# Can we wait a year before doing anything about AI search?

Canonical: https://fortitudemedia.ai/insights/first-mover-advantage-ai
Author: Ross Williams
Published: 2026-04-22
Last reviewed: 2026-09-24
Reviewed by: Ross Williams
Pillar: How buying is changing
Summary: Waiting a year on AI search costs more than a year. Why firms named early get named more, and the smallest start a founder-led B2B firm can make now.
Keywords: is it too late for ai search, first mover advantage ai search, how long to get recommended by chatgpt, when to start ai seo, ai search for small b2b firms, can we replace a competitor in chatgpt
Short answer: You can, but it is rarely the cheapest choice. Buyers shortlist before they contact anyone, and the signals that get a firm named, such as coverage, reviews, links and what a model learned in training, take months to build, so early starters hold more of them. Live search is more open: ChatGPT search, Perplexity, Copilot and Google's AI features read pages when asked, so a clear new page can be cited within weeks. Start small this quarter with one specialism, a fortnightly piece, three case studies and a monthly check.
What changed: Replaced the invented example with Helm Club, split what locks in from what does not, and added ID-Pal, timelines, a cost-in-hours table and 2026 sources.
Publisher: Fortitude Media Limited

When quoting, attribute to Fortitude Media and link to the canonical URL above.

---
Most founders we talk to are not sceptical about AI search. They accept that buyers now ask ChatGPT, Gemini, Copilot and Claude who to hire. Their question is about timing: is this urgent, or can it sit on next year's plan behind the new hire, the office move and the pricing review?

Waiting is not neutral, because buyers settle on a favourite early. The [6sense 2025 Buyer Experience Report](https://6sense.com/newsroom/the-timeline-for-influencing-b2b-buyers-is-shrinking-insights-from-6senses-2025-buyer-experience-report/) (November 2025, more than 4,000 buyers in North America, Europe and Asia Pacific) found that buying groups had already ranked their preferred vendors before first contact in 94% of cases, and bought from that first choice 77% of the time. More of that early research now happens inside an assistant. In [G2's April 2026 survey of 1,076 software buyers](https://www.prnewswire.com/news-releases/new-g2-research-half-of-b2b-software-buyers-now-start-their-research-with-ai-chatbots-302742807.html), 51% said they now start their research in an AI chatbot, up from 29% a year earlier. G2 surveyed software buyers, so the figure will differ for services, but the direction is clear.

So a year of waiting is a year in which shortlists in your field are drawn up without you. Some of the advantage early firms build sticks and some does not. This guide separates the two, shows what an early lead looks like for a real client, and sets out the smallest sensible start, in hours a week.

## Key takeaways

- You can wait a year, but buyers form shortlists before they contact anyone, so a year of absence is a year of shortlists drawn up without you.
- Some advantages build slowly and lock in, such as what a model learned in training, press coverage, reviews and links.
- Live search does not lock in: ChatGPT search, Perplexity, Copilot and Google's AI features fetch pages when a buyer asks, so a clear new page can be cited within weeks.
- It is not too late in 2026: a firm that is absent from AI answers today can break in, as ID-Pal did between May and September 2026.
- The smallest useful start is one specialism, one substantial piece a fortnight, three case studies with outcomes, consistent public facts and a fixed list of buyer questions checked every month.
- That start costs a few hours a week of senior time, which in our view is a smaller bill than the one for catching up later.

## Is it too late to start with AI search in 2026?

No. A rival that started last year holds some advantages you will have to work for, but most AI answers are still put together fresh each time a buyer asks.

Google's AI Overviews and AI Mode, ChatGPT search, Perplexity and Microsoft Copilot all search the web and build the answer from what they find. Ahrefs found in [March 2026](https://ahrefs.com/blog/ai-overview-citations-top-10) that only 38% of AI Overview citations came from pages in Google's top ten, down from 76% in July 2025, because Google now runs several related searches for one question (it calls this query fan-out). Ranking still helps, but a page that answers the follow-up questions well can be cited without being the top result.

Recency also helps a newcomer. [Seer Interactive's July 2026 study](https://www.seerinteractive.com/insights/study-content-recencys-impact-on-ai-visibility-in-2026) found that 75% of pages cited by AI assistants had been updated in the last 12 months. It covered four consumer and finance industries rather than B2B services, but it suggests a well-kept page from this year can compete with an older one.

