Building the Business Case

A GEO business case states the problem with evidence, the cost of a bounded programme, an upside estimate built from your own data with every assumption labelled, and the risks, then asks for a staged commitment with a clear decision point. Because AI visibility has no reliable industry benchmarks yet, honesty about uncertainty is what makes the case credible. This lesson walks through each part with a worked, illustrative model.

In this lesson

• The structure of a GEO business case
• Costing people, tools and content
• Building an upside estimate with labelled assumptions
• Presenting risks and a staged plan
• Exercise and quiz

What goes into a business case?

Keep it to one or two pages. Leaders read short, specific cases.

How do you present the evidence?

Use your own baseline: a manual check or tracked prompt set (see GEO 101). Illustrative:

• "In 40 discovery prompts across ChatGPT and Perplexity, we were recommended in 7; our two main competitors in 22 and 18."
• "Three answers quoted our 2023 pricing."
• "GA4 shows AI assistant sessions growing month on month from a small base."
• "Sales report prospects mentioning ChatGPT in discovery calls."

State sample sizes. A small sample is fine if you say so.

How do you cost the programme?

Present a pilot cost (for example one quarter) and a full-year cost separately.

How do you estimate the upside honestly?

Build a simple model from your own funnel, with every input labelled as data or assumption. Illustrative model, not real data:

Annual upside = 20 × uplift × 12 × 15% × €12,000:

The uplift is the weakest link, so say so: "We don't know how visibility translates to demos; the pilot will test this."

What risks should you include?

How do you structure the ask?

Ask for a stage, not a forever programme:

1. Pilot (one quarter): baseline, fix access, improve 5–10 priority pages, set up attribution.
2. Decision point: review visibility, accuracy and early outcomes against stated thresholds.
3. Scale (if justified): budget for the next two quarters.

Illustrative threshold: "Continue if unbranded recommendation rate on priority prompts rises meaningfully against control topics, or AI-sourced demos increase."

A quick example

Illustrative. A CMO rejects a first draft promising "3× AI traffic". The rewrite shows a pilot costing one-third of a content marketer's time plus a tool subscription, a three-scenario upside table with the uplift labelled as an assumption, and a review date. It's approved because it's honest and bounded.

Exercise

Write a business case covering cost, a labelled upside estimate and risk. Use the structure above, your own funnel data where possible and low, mid and high scenarios. Mark every assumption clearly, and end with a pilot ask and a decision date.

Quiz

1. What makes a GEO business case credible?

a) A guaranteed ROI b) Evidence, labelled assumptions and a staged ask c) Industry hype Answer: b

1. In the upside model, which input is usually the weakest?

a) Monthly demo requests b) The assumed uplift from better visibility c) Deal value Answer: b

1. What should a pilot include?

a) A decision point with thresholds b) No measurement c) A five-year commitment Answer: a

1. Why present low, mid and high scenarios?

a) To show uncertainty honestly b) To confuse leaders c) It's legally required Answer: a

1. Which is a mitigation for "improvement doesn't reach pipeline"?

a) Self-reported attribution and AI referral tracking b) Stop measuring c) Publish more thin pages Answer: a

Next: Lesson 3: Budgeting tools, people and content

Sources

• Google Search Central, Creating helpful content
• Google Analytics Help, About key events