# Building a GEO Business Case (With Template)

> A GEO business case template: problem, evidence, options, costs, measures and risks, with guidance on filling each section honestly.

Source: https://mentionx.ai/learn/careers/building-geo-business-case  
Author: Dipuck Jones, founder of MentionX  
Updated: 30 October 2026  
Level: Senior Leader

**A GEO business case explains why AI answers matter to your company, what you propose to do, what it will cost and how you'll know whether it's working. The strongest cases rely on evidence from your own market, compare at least two options, count people's time as a cost, and set a review point where the work can stop. Use the template below and fill it with your numbers, not industry averages you can't verify.**

*Checked 3 Nov 2026 (CET). "Our view" marks opinion. Examples are fictional.*

> **Key takeaways**
>
> - Build the case on your own evidence, ideally from a pilot, not on industry forecasts.
> - Compare at least two options plus doing nothing, using the same criteria for each.
> - Count people's time as a cost and label every estimate as an estimate.
> - Set a review date and say what result would mean scaling back.

*Checked on 31 October 2026. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.*

## When do you need a business case rather than a pitch?

A pitch wins permission for a small experiment. A business case asks for ongoing resources: a budget line, a hire, an agency, or a meaningful share of a team's time. Most companies want one when spending crosses a threshold or when work will run beyond a quarter.

If you haven't run a pilot yet, consider pitching one first. A business case built on three months of your own results is far more persuasive than one built on forecasts. *Our view:* the best GEO business cases read like a report on a pilot plus a proposal to scale what worked.

## What goes in the business case template?

Copy the sections below into your own document. Replace each prompt in italics with your content. Keep the whole case to three or four pages, plus appendices.

### Section 1: Summary

*Write three sentences: the problem, the proposal and the ask.* For example: "AI assistants rarely recommend us for the questions buyers ask before a demo. We propose a six-month programme to improve how we're described and cited. It needs a part-time specialist, writing support and a tracking tool."

### Section 2: The problem

*Describe what's happening in your market, in business terms.* Name the buyer questions where you're missing or described inaccurately, and the competitors that appear instead. Link the problem to something leadership already tracks, such as pipeline, win rate, brand perception or support volume.

### Section 3: The evidence

*Summarise your own data.* Include your prompt set size, engines tested, runs per prompt and dates. Show results by intent: brand, comparison, recommendation and problem questions. Add AI referral visits from your analytics if you've set them up. Put the raw data in an appendix.

If you ran a pilot, show before-and-after results with the same method, and be honest about sample sizes. Mixed results are fine if you explain them.

### Section 4: Options

*Compare at least two realistic options, plus doing nothing.* Typical options are: keep the work inside the existing SEO or content team with extra time; hire or assign a dedicated specialist; or bring in an agency or freelancer for a fixed period. For each, describe what would be done, who would do it, what it would cost and the main risk.

Doing nothing is a real option. Describe what you think happens if you don't act, without exaggerating. "Competitors continue to be recommended for comparison questions, and we lack data to judge the trend" is honest. "We lose half our pipeline" isn't, unless you can show it.

### Section 5: Costs

*List every cost, including time.* Break costs into people, content, tools and technical work. Use real quotes or internal rates, not guesses. If a number is an estimate, label it. Include one-off costs, like a technical audit, separately from recurring costs, like monthly tracking.

### Section 6: Measures

*Define what success looks like and when you'll review it.* Choose a small set: mention rate across your prompt set, accuracy of descriptions, citations of your pages, AI referral visits and, where you can connect them, leads or opportunities that began with AI research. State the starting values from your evidence section.

### Section 7: Risks and dependencies

*Be candid about what could go wrong.* Assistants change their behaviour often. Results take time. Some work depends on other teams, such as developers for technical fixes or legal for claim reviews. Say how you'll manage each risk.

### Section 8: Decision and timeline

*State exactly what you're asking for and when you'll report back.* For example: "Approve option 2 for six months, with a review at month three. If mention rate on comparison prompts hasn't improved, we'll scale back to a monthly monitoring routine."

## How do you estimate the value side?

This is where most cases overreach. You usually can't prove that AI answers drive a specific amount of revenue, because attribution is incomplete. Assistants often don't send clicks, and many buyers research in one place and convert in another.

Use ranges and stated assumptions instead of single numbers. Start with what you can measure, such as AI referral visits and their conversion rate on your site. Then add what you can reasonably infer, such as sales notes where buyers mention an assistant. Show a cautious, middle and optimistic case, and make clear which assumptions drive each.

*Our view:* a value section that says "we can't yet measure this precisely, here's how we'll learn more" builds more trust than a confident forecast. Finance teams spot invented precision quickly.

## How do you compare options fairly?

Use the same criteria for each option: cost, time to start, expertise, control, risk and how easy it is to stop. An agency may start fastest but cost more and leave less knowledge in-house. A dedicated hire builds long-term capability but takes months to recruit. Extra time for the current team is cheapest but competes with other priorities.

