When we first started hearing about generative engine optimisation framed as the next big thing in search, our reaction was scepticism mixed with déjà vu. The industry was naming and productising something before there were stable signals to optimise against. That is always a warning sign.
After testing GEO strategies with clients across different sectors, our thinking has shifted. Not towards hype, but towards nuance.
The most honest answer to "Should we invest in GEO?" is it depends. The important part is what it depends on: your query mix, buyer maturity, brand strength, and how much of your existing SEO value comes from fragile, top-of-funnel informational traffic.
Here is how to work that out using your own data.
What GEO Actually Means
GEO describes strategies intended to influence how large language models like ChatGPT, Google Gemini, Claude, and Perplexity retrieve, synthesise, and present information. The term itself was introduced by researchers at Princeton University in November 2023.
Context matters. AI platforms are now driving measurable referral traffic. Between September 2024 and February 2025, referral traffic from generative AI rose by 123%, with ChatGPT accounting for roughly 87% of all AI referral traffic across major industries.
But the hype often skips a crucial detail. AI platforms still drive only about 1% of overall web traffic. That is meaningful, but it is not an existential threat to search for most businesses today.
The risk is unevenly distributed, and that is what GEO conversations usually miss.
Set Up Proper Tracking First
Before reallocating a single hour from SEO to GEO, you need a measurement spine that answers one question clearly: if organic clicks go down, do we lose revenue, or do we just lose noise?
At a minimum, that means three things.
Segment organic traffic by intent
If you are still reviewing organic sessions in aggregate, GEO decisions will be guesswork. Classify your top queries and pages into informational, commercial investigation, and transactional intent. Track each segment separately.
AI answer engines disproportionately affect informational demand. If informational traffic drops 30% but commercial stays flat, GEO urgency is low.
Separate branded from non-branded demand
Branded search is your proxy for mental availability. Track branded impressions and clicks in Search Console and look for lagged correlation between informational content publication and branded demand four to twelve weeks later.
If AI answers are replacing clicks but still increasing awareness, branded demand should rise after informational traffic falls.
Connect organic to pipeline, not just leads
Lead volume is a trap here. Track pipeline influenced by organic search using first-touch or meaningful-touch attribution. Monitor conversion rates by intent segment.
If lost traffic was low-intent, pipeline should remain stable. If pipeline moves, the GEO risk is real.
Watch query-level volatility
AI impact is spiky, not smooth. Create a watchlist of 20 to 50 high-value informational queries. Track impressions, CTR, and ranking stability for each. If impressions are flat but CTR collapses, AI answers are intercepting demand.
Capture qualitative sales signals
This is often the earliest indicator. Ask sales teams to note when prospects mention ChatGPT, AI search, or being pre-educated. Mental availability tends to surface in sales conversations before dashboards.
Run a Small, Bounded Test
If your data shows some GEO risk but not enough to justify a wholesale shift, your first test should focus on signal detection, not performance uplift.
You are testing whether AI systems will choose you when you deliberately give them the chance.
Pick a single, high-stakes query cluster. One category, one buyer stage, five to ten closely related queries. AI answers generalise. Scattered testing teaches you nothing.
Use one canonical asset. It should already be authoritative, close to revenue, and capable of holding a point of view. This is not about GEO tactics yet. It is about representativeness.
Make one strategic change only: introduce a named, opinionated decision framework that defines the category, makes trade-offs explicit, and gives conditional recommendations. Consistency matters more than volume.
Define success before publishing. Look for visibility and language signals, not traffic. Are you named in AI answers? Is your framework paraphrased? Are competitors framed relative to your criteria?
Even one or two consistent inclusions are meaningful.
What Actually Drives GEO Outcomes
After testing across clients, patterns emerge.
Intent matters most. Informational queries are where AI answer engines intercept demand most aggressively. Commercial and high-consideration queries still rely heavily on traditional search journeys.
Brand strength outweighs optimisation tricks. In AI answers, being cited or synthesised correlates far more with authority and clarity than with new tactical tweaks. An Ahrefs analysis of 75,000 brands found a 0.664 correlation between brand web mentions and AI Overview mentions, compared to just 0.22 for backlinks.
This looks less like a new channel and more like a redistribution of rewards.
Traffic loss does not always equal value loss. Some displaced clicks never converted. Others, particularly in content-led demand generation, absolutely mattered. Without tying GEO back to downstream outcomes, the discussion stays superficial.
How to Bound the Investment Rationally
There is no universal percentage to shift from SEO to GEO, but there is a defensible way to anchor the decision.
Start from risked value, not budget. Quantify how much revenue is influenced by organic search, excluding branded navigational traffic. Map revenue-influencing pages to intent types and identify which queries trigger AI answers today.
Subtract what GEO cannot realistically defend, such as pure definitions or commodity information. What remains is your defendable value pool.
As a rule of thumb, GEO investment should not exceed the gross margin of the revenue it can plausibly protect or amplify. If $5M of pipeline is genuinely at risk and gross margin is 70%, the theoretical upper bound is $3.5M over time. You will never spend that, but it grounds the conversation in economics.
Then weight amplification upside. If tests show increased brand mentions, shorter education cycles, or rising branded demand despite traffic loss, GEO is not just defensive. It behaves like a brand accelerator.
Apply constraints. If positioning is unclear, cap GEO at 10 to 15% of SEO effort and focus on one or two categories. If positioning is sharp, 20 to 30% is realistic. Beyond that, GEO stops being a search initiative and becomes a brand strategy.
Knowing When to Act
In early 2026, selective urgency beats wholesale action.
Act now if you see two or more of the following: CTR collapse without ranking loss, commercial queries triggering AI answers, competitors being named consistently while you are absent, or sales reporting buyers who already know what they want.
AI systems do not rotate representation evenly. Early incumbents compound. Absence is recoverable early, but persistent absence becomes expensive.
Monitor without panic if organic revenue is stable, branded demand is flat or rising, or your category resists summarisation. In those cases, GEO is brand hygiene, not a growth lever.
One Practical Action to Take This Week
Do not start by writing new content or changing pages.
Instead, identify the one buying decision that usually happens before a sales conversation. Phrase it as a decision statement, not a keyword.
Ask sales what buyers need to decide before talking to you. Then explore how that decision is summarised in Google AI Overviews and one or two AI platforms your buyers use.
Ignore links. Look at who is named, how trade-offs are framed, and what criteria are used.
Be honest. If that answer stood alone, would it naturally point someone towards you?
Compare it with your best existing page. Do you define the decision, or just describe options? Do you impose criteria? Could your explanation be paraphrased without naming you?
If you can write one sentence that captures your point of view on that decision, you have identified your POV gap and your first meaningful GEO test. If you cannot, no amount of GEO work will stick yet.
AI answer engines do not reward activity. They reward decisions made legible. The organisations that adapt are not the ones chasing tactics, but the ones willing to sharpen positioning, reduce content sprawl, and accept that some problems are strategic, not technical.
If this is a conversation you are already having internally, it is often useful to pressure-test the thinking with the team at ADMATIC who are working through the same trade-offs in real client environments.