Does GEO Work, or Is It a Waste of Time and Money?

What does the evidence show?

For GEO:

• Mentions change behaviour. Profound's panel study found visits to a brand's site rose to 1.5 to 2.5 times the expected rate in the week after an AI mention. An academic panel study found named recommendations moved behaviour far more than passing mentions.
• Buyers use assistants. BrightLocal found 45% of consumers used AI tools for local business recommendations in 2026. Scorpion found 22% of homeowners used AI tools to research home service providers.
• Evidence moves answers. Peec's analysis of nearly 200,000 AI responses found brands ranked first in frequently cited lists were mentioned considerably more often.
• Content changes can matter. The research paper that named GEO found some changes raised visibility by up to 40% in a controlled benchmark.

Against the hype:

• Direct traffic is small. Conductor's 2026 benchmark put AI referrals at about 1% of site traffic.
• Answers are volatile. seoClarity recorded ChatGPT citation volumes falling by 86–94% in spring 2026 before rebounding.
• Conversion evidence is mixed. Some studies find AI visitors convert better than search visitors. A large study of ecommerce sites found ChatGPT referrals converted worse than organic search.
• Much of GEO is SEO. Google describes optimising for its AI features as still being SEO.

Is GEO worth investing in for businesses of different sizes?

Do Google rankings or AI visibility drive more business value?

Today, for most businesses, Google rankings still drive far more measurable traffic. AI visibility works differently. It shapes shortlists and brand perception, and much of its effect shows up later as branded search or direct visits. Those are credited to other channels. The two are not rivals: rankings feed many AI answers, and AI mentions send people back to search. Measure both, and expect AI's value to be undercounted by last-click reports.

Is AI visibility worth the investment for SMBs compared with Google SEO?

For small businesses, do the shared foundations first. That means a crawlable site, clear service and price pages, a complete Business Profile and genuine reviews. These help Google SEO and AI visibility at once, at low cost. Add the GEO-specific basics, such as Bing Places, Apple and Yelp profiles, allowing AI search crawlers, and a monthly check of what assistants say. Avoid expensive retainers unless they include honest measurement and clear deliverables.

Does GEO drive conversions now, or mainly position brands for future AI search?

Both, in different proportions. Some conversions happen now: buyers who see you named contact you, often after a later search. Ahrefs has reported that AI search sent about 0.5% of its visitors but 12.1% of its sign-ups. Much of the value is positioning, though: being in the shortlists and descriptions that more buyers will see as AI use grows. Plan for both. Track self-reported AI discovery and branded search now, and treat recognition in answers as a long-term asset.

Is B2B GEO being implemented, or is it still hype?

Both. Many B2B companies now track what assistants say, fix wrong descriptions, and work on review profiles and comparison content. Review platforms have published their own analyses of how AI answers use their data. But the market also has a lot of hype. Watch for these warning signs:

• guaranteed placements in ChatGPT or other assistants;
• scores that cannot be traced to stored answers and sources;
• visibility figures inflated by questions that name the brand;
• plans built on publishing large volumes of thin pages.

A real programme has a written method, tracks unbranded questions over repeated runs, and ties its work to specific sources and pages. Whether GEO tools measure anything real is covered in whether GEO tools are reliable or mostly measure noise.

How can you test whether GEO works for you?

Run a small, controlled test before committing a large budget:

1. Pick 20 buyer questions, mostly without your brand name, and split them into two similar groups.
2. Take a baseline: run each question three times in the assistants your buyers use, and record mentions and cited sources.
3. Improve the evidence for one group only, such as the relevant pages, profiles, reviews or list placements.
4. Re-run both groups after six to eight weeks.
5. Compare the change, and check enquiries that mention AI.

If the improved group moves and the other does not, you have evidence that the work helps in your market.

When is GEO a waste of money?

• When buyers in your market rarely use assistants to choose.
• When the basics are missing, such as a crawlable site or accurate profiles, and money goes to advanced tactics instead.
• When a provider cannot show the answers and sources behind its figures.
• When the goal is to game answers rather than improve the evidence.

What might an illustrative assessment look like?

Consider an illustrative 15-person accounting firm deciding whether to spend on GEO. It asks recent clients how they found the firm. Four of 30 mention ChatGPT or Google's AI answers. A check of 20 questions shows the firm named in 2 of 60 runs, with competitors cited from a local business directory and a review site. The firm spends a few days completing profiles, publishing pricing for common services, and asking every client for an honest review in their own words. It skips a proposed $4,000 monthly retainer. Three months later, it is named in 14 of 60 runs, and new client forms show a steady trickle of AI mentions. For this firm, GEO basics were clearly worth it; the expensive retainer was not needed yet.

What do people ask about whether GEO is worth it?

Isn't GEO impossible because AI answers are random?

Individual answers vary, but patterns across many runs are measurable, and they shift when the evidence changes.

How long before GEO pays off?

Search-based answers can change within weeks. Broader recognition and business effects usually take months.

Can we do GEO ourselves?

Much of it, yes: profiles, pages, reviews and monthly checks. Specialists help with measurement at scale and outreach.

What is the cheapest useful first step?

Ask 20 buyer questions in three assistants, record what they say about you and why, and fix the simplest gaps.