## What advantage does starting early actually give?

Mostly a head start on evidence. Assistants recommend firms they can see described, discussed and vouched for across the web, and that evidence takes time to build.

Most of what assistants cite comes from other people. [Muck Rack's May 2026 analysis](https://www.globenewswire.com/news-release/2026/05/07/3290268/0/en/generative-pulse-earned-media-consistently-drives-ai-citations-holding-at-84.html) of more than 25 million links cited by ChatGPT, Claude and Gemini found that 84% were earned media and 0.3% were paid. An [SE Ranking study reported by Search Engine Journal in November 2025](https://www.searchenginejournal.com/new-data-top-factors-influencing-chatgpt-citations/561954/) found that the number of referring domains, the other sites linking to a page, was the strongest predictor of ChatGPT citations. Coverage and links both build up slowly.

In our view that creates a loop: being named brings enquiries, enquiries become case studies, case studies earn coverage, and each is evidence for the next answer. No controlled study measures that loop, so treat it as our reading rather than a law. The wider change in buyer behaviour is covered in [how AI is changing the way B2B buyers choose suppliers](/insights/ai-search-change-your-industry); this guide is about timing.

## What locks in, and what does not?

Signals that take months to build lock in; answers fetched live do not. The difference tells you where a late start hurts and where it barely matters.

| Signal | Does it lock in? | Why | How fast it can change |
|---|---|---|---|
| What a model learned in training | Yes, until the next retrain | Memory reflects the web at training time | Months |
| Press and trade coverage | Yes, it builds up | Earned media is most of what assistants cite | Months, one article at a time |
| Reviews on G2, Clutch, Trustpilot and Google | Yes, it builds up | Review sites are cited in B2B answers | Months |
| Links and mentions from other sites | Yes, it builds up | Referring domains predict ChatGPT citations | Months |
| Live search (ChatGPT search, Perplexity, Copilot, Google's AI features) | No | Pages are fetched when a buyer asks | Weeks |
| What your own site says | No | Live search reads the current version | Weeks, once crawled |

The first four rows are where an early rival has an edge you cannot buy back quickly. In the same G2 survey, 45% of software buyers said citations of review sites were the most confidence-inspiring signal in an AI answer, and a rival with three years of reviews has something you cannot produce this quarter. Coverage builds the same way, as we set out in [how long PR takes to work and what a year should look like](/insights/compounding-effect-consistent-pr-12-months).

The last two rows are where a late start matters least, so put your first effort there: clear pages that answer buyers' questions, consistent facts and crawlers let in, while the slower evidence builds behind them.

## What does an early lead look like in practice?

It looks like a channel rivals are not using yet. Helm Club, a founder membership community running chaired peer groups for founders of growing UK businesses, is our clearest example.

"1 in 5 new members this year came through an AI recommendation." As our case study puts it: "A channel that did not exist before now brings one in five new members: founders who ask ChatGPT, Gemini or Claude which peer group to join are told about Helm."

The work behind it was specific: "We built the structured data AI engines read, then a comparison page for every question a founder asks when weighing up peer groups: Helm against YPO, Vistage and the rest." The timing lesson is in the result: "Every comparison page we have published ranks first for its question on Google and the AI engines. The pages not yet built are the ones where a competitor still writes the answer."

That is the first-mover effect in two sentences. For each buyer question, somebody's page is the answer, and a competitor holds the question until someone writes a better one. The full story is in [the Helm Club case study](/work/helm-club).

## How long does it take to get recommended by ChatGPT from scratch?

It depends on which part of ChatGPT is answering. When it searches the web live, a new page or a new article can show up in answers within weeks; when it answers from the model's own memory, the change arrives only when the model is retrained, which takes months.

The same split applies to Perplexity, Copilot, Gemini, Claude and Google's AI features. Clarity fixes are the quickest wins, such as a service page that says plainly what you do and for whom, because live search reads them straight away. Being named consistently for a competitive specialism takes longer, because it depends on coverage, reviews and links.

