appear at all.

The second front: “Agentic AI” changes buying, not just browsing

If conversational AI changes discovery, agentic AI changes the purchase itself.

HBR’s work on AI agents doing the shopping points to a shift many marketing teams are still discounting: customers won’t only ask AI what to buy—they’ll delegate the buying process end-to-end. The request becomes: “Find the best option under $X, ship it this week, optimize for reviews and return policy.” Then the agent executes.

That creates a new kind of competition. You still need to persuade humans with brand, story, and emotion. But you also need to persuade a machine that is:

  • Comparing specs and constraints
  • Checking availability and fulfillment speed
  • Evaluating price, warranty, return policy, and reputation signals
  • Preferring structured, reliable, machine-readable inputs

This is where marketing runs into ecommerce, operations, and product data—fast.

Comparison chart illustrating the difference between traditional SEO (links, rankings) and GEO (inclusion, citations, accurate summaries). What I’d do now if I were leading marketing in 2026

If you buy the “two fronts” model, the response is straightforward: build a GEO plan for discovery and a machine-readable commerce plan for agents.

Here’s the outline I’d put in motion.

1) Build “AI-citable” content, not just “searchable” content

I’d start with an audit focused on clarity and usefulness, not word count. Specifically:

  • Thin pages that say a lot without stating anything concrete
  • Missing comparison content (e.g., “X vs Y,” “best for,” “not recommended if”)
  • Weak FAQs, missing definitions, and vague use-case framing
  • Inconsistent product claims or positioning across channels

Then I’d rewrite the core pages to be explicit, structured, and reference-friendly—so a model can lift the right details without distorting them.

2) Treat structured data as a growth lever

Schema markup, consistent product feeds, and clean metadata used to feel like technical housekeeping. In an agent-driven world, they’re persuasion infrastructure. If an AI agent can’t reliably extract price, availability, shipping speed, or key specs, you’ve introduced friction into the very channel designed to remove it.

3) Shift measurement from clicks to “presence”

I’d start tracking signals that reflect how AI systems represent the brand:

  • Share of voice inside AI answers (mentions, citations, inclusion in shortlists)
  • Accuracy of AI summaries (misstatements become brand risk)
  • Downstream conversions from AI-assisted journeys
  • Category association (e.g., “best for X”), not just keyword rank

Conversions still matter, but attribution has to move beyond “last click.” Increasingly, there may be no click at all.

4) Double down on what AI can’t commoditize

Puntoni’s point is the right anchor: as AI makes information and transactions easier, the premium shifts to what’s harder to replicate—community, emotional connection, experience, trust, and identity. GEO helps you get discovered. Brand is what gets you chosen and remembered.

Where this leaves the next 12 months

Right now feels like the early SEO era: the rules are moving, and the teams that execute with discipline can lock in durable advantages. GEO is quickly shifting from “nice to have” to “baseline,” and agentic commerce is right behind it.

The winners won’t treat this like a channel tweak. They’ll treat it like a funnel redesign: discoverability in generative engines, machine-readable readiness for agents, and human differentiation everywhere AI makes offerings feel interchangeable.

FAQ

What is Generative Engine Optimization (GEO)?

GEO is the practice of making your content more likely to be included, cited, and accurately summarized in AI-generated answers from tools like ChatGPT and Perplexity.

How is GEO different from SEO?

SEO primarily optimizes for rankings and clicks. GEO optimizes for inclusion and faithful representation in generated responses—often in situations where the user never clicks through to a website.

What should marketers prioritize first?

Start by making core pages explicit and structured (definitions, FAQs, comparisons), then ensure product and business data is consistent and machine-readable across every channel.

Conclusion: GEO isn’t optional anymore—it’s the new visibility layer

HBR’s “two fronts” warning reads like a practical roadmap. If you want to stay visible during discovery, you need a real GEO strategy. If you want to stay competitive during purchase, you need to prepare for agents that evaluate products more like a spreadsheet than a shopper.

If you want a practical way to operationalize that shift—turning AI-era visibility into a repeatable content and optimization workflow—consider AIuthority as part of your GEO toolkit and broader AI marketing readiness plan.