# How do I get my board to approve budget for AI search?

Canonical: https://fortitudemedia.ai/insights/business-case-ai-optimisation-board
Author: David Adams
Published: 2026-04-22
Last reviewed: 2026-09-24
Reviewed by: David Adams
Pillar: Costs, returns and the business case
Summary: How to get board approval for AI search spend: evidence of what ChatGPT says today, a break-even from your own figures, monthly measures and a stop point.
Keywords: business case for ai search, board approval ai search budget, ai search roi, ai search kpis year one, is ai search an investment or a cost, how to justify ai seo spend
Short answer: Show the board what ChatGPT, Gemini and Copilot say when buyers ask for a firm like yours, and tie it to a number they already watch. Work out the break-even from your own gross profit per client, not fees. Ask for a fixed budget with monthly measures, a twelve-month review and stop conditions agreed in advance. Present the work as an asset judged over two to three years, even though it is booked as an expense. Evidence and a capped downside win approval.
What changed: Merged the cost-or-investment guide, fixed the break-even to use gross profit, added the ID-Pal before and after, year-one measures, budget source and 2026 sources.
Publisher: Fortitude Media Limited

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

---
You have typed a few of your buyers' questions into ChatGPT and watched it name two competitors and not you. Now you need the board, your co-founders or the chair to agree to spend money on something that will not show up in the pipeline report for months.

Boards are not hostile to new ideas, only to vague ones. A paper that shows what buyers see today, ties it to a number the board already watches, and sets out a cost, a measure and a point at which you would stop usually gets approved.

Market figures help, but they will not carry the paper on their own. 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, and [Forrester reported in January 2026](https://www.forrester.com/blogs/b2b_buyers_make_zero_click_buying_number_one/) that 94% of business buyers use AI somewhere in their buying process. Both studies are international and lean towards technology purchases. What wins the vote is what the assistants say about your own market, which you can gather this week.

## Key takeaways

- Boards approve specific, evidenced proposals, so show them what ChatGPT, Gemini and Copilot say about your market today rather than forecasts about AI.
- Tie the proposal to a number the board already watches, such as enquiries, win rate or revenue from new clients.
- Work out the break-even on gross profit, not fees, because a finance director will redo the sum with margin.
- Present the work as an asset that keeps working after you stop paying, judged over two to three years, even though most of it is booked as an expense.
- Promise measures you can report every month, such as how often you are named for a fixed list of buyer questions, rather than a guaranteed return.
- Say in advance what would make you stop, because a defined exit makes a cautious board far more willing to start.

## What does a board need to see before it approves this?

Four things: evidence that the problem is real in your market, a plan with a cost, a way to tell whether it is working, and a point at which you would stop. Most proposals about AI search offer only the first, and offer it as a general trend rather than as evidence about your firm.

A new office comes with a lease, a hiring plan, a revenue target and a review date; this proposal should look the same. The board's real fear is money spent on activity with no visible result and nobody accountable for it, so answer that fear directly.

## How to build the case, step by step

### Step 1: Show the board what buyers see today

Write down 20 questions a buyer would ask when looking for a firm like yours, in their words: "best employment law firm for tech start-ups in Manchester", "who can help us get ISO 27001 certified". Ask each in ChatGPT, Gemini, Copilot, Perplexity, Claude and Google's AI Mode, in a fresh chat and more than once, because answers vary from run to run. Record who is named and how you are described, and add two or three screenshots: nothing persuades a board faster than a competitor named for the work you do best.

Being known is not the same as being recommended. ID-Pal, an identity verification firm we work with, had bought the UK verification platform NorthRow and folded it into its own site. Our baseline found this: "Across 36 recorded AI answers to the questions ID-Pal's buyers actually ask, ID-Pal was named three times, and only for its brand or its Salesforce integration. AI knew who it was, and had no reason to recommend it." After the fixes, the same questions were asked every month:

| What we tracked for ID-Pal | Result |
|---|---|
| UK buyer questions where AI names ID-Pal | none in May, two in August, four in September 2026 |
| AI answers still recommending NorthRow as a rival | six in May, one in September 2026 |
| Accuracy | 6 of 6 correct answers to 'what does ID-Pal do?', every engine, every run |

That is evidence a board can follow. The full story is in the [ID-Pal case study](/work/id-pal), and [benchmarking how often you are named against competitors](/insights/benchmarking-ai-visibility-competitors) explains how to make your own comparison fair.

### Step 2: Tie it to a number the board already watches

Connect the evidence to a number that comes up at every meeting: new enquiries, proposal volume, win rate or revenue from new clients. If web enquiries have fallen while rankings held, explain why: in a [US study by Pew Research Center (July 2025)](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/), people clicked a traditional result on 8% of visits when a Google AI summary appeared, against 15% when none did. If prospects arrive with a shortlist already formed, or you lost a pitch to a firm named in every answer, say so. The aim is to move from "AI is interesting" to "this affects a number we already worry about".

