GEO can be a worthwhile investment when AI answers participate in discovery or selection in your category and your company can serve an identifiable opportunity. Connect commercial questions, a substantiated offer, execution capacity and measurable SOV progress.
You do not need to wait for a universal return formula to investigate the channel. You need a proportionate plan with a baseline, budget, owners and review criteria.
Seven investment criteria
1. Are there questions influencing purchases?
Map problems, alternatives, comparisons and requirements customers actually discuss. A conceptual question may build category understanding; a supplier request belongs to another stage. Do not estimate commercial intent from the number of possible articles.
2. Does the offer fit the need?
A company need not be best for everyone. Identify applications where its differences matter and explain who the solution serves. A specific scope may produce more useful content than a generic contest for “best company”.
3. Is the brand correctly identified?
Check category, relevant location, products and namesakes. If the AI describes another company, correct identity in controlled sources and seek consistency in legitimate references. Named recognition differs from spontaneous discovery; track them separately.
4. Are there assets and evidence to support the answer?
Documentation, authorized cases, examples and operational knowledge support verifiable content. Before scaling, establish which claims can be substantiated. One decision-useful page may be more valuable than multiple versions of the same presentation.
5. Can content be found and read?
Check access, HTML, links and organization. Vercel’s crawler study reinforces that verification. Consistently presenting useful content to people and mechanisms also helps distinguish technical implementation from deceptive cloaking.
6. Can you execute and monitor?
Assign content approval, evidence provision, website fixes and analysis. Without execution capacity, a dashboard may collect unresolved gaps. In a SWAS model, software and specialists should connect to deliverables and responsibilities.
7. Is there a criterion for expansion or adjustment?
Track SOV within a common universe, by intention and engine. Monitor business indicators with appropriate attribution too. Agree in advance what warrants expansion, revising a hypothesis or reducing work that does not contribute.
Designing the first cycle
| Stage | Decision input |
|---|---|
| Baseline | Questions, engines, market, language, formula and starting presence |
| Prioritization | Gaps linked to real needs and differentiators |
| Execution | Recorded content, identity, access or evidence changes |
| New round | Common-core comparison with failures and variation visible |
| Review | Next actions and budget decision |
Fit the cycle to the budget and team capacity. No fixed share must be transferred from SEO. Preserve work generating discovery and conversion, and size the additional investment.
Evaluating a commercial proposal
Ask which components are software, which are service and what will be executed. Request an SOV definition, baseline, period and reviewable evidence. Compare like outcomes: source, mention and recommendation are not equivalent.
The founding GEO paper and review arXiv 2607.14035 inform hypotheses and limits; they do not calculate your company’s financial returns. A proposal may include an SOV-growth guarantee under eligible contractual conditions without promising revenue or fixed rankings.
When to adjust priorities
If the offer is unclear, essential evidence is missing or nobody can execute changes, address those dependencies first. Where opportunity and execution capacity exist, monitored work supports learning and disciplined expansion.
An old baseline—including murmur’s August 2026 Brazilian campaign, conducted in Portuguese before this collection—does not replace a current assessment of your company. Decide using the project’s market, language and timing.
Frequently asked questions
Is it worth investing when strong leaders already exist?
It may be worthwhile for specific intentions where your offer has fit and evidence. Assess the concrete opportunity rather than absolute leadership.
How long until results appear?
The execution timeline can be defined; each engine’s responses must be monitored. Contracts should clarify the evaluation period and guarantee conditions.
How can I tell whether the program is progressing?
Check deliverables, correct identification, retrieved sources and changes in recommendations and SOV within the common universe. Connect signals to opportunities without automatically equating them with sales.
What do I need to start?
A clear offer, access to assets, relevant questions, a baseline and execution and approval owners. The initial scope can expand as results develop.
How murmur works
murmur works through a SWAS model, combining proprietary measurement software with specialists who diagnose, plan and execute GEO to grow Share of Voice (SOV). Eligible contract models may include an SOV-growth guarantee, with the metric, baseline, period, scope and conditions agreed in writing. This does not guarantee an individual answer, a fixed position or sales.
Author: Mateus Gomes, murmur operator. This article explains our approach and discloses our commercial interest.
See also
- What is generative search and how does it change marketing in 2026?
- I lead growth: how much of my SEO budget should I move to GEO?
- Why does AI recommend some brands and not others in 2026?
- Is GEO a fad or does it really work in 2026?
- How to tell whether a GEO agency has real results
- How to measure whether AI cites your brand