Key Takeaways
- Re-check only what actually moves. Three things change fast enough to be worth a monthly look: who gets named for your best question, whether you appear at all, and which pages the live tools are citing. Everything else changes slower than you can act on it.
- Keep the questions fixed between runs. A monthly check is only readable if the questions stay the same. Change the wording and you have measured two different things, which is how people convince themselves of movement that never happened.
- Treat a one-run change as noise. These answers vary between runs even with no change on your side. A single month up or down means very little. Three months pointing the same way is a signal worth acting on.
- Put the slow things on a trigger instead of a calendar. Your own description, your listings and your service pages do not need a monthly review. They need reviewing when your business changes, which is a different rule and a much cheaper one.
- Cap the whole thing at twenty minutes. A monthly check that takes an afternoon gets skipped by March. Five questions, two assistants, one line written down per question is enough to catch every change that matters.
Most people who start checking whether AI recommends them do it obsessively for two weeks, then not at all for six months. Both halves of that are reasonable responses to not knowing what is worth watching.
Almost nothing in this field moves fast enough to reward weekly attention, and almost nothing is stable enough to ignore for a year. The useful cadence sits between the two, and it is different for different things.
This post sorts the work into what genuinely changes month to month, what belongs on a slower cycle, and what you can stop looking at entirely once it is right.
What should you re-check every month?
Three things. Who gets named for your most valuable buyer question, whether your own name appears anywhere in those answers, and which pages the live search tools are citing when they answer. Nothing else changes fast enough to be worth a monthly look.
The reason those three move and the rest do not is worth understanding. All three depend on what is currently being read on the open web, which changes whenever anybody in your field publishes something. The rest depends on decisions you made, and your decisions do not change while you are not looking.
Google's description of its AI features in Search is explicit that the answers are generated from what its index currently holds, which is a moving target. That is the part worth re-reading. Your own pages are not.
The three things worth a monthly re-run
Start with who gets named for your best question, meaning the one with the most money behind it. That list is the competitive picture in a single line, and a new name appearing in it is the earliest warning you will get.
Second, whether you appear at all, anywhere in those answers. Not your rank or your position, which are not stable enough to track. Present or absent is a reading you can actually trust from one month to the next.
Third, which pages the live tools cite. Those citations are visible, which makes them the cheapest evidence in this entire field. If the same three outside pages keep getting quoted for your question, those pages are the market, and getting described on one of them is the shortest route in.
Who gets named for your top question
Moves week to week
Monthly
Whether your own name appears at all
Moves week to week
Monthly
Which pages the live tools cite
Moves week to week
Monthly
How competitors describe themselves
Moves a few times a year
Quarterly
Your directory and listing entries
Moves when you move it
Quarterly
The questions your buyers actually ask
Moves with the market
Twice a year
Whether your own paragraph is accurate
Moves when your business does
On change only
The dashed line in every track is the one month mark. A bar that ends before it has gone stale within the month and is worth re-running; a bar that runs past it has not. The lengths are an illustrative shape drawn from how these things behave, not measured figures.
What belongs on a quarterly cycle instead
How your competitors describe themselves, and your own directory and listing entries. Both change a handful of times a year, so a monthly review finds nothing eleven times out of twelve and trains you to stop looking.
Competitor descriptions are worth a real read once a quarter, because when somebody rewrites their positioning it usually shows up in the answers a month or two later. Reading the change early is what turns it into a decision rather than a surprise.