### Step 3: Do the break-even sum on gross profit

Use four figures your finance director already trusts: the fees from a typical new client over its lifetime, your gross margin on that work, your win rate and the proposed annual spend. Do not borrow return multiples from a vendor's slide.

Example: a typical new client pays £60,000 in fees over three years and your gross margin is 50%, so each client adds £30,000 of gross profit. The proposal is to spend £40,000 a year, or £120,000 over three years. To break even, the work needs four extra clients across the three years. At a win rate of one in three, that is about twelve extra proposals, or four a year.

The board is then judging whether being named by AI assistants will bring four more proposals a year. Done on fees, the same sum seems to need only two clients, and a finance director who redoes it will doubt the rest of the paper. In our experience the first year is when spend runs ahead of return, so show the break-even over the full period. Our guide to [what it costs you not to be recommended](/insights/cost-of-inaction-losing-right-now) sizes the other side of the ledger.

### Step 4: Set out the plan and what you are building

Describe the work plainly: a website that says what you do, for whom and where; useful answers to buyers' questions; your experts quoted where your buyers read; and one consistent description of the firm across the web. Name who does each part, and say how much senior time it needs: a monthly interview with two or three experts, with someone else writing, is a far smaller ask than an article a week.

Then explain why this belongs with investments. The test is what happens the month after you stop paying. Google Ads and sponsored posts rent attention, and the results stop with the invoice. The work that gets you named builds five things that keep working:

- **A library of answers** to the questions your buyers ask, with your view and your evidence.
- **A record of what others say about you:** coverage, guest articles, awards and reviews.
- **One consistent description** of your services, sectors, locations and people, which matters because [AI assistants cross-check what is said about a firm](/insights/ai-cross-references-business-sources).
- **Visible expertise:** named people with real credentials next to your firm's name.
- **A baseline and a record** of how often you are named, so you can see the asset grow.

None guarantees a recommendation, but the earned parts carry most weight. [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 84% came from earned media and 0.3% from paid content. And the asset can bring in business. For Helm Club, a founder membership community we work with, "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." ([Helm Club case study](/work/helm-club))

The argument fails when the work builds nothing, as [why a cheap AI search package rarely works](/insights/cheap-ai-optimisation-doesnt-work) explains.

### Step 5: Agree what you will report and the milestones

Promise measures, not outcomes you do not control. These are the ones we would put in a year-one paper:

| Measure | What it tells the board | When it tends to move |
|---|---|---|
| Share of a fixed list of 20 to 50 buyer questions where you are named, counted only when it repeats | Whether the asset is growing | Within weeks in assistants that search the web live; months in answers from a model's memory |
| Accuracy of how assistants describe you | Whether your facts are consistent | In our experience, often first |
| Visits from AI assistants in Google Analytics | Supporting evidence | Alongside mentions |
| Enquiries that mention an AI assistant, and their pipeline | The business result | Usually later in the year |

Set milestones as targets, not promises: baseline and website fixes in the first quarter; more accurate descriptions and mentions for some specialist questions in the second; a rising share of mentions and the first AI-sourced enquiries in the second half; then a formal twelve-month review. Our guide to [building an AI visibility dashboard and board report](/insights/build-ai-visibility-dashboard) has a one-page format you can copy, and [how to tell which leads came from ChatGPT](/insights/attribution-ai-era-track-leads) covers the enquiry figures.

### Step 6: Say what would make you stop

Most proposals leave this out, and it wins over cautious boards. For example: if the share of mentions across your fixed questions has not moved after nine months, the approach changes; if it has not moved after twelve, the budget goes back to the board. A defined exit turns an open-ended request into a controlled trial, and separates "this is not working" from "this is taking the time we said it would".

## Where should the budget come from?

From whichever line the board expects to produce new clients, without switching off paid search, which brings enquiries quickly while slower work builds.

Some of the money may already sit in your SEO budget. Google's own [guide to optimising for its generative AI features](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide) (May 2026, updated July 2026) says the work is still SEO, so a capable SEO provider may already cover part of the website side. SEO budgets rarely cover earned coverage and reviews. We compare the routes in [whether being named by ChatGPT is cheaper than Google Ads](/insights/economics-ai-visibility-vs-advertising).

## What objections should you expect?

### "How do we know our buyers use these tools?"

Show your Step 1 evidence and ask sales how many recent prospects mentioned an assistant. Many buyers have one at work: [Microsoft 365 Copilot passed 30 million paid business seats in July 2026](https://www.microsoft.com/en-us/investor/earnings/fy-2026-q4/press-release-webcast), and Forrester found business buyers four times as likely as consumers to use Copilot.

### "We rank well on Google, so isn't that enough?"

Ranking helps but does not guarantee a mention. [Ahrefs found in March 2026](https://ahrefs.com/blog/ai-overview-citations-top-10) that 38% of Google AI Overview citations came from top-ten pages, down from 76% in July 2025, because Google now runs several related searches for one question. Covering the follow-up questions matters too, as we explain in [why ranking first on Google does not mean ChatGPT mentions you](/insights/google-rankings-wont-protect-you).

