How to report AI visibility to clients, with a template

An agency owner asks this before the first client report of the month. This guide gives the five numbers a client report needs, their order, and a template. The example is our own project, with the numbers of its last 28 days. For the tools that print your logo on the report, read our guide on white-label reports. Our guide for agencies compares the tools on client projects, seats and client reports. For what an agency charges for the work, read our guide on agency or tool.

How should an agency report AI visibility to clients?

Report one closed calendar month, with five numbers in a fixed order, and each number's change on the calendar month before. One page carries the numbers. Our own figures below use a trailing 28-day window instead, because the month of the read is still open. Our monthly report draws the first four as tiles and states the fifth under Sources. Keep the stored answers behind the numbers at hand.

The five numbers, in the order the report prints them:

1. Visibility. The share of analyzed answers that name the client's brand.
2. Share of Voice. The brand's place among the monitored brands by mentions. Our guide on Share of Answer and Share of Voice defines it with its denominator.
3. Sentiment. The share of the brand's mentions that read positive, with the counts of neutral and negative mentions beside it.
4. Recommendation. The share of the brand's mentions that recommend it, with the count of recommended mentions beside it.
5. Own site cited. The client's domain among the domains the answers cite, by citations.

Each number states its base. On our project, in the 28 days to 30 October 2026, visibility reads 14 of 8,953 answers, not 0.2% alone. A share without its base reads the same for a month of few answers and a month of many.

What do the numbers look like on a real project?

On our own project, in the 28 days to 30 October 2026, on 7 AI platforms, the last three numbers read as the table shows. Our guide on markets holds the visibility and Share of Voice of the same 28 days, split by market.

The three numbers rest on 14 mentions, so one mention more or less moves sentiment by about 7 points. The base tells the client how far to trust the share.

A new client has no stored answers, so its first report has no month before. Read the client's brand in today's answers first, so the agency knows the base the first report starts from.

One page per platform, or one pooled view?

Both, in that order: the pooled numbers on the first page, and one line per platform below them. The platform lines show where the pooled number comes from, because the platforms differ. In our 28 days to 30 October 2026, Copilot named a monitored brand in 67% of its answers, the highest share of any platform. Our guide on AI visibility tools gives the lowest share and the platform behind it.

Our report prints one line per platform under the tiles, with its visibility and its answer count. A platform with no answer in the month prints "Not measured this month" instead.

The same split holds for markets. Pick one market per report, or state "All markets" in the header. Our guide on markets shows why the same prompt answers differently per country, and how to roll several markets into one scorecard.

How do you show the change against the month before?

State each change in points, next to its number, against the calendar month before. Four rules apply:

1. Compare shares, not counts. A calendar month runs 28 to 31 days, so a count of answers changes with the length alone. A share keeps its meaning across months of unequal length. A trailing window compares with the window of the same length before it.
2. Show no change when the month before holds fewer than 7 measured days. A client's first month has no month before, so its tiles carry no arrow.
3. Write points, not percent. Our visibility rose from 0.0% to 0.2% in the 28 days to 30 October 2026, against the 28 days before. That is 0.2 points. Written as a percent change, a rise from zero has no value at all.
4. Date the read. A new competitor or a new spelling back-fills old answers, so a figure can change after you sent the PDF. The coverage line states the day of the read and the competitors monitored on that day.

Our own project shows how far the numbers move. In the 28 days to 30 October 2026, the five rivals we monitor each rose by 1.8 to 15.2 points on the 28 days before. In the week of 19 to 25 October 2026, three of them fell, and none by more than 1.3 points. On our project, the weekly moves were smaller than the monthly ones. Pick one window and keep it. The report repeats each month, and each month compares with the one before it.

Rule 2 asks for 7 measured days in the month before. By hand, that means the same prompts on every platform, asked again on each of those days. A tool asks them each day on a schedule, so the month before fills with no daily work from the agency.

What goes in the template?

The template is one page, in this order. The header carries the agency's logo and name, the client's brand and domain, the month and the market. Under it, our monthly report prints these sections, and a project turns each one on or off.

The summary answers one question: do AI answers name the brand more often than its competitors? Our report writes it from the figures. It opens with Yes or No and the leader's Share of Voice, then the brand's share and its place. It ends with the brand's visibility, its base, and the change in points on the month before, by the month's name. On our project the two parts read as follows. In the 28 days to 30 October 2026, Profound leads with 29.3% Share of Voice, and FullReach AI has 0.1%, 6th of 6. In the same 28 days, 0.2% of 8,953 answers name FullReach AI, up 0.2 points on the 28 days before. The note is the one part a person writes. Keep it to what the agency did and what it does next.

Copy the template below into a document or a sheet, and fill each bracket.

[Agency logo] [Agency name]
AI visibility report: [Client brand], [client domain]
Month: [Month and year]. Market: [Market, or All markets]

SUMMARY
[Yes or No]. [Leader] leads with [x.x]% Share of Voice.
[Brand] has [x.x]%, [place] of [brands].
[x.x]% of [answers] answers name [Brand], [up or down] [x.x] points on [month before].

NOTE
What we did: [one or two sentences]
What comes next: [one sentence]

COVERAGE
Days measured: [n]. Analyzed answers: [n]. Platforms: [n]. Prompts: [n].
Markets: [list]. Competitors monitored: [list]. Read on: [date].

NUMBERS: value | base | change in points on [month before]
Visibility:      [x.x]% | [n] of [n] answers      | [+/- x.x]
Share of Voice:  [x.x]% | [place] of [n] brands   | [+/- x.x]
Sentiment:       [x]% positive | [n] neutral, [n] negative | [+/- x]
Recommendation:  [x]% recommended | [n] of [n] mentions | [+/- x]
Own site cited:  [place] of [n] domains | [n] citations | [+/- places]

PLATFORMS: platform | visibility | analyzed answers
[Platform] | [x.x]% | [n]

COMPETITORS: brand | visibility | Share of Voice | change in points
[Brand] | [x.x]% | [x.x]% | [+/- x.x]

DEFINITIONS
[One sentence per number.]

What to do next

1. Copy the template. Fill its header and its coverage line for one closed month and one market.
2. Put the five numbers in the order above, each with its base and its change in points.
3. Write the note in two or three sentences. Send the PDF, and keep the answers behind each number ready for the client's questions.