Listings deserve a quarterly pass for a duller reason. They rot. Hours, service names and specialties drift out of date, and an assistant reading a stale listing will repeat it confidently. What to do with any of these readings is worked through in what to do with your AI test results.
| What you are checking | How fast it moves | Cadence | What a change means |
|---|---|---|---|
| Who gets named for your top question | Week to week | Monthly | Somebody published something that answered it better |
| Whether your name appears at all | Week to week | Monthly | You entered or left the set of businesses it will name |
| Which pages the live tools cite | Week to week | Monthly | The pages that define your market have shifted |
| How competitors describe themselves | A few times a year | Quarterly | A repositioning that will reach the answers in a month or two |
| Your directory and listing entries | Only when you move them | Quarterly | Usually rot: stale hours, old service names, wrong specialty |
| The questions your buyers ask | With the market | Twice a year | Your question set is measuring the wrong thing |
| Whether your own paragraph is accurate | Only when your business changes | On change only | You changed what you do and the description did not follow |
What never needs checking twice
Your own description, once it is right. It does not decay, it does not get quietly rewritten, and re-reading it monthly produces nothing except the urge to fiddle with wording that was already working.
The same goes for your service pages and your proof pages. These belong on a trigger rather than a calendar: when you change what you do, who you serve, or what you charge, update them that week. Otherwise leave them alone.
There is one exception worth naming. If an assistant starts describing you inaccurately, the fix is upstream in those pages even though they did not change, because something else about you did. The signals to watch for are covered in how to tell if AI visibility is improving.
The twenty minute version
Pick five questions and never change their wording. Ask them in the same one or two assistants each month, three times each, and write one line per question: who got named, and whether you were one of them.
Do it on the same day each month. Answers vary between runs on their own, so a consistent routine is what makes the variation readable rather than alarming. OpenAI's own description of what ChatGPT is is a useful reminder here: it generates a response each time rather than looking up a stored one, so two runs of the same question are two samples, not a contradiction.
Where the live tools are concerned, note the cited pages as well as the names, because the sources are shown. Perplexity documents that behavior in its help center, and it turns a vague question into a list you can act on. The wider question of how often any of this is worth doing is covered in how often you should check if AI recommends your business.
One honest limit. If you have made no changes and nobody in your field has either, a monthly check will show you the same answer for months, and that is a correct result rather than a broken process. Checking more often will not make it move.
If you would rather not run it yourself, the Watch re-runs your board every week and sends a digest of what moved, at $399 a month or $300 a month added with the audit. Or start with the free check: it asks ChatGPT 25 real buyer questions from your niche at $0 and emails a plain-English readout. It runs as soon as you ask for it, and the report lands within the hour between 7am and 5pm Central, or by 10am the next morning outside those hours.
Questions we hear the most
What should I re-check every month for AI visibility?
Three things: who gets named for your most valuable buyer question, whether your own name appears anywhere in those answers, and which pages the live search tools are citing. Everything else changes slower than a month.
What can I safely leave alone?
Your own description, your listing entries, your service pages and your proof pages. Those change when your business changes, not on a calendar, so review them on that trigger instead of re-reading them every month.
How many questions should a monthly check cover?
Five is plenty for a monthly run, chosen from the questions with real money behind them. Keep the wording identical every time, because a reworded question measures something different and quietly resets your baseline.
What counts as a real change rather than noise?
A move that holds for three consecutive months, or a name that appears or disappears across several questions at once. Assistants vary between runs on their own, so a single month in either direction proves very little.
Should I check every AI tool every month?
No. Pick the one or two your buyers actually use and stay with them. Adding tools multiplies the work and mostly produces the same finding, which is that the same handful of well-described businesses keep getting named.
What do I do when a competitor appears where I used to be?
Look at what changed on their side rather than yours. Usually a new page answering that exact question, or a new outside mention. That page is the thing to study, and often the thing to answer better.
Can AnswerHalo run the monthly check for me?
Yes. The Watch re-runs your board every week and sends you a digest of what moved, for $399 a month, or $300 a month when it is added with the audit. Nothing to log into, and no call at any point.
How do I get a first reading to compare against?
The AnswerHalo free check asks ChatGPT 25 real buyer questions from your niche and emails you a plain-English report: your score, who got named in your place, and the one fix to start with. It costs $0 and it runs as soon as you ask for it.
Your buyers are already asking. Find out what AI tells them.
The free check asks ChatGPT 25 real buyer questions about your niche and sends you the report within the hour, at $0. If it shows you are already getting named everywhere, we will say so in plain words.
Get my free check →