Two checks come first. OpenAI says a site must allow its crawler, OAI-SearchBot, to appear in [ChatGPT search](https://help.openai.com/en/articles/9237897-chatgpt-search), and Cloudflare has blocked AI crawlers by default on new domains since July 2025. Our guide to [how AI crawlers read your website](/insights/ai-crawlers-vs-google-spiders) covers the rest, and [how often you need to publish](/insights/how-often-publish-content-ai) covers the rhythm.

## Do AI models only learn about my firm when they are retrained?

No. Retraining is one route, and it is slow. The other is live retrieval: when ChatGPT search, Perplexity, Copilot or Google's AI features look something up, they read your pages and other people's pages as they stand that day.

Evidence you publish this year reaches live search once crawled, and joins the web the next models learn from. A firm that waits a year gives up a year of live answers and leaves the model's picture of its field to whoever published in the meantime.

## What does waiting cost while we are not named?

Absence from AI answers sends no signal: you do not lose the pitch, you never hear about it. We set out how to put a figure on that for your own firm in [what it is costing you not to be recommended by ChatGPT](/insights/cost-of-inaction-losing-right-now).

## What is the smallest sensible start?

One specialism, done properly, at a rhythm you can keep after the first quarter, which is where most firms stop. In our experience this is the minimum that moves anything.

### Step 1: Choose one specialism

Pick the area where you are strongest and buyers ask questions, and agree it with your partners or board.

### Step 2: Rewrite that service page

Say plainly what you do, for whom and with what evidence, and put the direct answer in the first paragraph. Kevin Indig's analysis of 1.2 million ChatGPT answers, [reported by Search Engine Land in February 2026](https://searchengineland.com/chatgpt-citations-content-study-469483), found 44% of citations came from the first 30% of a page.

### Step 3: Publish one substantial piece a fortnight

Have it written, or at least shaped, by the person who does the work. Answer the questions buyers ask, including the follow-up questions they ask next.

### Step 4: Ask three clients for case studies with outcomes

Use numbers wherever the client allows. Case studies are evidence others can cite.

### Step 5: Make your public facts agree

Your website, Companies House, LinkedIn, Google Business Profile, directories and review sites should give the same name, description and services. Our guide to [why ChatGPT describes a firm wrongly](/insights/ai-conflicting-information-online) shows how to find and fix the conflicts.

### Step 6: Check a fixed list of buyer questions every month

Write down 20 to 50 questions your buyers ask, run each several times a month in a fresh session in ChatGPT, Gemini and Copilot, and count a win only when it repeats, because answers vary from run to run. The method is in [how to find out whether ChatGPT recommends your firm](/insights/track-ai-recommending-your-business).

## What is the minimum budget for the smallest start?

Mostly time. Done in-house, the minimum start takes a few hours a week of senior time plus some editing; with outside help, you swap some of those hours for a fee. Spend the first hours on what live search reads within weeks.

For example, a firm running the six steps in-house might plan as follows. The figures are round and illustrative, not measured.

| Task | Who | Rough time |
|---|---|---|
| Choose the specialism and rewrite one service page | Founder or partner, plus an editor | Two to three days, once |
| One substantial piece a fortnight | The expert for an interview or draft, an editor to finish | About half a day a fortnight |
| Three case studies with outcomes | Client lead, plus an editor | A few hours each, over the first quarter |
| Consistent public facts | Whoever manages the website and listings | A day, once, then a short check each quarter |
| Monthly check of buyer questions | Anyone careful, with a spreadsheet | Two to three hours a month |

There are three ways to carry that load.

| Route | What you get | What it costs |
|---|---|---|
| Do it yourself | The six steps above, at your own pace | Senior time, as in the table above |
| [Paid Discovery](/discovery), then your own team | One week, every finding under five lenses, ranked, with fixes written so your own team can ship them | A one-off fee, listed on [what it costs](/what-it-costs), then your team's time |
| An ongoing programme | We run the publishing, coverage and site work through [Reputation That Wins](/reputation-that-wins) and [Website That Sells](/website-that-sells) | A monthly fee, listed on [what it costs](/what-it-costs) |

For the ranges other UK providers charge, see [how much AI search optimisation costs in the UK](/insights/ai-search-optimisation-cost-uk).

