How should I price AI visibility as a line item?

Price the work and the scope, not the tool and not an outcome.

3 ways agencies package it

1. Fold it into the retainer. Tracking and reporting become part of the monthly report, and fixes come out of the existing hours. This is simple and suits clients who already trust you. The risk is unpaid growth in scope.
2. A monthly add-on. A fixed fee covers a defined set: a number of prompts, the engines tracked, the markets, a monthly review, and a set number of fixes. This is the cleanest to explain and to measure.
3. An audit, then monitoring. A one-time fee for prompt research, a baseline, and a prioritized plan, followed by a smaller monthly fee for tracking and updates.

What to define in the scope

• How many prompts and which engines.
• Which locations or countries.
• How often results are checked and reported.
• What a fix means: page edits, schema, citation work, profile updates.
• What is not included.

What drives your cost

Your time is the main cost: prompt research, analysis, content and technical fixes, and client calls. Tool cost is real but usually the smaller part, so a price built on "the tool costs X" leaves money on the table. Check how your platform meters work, because caps matter. Doily's Agency plan has pooled credits and hard caps with no surprise overage, so you can estimate what each client's tracking uses.

What not to sell

Do not sell mentions, citations, or a position in an answer. Answers change and no vendor controls them. Sell the work and the visibility into it.

Start with the audit offer for new clients. It gives you a baseline and a reason for the monthly fee.

A cost floor to start from.

Price equals cost divided by 1 minus your target margin. With example numbers: if tooling comes to $25 per client and the work takes 10 hours at an internal cost of $50 an hour, cost is $525. At a 40% target margin the price is $875, since 525 divided by 0.6 is 875.

Those figures are an example. Track 2 or 3 clients for a month before you set a price list, because estimates from memory run low.

Put the limits in writing.

Say which prompts, engines, and locations are covered, how often you report, that AI answers vary between runs and are reported as rates over many checks, and what is billed extra. Promise the work, the method, and the report, not a placement.

Retainer, project, or both.

A monthly retainer fits ongoing monitoring and steady improvement. A fixed fee fits the baseline audit and any one time build. Many agencies end up with both: a project fee to start, then a retainer.

Set a minimum term such as 3 months, since 1 month of data cannot show a trend, and revisit the price after 6 months with real hours in hand.