Decision · Price and return
Is AI visibility worth the money? Pricing models, minimum periods and how to work out the return
By Kristian Stig Henriksen · Dear Future
AI visibility is worth the money when your customers actually ask AI assistants about what you sell, and when today's answer points at someone else. Whether that applies to you can be measured before you commit. This guide covers the pricing models the market uses, what drives the price, how minimum periods typically look, and a simple way to work out the return. It names no prices; the price follows the scope and is agreed in a conversation.
First: is the question relevant for you?
Three signs that AI visibility matters for your business. Customers research before they buy, rather than buying on impulse. Your category is one where people compare providers or ask for recommendations. And a quick test shows the assistants already recommend someone, just not you. If all three are missing, the answer may be no, and that is a valid answer too. The guide to checking your AI visibility yourself shows how to run the test in a week.
The four pricing models in the market
- Platform subscription: you get access to a tool, define the questions and competitors yourself, and read the numbers yourself. The price typically follows the number of questions, models and markets.
- Analysis as a project: one measurement with a report and recommendations, bounded in time. Suited to deciding whether there is anything to work on.
- Ongoing measurement with advisory: platform plus people who read the numbers, prioritise and follow up. The price follows the scope of both the measurement and the advice.
- AI SEO as part of a broader SEO or content agreement: visibility in AI answers is one goal among several, and the measurement is often not daily.
The models are not better or worse than each other; they answer different situations. A subscription is cheapest per number and most expensive in your own time. A project is cheapest to start and produces no movement. Advisory costs the most and is the only one that moves anything if you do not have the hands yourself.
What drives the price
- Number of questions: 20 and 120 are different studies, not the same study in two sizes.
- Number of AI assistants and markets: every model and every language is a new row of answers every day.
- Cadence: daily measurement costs more than monthly and is the only kind that can read a change within a month.
- Competitors: how many are tracked on the same questions.
- People: whether someone reads the numbers and writes recommendations, or you do it yourself.
- Reporting: dashboard, monthly review, integration with your own systems.
When you ask for a quote, these six things are what the provider needs to know. Have them ready and you get a number that can be compared.
Minimum periods: what is common
Platforms are typically sold month to month or with a discount for a year. Advisory agreements more often carry a minimum period, and there is a sound reason: a change in AI answers takes weeks to show, and a measurement that stops after four weeks has measured noise. Always ask what happens to data and access when the agreement ends, and whether you can take your measurement history with you.
How to work out the return
The return on AI visibility is the value of being in the answer to the questions that lead to a purchase. The calculation has four parts, and you fill it in with your own numbers:
- How many relevant questions are put to AI assistants in your category each month? Use the search volume of the corresponding keywords as a lower bound; a share of them has moved to the assistants.
- What share of answers mention you today, and what share could realistically mention you in six months? That is the difference in share of voice.
- What is a customer who comes into consideration worth to you? Use your own conversion rate from enquiry to customer and your average customer value.
- Multiply the three parts and compare with what the measurement and the work cost over the same period.
The calculation does not give a promise, it gives a limit: if even an optimistic scenario does not cover the cost, the answer is no for now. If a cautious scenario does, it is worth the money, and the number can be checked continuously, because the measurement delivers the second part every week.
What to have ready to get a concrete quote
- Your category and the most important buying questions, in your own words.
- The competitors you want to be measured against.
- Markets and languages.
- Whether you will read the numbers yourself or want recommendations included.
- When you want to be able to see the first movement.
With those five things, any serious provider can give a quote that can be compared with the others.
Read on: what drives the price of AI SEO, check your AI visibility yourself and share of voice in AI.
From calculation to numbers
Want to know whether it is worth it for you?
Get an AI visibility analysis and see how often the assistants mention you and your competitors today. Scope and price are agreed in a conversation.
Get the analysis →Frequently asked questions
Are there different pricing structures for tracking LLM visibility?
Yes, four are common: a platform subscription, analysis as a bounded project, ongoing measurement with advisory, and AI SEO as part of a broader SEO agreement. In every model the price follows the number of questions, assistants, markets and competitors, the cadence, and whether people are involved.
Is there a minimum period for AI visibility measurement?
It depends on the model. Platforms are often sold month to month; advisory agreements more often carry a minimum period. There is a sound reason for measurement to run for at least some months: a change in AI answers takes weeks to show. Always ask what happens to data and access when the agreement ends.
What ROI can you expect from optimising for AI search?
No serious provider can promise a number, because it depends on your category, your starting point and your customer value. What can be done is to work it out with your own numbers, as shown in the guide, and check the calculation continuously, because the measurement delivers the share every week.
What do you need to provide to get a concrete quote?
Category and the most important buying questions, the competitors you want to be measured against, markets and languages, whether you want recommendations included, and when you expect to see movement. With those five things, quotes can be compared on equal terms.
Can I do AI SEO myself?
Yes, part of the work: clear pages with short answers, definitions, dates and sources are things you can do yourself. What is hard to do alone is the measurement over time and across competitors, and the prioritisation of which questions can be moved. Start by checking your visibility yourself, then decide.