### "Can't we just buy an advert in ChatGPT?"

You can buy an advert, not a place in the answer. ChatGPT has shown labelled ads to UK users on its Free and Go plans since 11 August 2026, and [OpenAI says](https://openai.com/index/testing-ads-in-chatgpt/) the ads do not change the answer itself. Being named is still earned.

### "Most of our work comes from referrals."

In our experience, referred buyers often check the name they were given with an assistant before they call. If it describes you vaguely or suggests alternatives, some referrals leak away. This often persuades a sceptical co-founder, because it is about keeping work you already win.

### "Can't we wait, or do it in-house?"

You can wait, but competitors who start now will have a year of coverage and reviews you cannot buy back quickly; see [whether you can wait a year](/insights/first-mover-advantage-ai). Your team may do much of the work if it has the time, press contacts and technical knowledge; if not, [how to choose a firm to help](/insights/how-to-compare-ai-optimisation-providers) lists the questions to ask.

## How much should you ask for?

A sum you can defend line by line, fixed for twelve months, with the break-even beside it. [What AI search optimisation costs in the UK](/insights/ai-search-optimisation-cost-uk) sets out published market ranges, and our own prices are on [what it costs](/what-it-costs).

If the board is not ready for a year, ask for a fixed-scope first step such as our [Paid Discovery](/discovery): one week, every finding ranked, and fixes written so your own team can ship them.

## What should the board paper look like?

Two pages, in this order:

1. **The situation:** buyers in our market ask AI assistants who to hire, and here is what they are told.
2. **The evidence:** the Step 1 table and two screenshots.
3. **The commercial link:** the board number this affects.
4. **The proposal:** what, by whom, over what period, and the five assets it builds.
5. **The cost and break-even:** the annual figure and extra clients needed, on gross profit.
6. **The measures:** the monthly report and milestones.
7. **The review and exit:** the twelve-month review and stop conditions.
8. **The decision sought:** one sentence.

## What to do this month

1. **Run the questions.** Put 20 buyer questions to six assistants, more than once each, and record who is named and how you are described.
2. **Talk to sales** about prospects who mentioned an AI assistant and pitches lost to firms named in every answer.
3. **Do the break-even sum** on client fees, gross margin, win rate and proposed spend.
4. **Write the two-page paper**, including the measures table and stop conditions.
5. **Pre-brief the most sceptical director**, so their questions are answered in the paper rather than in the room.

## Frequently asked questions

### How long should we ask the board to commit for?

Two to three years in principle, with a formal review at twelve months and monthly reporting throughout. A shorter commitment tends to end just as results start, and the review gives the board a firm point to change course.

### What return can we expect in the first twelve months?

Nobody can give you a reliable figure. In our experience year one brings more accurate descriptions, mentions for some specialist questions and the first enquiries that mention an assistant, while spend runs ahead of return. Judge year one on the leading measures and the return against the three-year break-even.

### What if the board wants a guaranteed return?

Nobody can guarantee what ChatGPT or Gemini will say. A provider can guarantee what it controls: effort, reporting and, for outbound work, meetings held. That is why our outbound service, [Pipeline That Fills](/pipeline-that-fills), carries a guarantee tied to meetings held, while we make no promise about AI answers. Offer the board control instead, and be wary of anyone who guarantees a place in ChatGPT's answers.

### Can we put this spend on the balance sheet as an asset?

Usually not. Marketing spend, including content and PR, is normally treated as an expense, and your accountant decides how each item is booked. Calling it an investment is about how you judge it: over a longer period, with early signals, expecting what you build to keep working. It is a management view, not an accounting one.

### What if we try it and nothing changes?

Then your stop conditions do their job: change the approach at nine months and stop at twelve if nothing has moved. You will still have a clearer website, useful articles and a record of what buyers see.

### Does this make sense for a firm our size?

For most founder-led B2B firms with £1m to £20m+ revenue, the break-even needs only a handful of extra clients over three years. The harder question is whether you can hold your nerve through year one. If you need enquiries next month, fund adverts or outbound first and run this alongside.

A board that sees real evidence, a clear cost and a defined exit will usually give you room to try. If you want that evidence gathered before the meeting, [book a free Growth Review](/growth-review): 30 minutes with David Adams, and three fixes in writing afterwards.

## Sources

- [G2, New G2 research: half of B2B software buyers now start their research 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)
- [Forrester, B2B buyers make zero-click buying number one (January 2026)](https://www.forrester.com/blogs/b2b_buyers_make_zero_click_buying_number_one/)
- [Pew Research Center, Google users are less likely to click on links when an AI summary appears (July 2025)](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/)
- [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)
- [Ahrefs, AI Overview citations from top-ten pages (March 2026)](https://ahrefs.com/blog/ai-overview-citations-top-10)
- [OpenAI, Testing ads in ChatGPT (updated August 2026)](https://openai.com/index/testing-ads-in-chatgpt/)
- [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)
- [Microsoft, FY26 fourth quarter earnings press release (July 2026)](https://www.microsoft.com/en-us/investor/earnings/fy-2026-q4/press-release-webcast)