Monthly AI visibility update for leadership: a one-page template built from weekly reports

A monthly AI visibility update for leadership is a one-page document that answers four questions: what should we decide, what changed, what are we still unsure about, and what do we need from you. It is not the four weekly reports stapled together. Executives do not need the weekly detail; they need the few observations that survived a month of repetition, the limits of those observations, and a clear request. Keep the detailed weekly report as an appendix for anyone who wants to check the work.

In short

• Lead with the decision or request, not the metrics. If nothing needs deciding, say so in one line.
• Show the month as four weekly readings side by side, with the population behind each, instead of a single blended number.
• Keep a short "what we still do not know" section. It protects you from being asked to promise an outcome.
• Every resource request names what it buys and what you will look at afterwards to judge it.
• The weekly reports stay as the appendix; the update links to them and does not repeat them.

What is a monthly leadership update, and how is it different from a weekly report?

A monthly leadership update is a one-page summary written for people who make budget and priority decisions, not for people who run the monitoring. A weekly report (how to write one) is an operating document for the team: it records what was measured, what changed and what to do next. The monthly update sits one level above it and does a different job.

What do leaders actually need to see?

Leaders need the decision, the direction of travel with its caveats, the open uncertainties, and the ask. They do not need definitions of every metric, prompt lists or per-engine tables. Those belong in the appendix.

Give each section one job:

• Decision or request answers "why am I reading this?" in the first two lines.
• The month in four readings shows movement without hiding a single bad week inside an average.
• What we did lists finished work, so effort is visible without being presented as the cause of any change.
• What we still do not know states the limits before someone else finds them.
• Resource requests turns the report into a decision with a cost attached.

How do you condense four weekly reports without misleading anyone?

Show the four weekly readings next to each other, keep each reading's denominator, and do not average them. The safest monthly view is the weekly view made shorter, not a new calculation.

Why not blend the month into one figure? Because the weeks may not describe the same population. The prompt set can change, an engine can be added or paused, and analysis lags collection. In DiscoveredBy's weekly report, every reading measures one Monday-to-Sunday week in UTC, and when an engine was not measured in both weeks, the change is compared over the engines measured in both (Weekly report). A monthly figure you compute yourself does not carry those protections, so a blend can hide a mid-month change in what was measured. If you want a clean month-over-month comparison, fix the prompt set for the period first (use a fixed prompt cohort).

Use the same four readings the weekly report and the client digest carry: brand visibility, citation rate, citation share and average citation position. Define them once, in the appendix. The metrics reference defines brand visibility and citation rate; for the other two, copy the wording from the Weekly report screen rather than paraphrasing. Two definitions matter most for a leadership audience:

• Brand visibility is the share of analysed answers that name your brand.
• Citation rate is the share of collected answers that cite your domain.

They divide by different populations (analysed versus collected answers), so print both counts. The platform marks a metric built from fewer than 30 observations as provisional, so label any slice under that rather than reading a change into it. The weekly report shows the percentages; take the counts from the Explorer for the same Monday-to-Sunday dates, and say so in the update. They may not match the report exactly if data landed after the report was generated.

Enlarge image

The one-page template

Copy this template and fill it in once a month. The bracketed text says what belongs in each slot; delete it as you go. Write the decision section last, even though it appears first.

MONTHLY AI VISIBILITY UPDATE
Month: [month, year]     Weeks covered: [Monday date] to [Sunday date], UTC
Prepared by: [name]      Prior update: [date]

1. DECISION OR REQUEST (two to three lines)
[The one thing we need from you, and by when. If nothing: "No decision
needed this month. Nothing below requires action from you."]

2. THE MONTH IN FOUR READINGS
| Reading | Week 1 | Week 2 | Week 3 | Week 4 | Read it as |
| ------- | ------ | ------ | ------ | ------ | ---------- |
| Brand visibility (n analysed) | | | | | [flat / up / down / too small to read] |
| Citation rate (n collected)   | | | | | |
| Citation share                | | | | | |
| Avg citation position         | | | | | |
Comparable across the month? [yes / no; if no, say what changed: prompts,
engines, or collection gaps.]