## What to do this month

1. Pick the one specialism you will build first and agree it with your partners or board.
2. Rewrite that service page in plain English: what you do, who for, why you and with what evidence, with the answer in the first paragraph.
3. Check that your robots.txt file and any Cloudflare setting let in OAI-SearchBot, Claude-SearchBot, PerplexityBot and Bingbot.
4. Put two publishing dates a month in the diary for the next six months, with a named author for each.
5. Email your three best clients this week and ask for a case study with numbers.
6. Write your list of 20 to 50 buyer questions and run the first check, so you have a baseline to measure against.

## Frequently asked questions

### Our competitors are not doing anything. Does that mean we can wait?

Only if you are sure, and you usually cannot be. Firms doing this well do not announce it; the first sign is often a rival appearing in answers you expected to be in. If nobody in your sector has started, being first is the cheapest position to take, because every buyer question is still open.

### If a competitor started a year ago, is it too late for us?

No. ID-Pal shows a firm going from absent to named. At the start, "AI could describe ID-Pal. It never recommended it." The result: "UK buyer questions where AI names ID-Pal: none in May, two in August, four in September 2026." The work included "Five industry pages rebuilt: fund administration, wealth management, legal, accountancy and estate agency." The pattern is plain: "Where a page has been rebuilt, ID-Pal appears. Where it has not yet, it does not, which tells us exactly what to build next." Be narrower than the rival, and pick the questions where you can become the best answer. The detail is in [the ID-Pal case study](/work/id-pal).

### Once a competitor is recommended by ChatGPT, can we ever replace them?

Yes, or at least join them. Assistants usually name several firms, answers vary from run to run, and live search reads the web again each time, so a named rival holds a strong position rather than a permanent one. The way in is through the questions they answer badly or not at all: a narrower specialism, a sector they do not serve, or a follow-up question nobody has written a good page for.

### Is this just SEO with a new name?

Partly. Google's [guide to its generative AI features](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide), last updated in July 2026, says optimising for them is still SEO, and good search work is the foundation. Ranking helps without guaranteeing a mention, though: the Ahrefs figures above show most AI Overview citations now come from outside the top ten. Being named also depends on evidence from others, as we explain in [why ranking first on Google does not mean ChatGPT mentions you](/insights/google-rankings-wont-protect-you).

### Can we do this ourselves or do we need outside help?

You can do the minimum start yourself if a partner will write and someone will keep the rhythm. Firms tend to need help with keeping going, winning coverage where they have no relationships, and structuring the site so assistants read it well. Do not let waiting for help become another reason to wait.

Waiting a year is a decision, and usually the most expensive one available. If you would like to know where you stand before you decide, [book a free Growth Review](/growth-review): 30 minutes with David Adams, and three fixes in writing afterwards.

## Sources

- [6sense, 2025 Buyer Experience Report (November 2025)](https://6sense.com/newsroom/the-timeline-for-influencing-b2b-buyers-is-shrinking-insights-from-6senses-2025-buyer-experience-report/)
- [G2, The Answer Economy: half of B2B software buyers now start with AI chatbots (April 2026)](https://www.prnewswire.com/news-releases/new-g2-research-half-of-b2b-software-buyers-now-start-their-research-with-ai-chatbots-302742807.html)
- [Ahrefs, Only 38% of AI Overview citations come from the top ten (March 2026)](https://ahrefs.com/blog/ai-overview-citations-top-10)
- [Seer Interactive, Content recency's impact on AI visibility in 2026 (July 2026)](https://www.seerinteractive.com/insights/study-content-recencys-impact-on-ai-visibility-in-2026)
- [Muck Rack, What is AI reading? Earned media drives 84% of AI citations (May 2026)](https://www.globenewswire.com/news-release/2026/05/07/3290268/0/en/generative-pulse-earned-media-consistently-drives-ai-citations-holding-at-84.html)
- [Search Engine Journal, SE Ranking data on the top factors influencing ChatGPT citations (November 2025)](https://www.searchenginejournal.com/new-data-top-factors-influencing-chatgpt-citations/561954/)
- [Search Engine Land, Kevin Indig's study of 1.2 million ChatGPT citations (February 2026)](https://searchengineland.com/chatgpt-citations-content-study-469483)
- [Google Search Central, Optimizing your website for generative AI features on Google Search (July 2026)](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)