A simple scoring table can help, as long as you explain the scores. Recommend one option and say why, but present the others honestly. Leaders trust cases that show the trade-offs.

## Who should review the case before it goes up?

Ask finance to check your cost figures and format. Ask the web or engineering lead to confirm technical work is realistic. Ask sales whether the problem description matches what they hear. Ask your manager whether the ask fits current priorities. Each review catches a different kind of weakness.

Keep a list of questions reviewers raised and how you addressed them. You'll often hear the same questions in the decision meeting.

## How do you present the case?

Send it ahead of the meeting with a short note: what you're asking for and the decision date. In the meeting, spend a minute on the problem, two on the evidence and options, and the rest on questions. Bring a printed copy of your evidence table and one screenshot of an assistant recommending a competitor.

After the meeting, send a summary of the decision, any conditions and the review date. Then deliver your first update on time. Your next business case will be judged partly on how well you reported on this one.

## From MentionX data: evidence by question type

**Method.** Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.

MentionX's anonymised aggregates show why a business case should break evidence down by intent:

| Intent | Brand mentioned | Share of voice |
| --- | --- | --- |
| Brand-specific | 100% | 64.3 |
| Comparison | 91.9% | 44.6 |
| Recommendation | 29.2% | 26.8 |
| Best-of | 26.6% | 22.7 |
| Problem | 5.7% | 7.7 |

*Our view:* if your own data looks like this, your case is about the bottom three rows. Being named on brand questions doesn't help when buyers ask problem questions that never mention you.

**Limits.** A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.

## Worked example: Ledgerline's six-month case

*Ledgerline is a fictional accounts-payable tool. Figures are illustrative.*

After a 90-day pilot, the content lead at Ledgerline, a fictional accounts-payable tool, writes a business case. Her evidence shows Ledgerline named in more of its comparison prompts than at baseline, but still rarely on problem questions. AI referral visits are small but convert well on the site.

She compares three options: more time for the SEO lead, a part-time specialist, or a six-month agency engagement. She recommends the specialist, shows every cost with finance-checked rates, and offers a cautious, middle and optimistic value range with assumptions stated. The CFO approves it with a month-three review, citing the honest value section as the reason.

> **Pro move:** Add a row to your options table for "how easy is it to stop?" Leaders approve faster when they can see the exit.

## What does a filled-in summary look like?

Here's an example summary written for a fictional company, to show the tone. "Buyers increasingly ask AI assistants which accounts-payable tool to choose. In our tests of 15 buyer questions across three assistants, we were named on most brand questions but rarely on problem questions, where Competitor A and Competitor B appeared instead. We propose a six-month programme led by a part-time specialist to improve how we're described and cited, at a cost set out in section 5, with a review at month three."

Notice what it does. It states the problem in buyer terms, gives the size of the test so readers can judge it, names competitors only as needed, and points to the cost rather than burying it. It doesn't predict revenue. Most leaders decide whether to read further based on this paragraph, so rewrite it until a colleague outside marketing understands it on first reading.

## How do you keep the case current?

Treat the business case as a living document for the life of the programme. Each month, update the evidence table using the same prompts and method, and add a short note on what changed and why you think it changed. At the review date, compare results with the measures you set in section 6 and recommend one of three paths: scale up, continue as planned or scale back.

If assistants change in ways that affect your measures, such as an engine adding or removing citations, record the date and explain the effect. Leaders accept that the field moves quickly. What they don't accept is being surprised. A short, regular update keeps them informed and makes your next request for resources much easier to approve.

For the budget detail behind section 5, see [how to budget for GEO](https://mentionx.ai/learn/careers/budget-geo-tools-people-content).

When a review approves continued work, update the summary and decision sections first, then archive the old version with its date. A clear history of decisions protects the programme when leadership or priorities change.

## Common questions

**How long should a GEO business case be?** Three or four pages plus appendices with raw data.

**Can I use industry statistics?** Only from sources you can verify. Your own data is more persuasive.

**How do I estimate revenue impact?** Use ranges with stated assumptions, starting from what you can measure.

**Should doing nothing be an option?** Yes. Describe its likely effect honestly.

## Sources

- Google Search Central, [AI features and your website](https://developers.google.com/search/docs/appearance/ai-features)
- Google Analytics Help, [Default channel group](https://support.google.com/analytics/answer/13051316)
- Google Search Console Help, [Performance report](https://support.google.com/webmasters/answer/7576553)
All accessed 3 Nov 2026.

## Frequently asked questions

### How long should a GEO business case be?

Three or four pages plus appendices with raw data.

### Can I use industry statistics?

Only from sources you can verify. Your own data is more persuasive.

### How do I estimate revenue impact?

Use ranges with stated assumptions, starting from what you can measure.

---

Written and reviewed by Dipuck Jones for MentionX Learn. The HTML version of this page: https://mentionx.ai/learn/careers/building-geo-business-case