Our position: charge for the quality of the decision and the work it unlocks, not for a cheap pile of prompts.
What you should leave with
- Separate diagnosis, implementation, and monitoring.
- Price review labor explicitly.
- Define scope with prompt families and repeats.
- Use acceptance criteria the agency controls.
Which AI visibility offers should be priced separately?
Separate the free or low-cost sales sample, full baseline audit, implementation plan, and monitoring program. Each has a different promise, labor profile, and evidence standard; bundling them hides margin and confuses the client.
A sales sample creates urgency and should remain directional. A full audit creates a reviewed baseline and source diagnosis. Implementation changes first- and third-party evidence. Monitoring reruns the stable prompt set, detects regressions, and updates priorities. Treating all four as ‘reporting’ underprices the valuable work.
Make the boundary visible in proposals. State which platforms, markets, languages, segments, competitors, prompt families, repeats, and human-review stages are included. A client can compare scope fairly, and the delivery team can protect quality when requests expand.
- Sales evidence sample
- Reviewed baseline audit
- 90-day fix sprint or retainer
- Recurring monitoring and refresh
Evidence used in this section
What actually drives delivery cost?
Cost rises with market and language count, prompt diversity, repeat runs, source investigation, entity ambiguity, reviewer time, and the specificity of the fix plan. Collection volume alone is a poor pricing unit.
A hundred generic prompts can be cheaper and less useful than thirty carefully designed buying questions. Human work sits in market modeling, classification, source comparison, error verification, and recommendation writing. Show that logic so the price is not compared with a consumer chatbot subscription.
Regulated or reputation-sensitive categories require deeper claim verification and senior review. Multi-location businesses add entity and regional complexity. Price these factors as named scope components instead of quietly compressing quality to preserve a flat fee.
- Prompt and segment design
- Repeat-run and platform breadth
- Manual classification and fact checking
- Source investigation and implementation detail
Evidence used in this section
How do you build a price from scope?
Estimate strategy, collection, review, investigation, reporting, and client-management hours; add external costs and margin; then package the result around a clear outcome. Use change orders for new markets or material prompt expansion.
Start with an internal work model even if the client sees a fixed fee. This exposes unprofitable assumptions, especially around source research and revisions. Keep a complexity buffer for ambiguous brand names, inaccessible pages, and client data that arrives late.
For retainers, reserve capacity for fixes as well as reruns. Monitoring without execution becomes an expensive weather report. State how many prioritized tasks, content briefs, corrections, or implementation hours are included and how unused capacity works.
- STEP 1
Define
List markets, platforms, prompt families, competitors, repeats, and outputs.
- STEP 2
Estimate
Model strategy, review, investigation, reporting, and account time separately.
- STEP 3
Package
Tie the fee to a reviewed baseline, implementation sprint, or monitoring outcome.
- STEP 4
Protect
Write expansion triggers, revision limits, dependencies, and acceptance criteria.
Evidence used in this section
What pricing structure fits each client stage?
Use a fixed fee for the baseline, a milestone or monthly fee for implementation, and a subscription for monitoring. Credits or per-prompt pricing can support software access but should not replace a scoped professional service outcome.
A one-time audit is easiest to approve when the client is uncertain. A 90-day sprint fits a specific backlog. A monthly plan fits categories where sources, products, competitors, and answers change often. Match commitment to the pace of useful decisions.
Discount only when the delivery model gets cheaper, such as a standardized vertical, annual monitoring commitment, or agency volume. Do not cut review depth invisibly; reduce scope and explain the consequence.
| Model | Best fit | Scope anchor |
|---|---|---|
| Fixed audit | First baseline or due diligence | Defined market and reviewed prompt set |
| Sprint/retainer | Known evidence backlog | Deliverables, owners, and milestones |
| Subscription | Ongoing change detection | Stable core plus planned exploration |
Evidence used in this section
Which pricing claims create the wrong expectation?
Do not price or sell guaranteed rankings, guaranteed traffic, or a fixed number of recommendations. Price the auditable work: test design, evidence collection, corrections, content, corroboration, and retesting.
Guarantees invite the team to manipulate the measurement or choose easy prompts. Use acceptance criteria such as approved scope, reviewed classifications, corrected profiles, published pages, and documented repeat tests. These are concrete without overstating control.
Avoid ‘unlimited’ monitoring unless compute, review, and support boundaries truly support it. An apparently simple promise can create a queue of low-value custom tests that crowds out the careful work clients are paying for.
- No recommendation guarantee
- No fixed traffic forecast
- No hidden cap behind unlimited language
- No implementation promise without client dependencies
Method boundary: Pricing examples are commercial design guidance, not a universal market rate. Your labor model, vertical, geography, and service depth determine a sustainable fee.
Evidence used in this section
Questions that change the decision
Frequently asked questions
Should agencies charge per prompt?
Use prompt count as one scope input, not the main value metric. Design, repeat policy, review, and diagnosis can make two audits with the same count radically different.
Can the first audit be credited toward a retainer?
Yes, when the credit is built into margin and the client begins implementation promptly. Keep the audit valuable as a standalone deliverable.
What belongs in a monthly minimum?
Include the stable rerun set, change analysis, source monitoring, a client review, and a defined amount of prioritized fix work or strategic support.
How should multi-location clients be priced?
Price the additional entity, regional prompt, competitor, and source-review complexity. Sampling representative markets can control cost when full location coverage is unnecessary.
Primary sources and research
Platform documentation supports factual statements. Where we describe an audit method or prioritization rule, that is AnswerMentions' operating judgment and is labeled as such.
- [1]FTC: advertising and marketing basicsThe FTC states that advertising claims must be truthful, non-deceptive, and supported by evidence when appropriate.
- [2]NIST: AI Risk Management FrameworkNIST frames AI risk work around governance, mapping, measurement, and management, a useful model for separating observations from decisions.
- [3]OpenAI Help: accuracy and citationsOpenAI warns that ChatGPT can produce incorrect facts and fabricated references, so consequential claims should be checked against reliable sources.
- [4]Google Search Central: creating helpful, reliable contentGoogle recommends original information, substantial analysis, clear sourcing, and content that leaves a visitor feeling they learned enough to achieve the goal.
- [5]Aggarwal et al.: Generative Engine OptimizationThe KDD 2024 paper evaluates generative-engine visibility in a controlled benchmark; it is evidence that visibility can be studied, not a universal ranking recipe.
- [6]Google Search Central: spam policiesGoogle treats scaled pages made primarily to manipulate rankings as abuse, regardless of whether automation, people, or both produced them.