3. WHAT CHANGED THAT MATTERS (at most three bullets)
- [An observation that repeated across at least two weeks, with the prompt and engine.]

4. WHAT WE DID
- Published or changed: [pages, with go-live dates]
- Not done, and why: [items]

5. WHAT WE STILL DO NOT KNOW
- [Each open question, what would answer it, and when we will look.]

6. RESOURCE REQUESTS
| # | Request | What it buys | How we will judge it | Needed by |
| - | ------- | ------------ | -------------------- | --------- |

APPENDIX
- Weekly reports: [list the four, with links or dates]
- Definitions: brand visibility, citation rate, citation share, average
  citation position.

How do you write the "what changed" section?

Include a change only if it appeared in at least two of the four weekly reports, or if it is a single event that leadership must know about. A one-week wobble goes in the appendix.

Each bullet should carry four things: what you observed, in which prompts and engines, how many weeks it held, and what you have not established. Compare these two versions.

• Weak: "Visibility improved this month."
• Better: "Quillstone was named in more analysed answers in week 4 than in week 1 (33.0% against 31.4%). The difference is 1.6 points on 280 to 300 analysed answers a week, which we treat as a small movement, not a trend."

The second version can be checked and does not oversell. Stated reasons, citations and mentions describe what an answer contained; they do not prove why an engine produced it, so avoid "the engines now prefer us".

What belongs under "what we still do not know"?

Anything that limits how far a decision can lean on the data. Naming these openly reads as rigour, and it heads off the request to forecast an outcome you cannot measure.

Typical entries:

• Cause. Pages changed during the month and readings moved, but a before-and-after alone does not show the change caused the movement (did your content update help?).
• Collection gaps. If an engine did not run on some days, the month is less complete than it looks (check which days each engine actually ran).
• Small slices. A single engine or country under 30 observations is provisional.
• Business impact. Answer presence and measured referral traffic are different things, and revenue effects are not measured by the visibility readings.
• Pages too new to read. Recently published pages have had little time to appear in answers; give the date you will check them.

Each entry should say what would resolve it and when you will look. "We will know more" is not an entry.

How should you frame resource requests?

Frame each request as a bounded piece of work with a way to judge it afterwards, not as a promise of a result. Leadership can approve "two weeks of editor time to correct three pages" far more easily than "invest in AI visibility".

Use this pattern for every row of section 6:

1. The request: hours, budget or an approval, stated plainly.
2. What it buys: the specific pages, prompts or fixes.
3. How we will judge it: the readings and the dates you will compare. Describe it as a test of a hypothesis, not a guarantee.
4. Needed by: the date after which the work slips.

If your team keeps open work in DiscoveredBy's Actions → To do list, which gathers open citation gaps, optimization drafts, alerts and other records in one prioritised list (Actions), you can size the request from it. Ownership and deadlines outside those records stay in your own tracker.

Worked example: Quillstone's September update

Quillstone sells document-review software to mid-sized legal and compliance teams. Northfold Digital writes weekly reports for Quillstone's marketing lead and a monthly update for the leadership team. This is the update for the four weeks from Monday 31 August to Sunday 27 September 2026. Percentages are as the weekly reports show them; the counts behind them were taken from the Explorer for the same dates.

MONTHLY AI VISIBILITY UPDATE
Month: September 2026   Weeks covered: 31 Aug to 27 Sep 2026, UTC
Prepared by: Northfold Digital

1. DECISION OR REQUEST
Approve the corrected pricing copy on the public pricing page by 2 October.
Two of the four weeks showed an answer quoting an outdated price from a
third-party listing. No other decision is needed this month.

The four weekly readings:

Check the arithmetic: 88 / 280 = 31.4%, 92 / 288 = 31.9%, 90 / 290 = 31.0%, 99 / 300 = 33.0%; 42 / 284 = 14.8%, 45 / 300 = 15.0%, 44 / 299 = 14.7%, 45 / 300 = 15.0%. Citation share and average citation position would sit in the same table in a real update and are left out of this excerpt.

2. (continued) Read it as: brand visibility moved from 31.4% to 33.0%
(+1.6 points) between the first and last week, with a dip in week 3.
Citation rate stayed within 0.3 points. We treat both as flat to slightly up.
Comparable across the month: yes. Same prompt set and same engines
every week.

3. WHAT CHANGED THAT MATTERS
- The pricing prompt on Perplexity quoted an outdated price in weeks 2 and 3.
  It did not in week 4, the week after a correction request went to the
  listing site. One week is not enough to call this resolved.
- Two comparison roundups that appeared as sources in all four weeks do
  not mention Quillstone.
- Brieflane was named in more analysed answers than Quillstone in all
  four weeks. The gap did not change.

4. WHAT WE DID
- Published: "Document review checklist for compliance teams" (10 Sep).
- Changed: the redlining guide (10 Sep), the security page (17 Sep).
- Not done: the pricing page fix (awaiting approval; see section 1).

5. WHAT WE STILL DO NOT KNOW
- Whether the week 4 rise reflects our page changes. Two pages went live
  on 10 and 17 Sep; we will compare the same prompts on 12 October.
- Whether pricing answers are fixed. One clean week; we will re-check
  weekly for four more.
- Business impact. These readings do not measure pipeline.

6. RESOURCE REQUESTS

Notice what the update leaves out: prompt lists, per-engine tables, answer excerpts and metric definitions. All of it is in the appendix, one link away.

How do you use DiscoveredBy to build the update?

Use the weekly reports as your source, and assemble the monthly page by hand. Depending on your plan (plans and limits), DiscoveredBy builds a dated weekly report for each project each week, skipping a project that has no data for that week, and keeps past weeks available from a dropdown (Weekly report). It does not build a monthly rollup for you, so treat the four saved weeks as your inputs.

Some supporting features are worth knowing:

• The Client reports overview lists each accessible project's report for one common completed week, so an agency can pull the same week across clients. Its readings are separate per project, not an average or a ranking (Client reports).
• If you want a supporting view of several measures for the appendix, a dashboard holds up to 12 saved Explorer views. A shared dashboard link is a frozen snapshot of the results at the time it was made (Dashboards and sharing).
• Client report links share a fixed digest of one weekly report and expire after 1, 7 or 30 days; check what the preview contains before sending (Client report links). Anyone holding a link can open it, so treat it as an access credential.

Enlarge image

Common mistakes and limits

• Averaging four weeks. A blended number hides changes in what was measured. Show the weeks.
• Leading with a metric. If the first line is a percentage, the reader has no idea what to do with it.
• Implying cause. Work done and readings moved in the same month is a sequence, not a proof.
• Promising a number next month. Visibility readings are observations of answers; do not forecast them, and do not convert them into revenue.
• Dropping the uncertainty section in a good month. It matters more when results look good.
• Repeating the weekly report. If the update needs three pages, your appendix is in the wrong place.

What this update cannot tell leadership: why an engine wrote what it wrote, what customers did after seeing an answer, or how a metric will move next.

Frequently asked questions

How long should a monthly AI visibility update be?

One page for the update, plus an appendix. If it runs longer, move detail into the appendix. A reader should be able to find the decision or request in the first two lines.

Should I include competitor comparisons?

Include one only if it changes a decision, and only if it is like for like. Share of voice depends on the competitors you chose to track, so adding a competitor can lower it without any change in the answers (metrics reference).

What if nothing changed this month?

Say so in the decision line, then use the rest of the page for stability, work done and the next test. A flat month with a clear next step is a complete update.

Can I send leadership a live link instead of a document?

You can share a fixed snapshot of a weekly digest or a dashboard, but both are frozen at creation and expire. Neither is a monthly summary, so the one-page document is still the thing that carries the decision.

How is this different from a quarterly review?

A monthly update supports ongoing decisions; a quarterly review looks back over a longer period with the client or leadership (review a quarter of AI search work). Build the quarter from the monthly updates.

Next step

Once your weekly reports are running, you can write the monthly update from them by hand. Sign in or start a project, run a few weeks of monitoring, and use the template above on the first month you have four comparable weekly reports. See alerts and reports for how the weekly report fits with the rest of the platform.