AI Nudge · included in every software plan

Making sure machines can read your business correctly.

Search engines and AI assistants answer questions about your business using whatever they can parse about it. AI Nudge checks what they can actually see, tells you exactly what is wrong, fixes what it can, and then keeps checking every week — because this is a slow, compounding process, not a switch.

Straight up

What this is, and what it is not.

There is a lot of nonsense sold under this heading. Here is the honest boundary of what this feature does, written before you pay for anything.

What it does

  • Checks what machines can read about you — your details, your pages, your structured data, and whether crawlers are even allowed in.
  • Names every problem and tells you the steps to fix it, in order of how much it costs you.
  • Fixes what is ours to fix and tells participating engines the moment a page we host changes.
  • Watches it weekly, forever, so a redesign or a stray setting cannot quietly undo the work.
  • Shows the denominator — you always see what was assessed and what was not.

What it does not

  • It does not rank you in ChatGPT. Nobody can sell that honestly — there is no compliant way to even measure it, let alone promise it.
  • It does not write your reviews or influence what anyone says.
  • It does not edit your website. We have no login to it and never ask for one. We write the fix and hand you a paste-ready snippet; someone with access pastes it in.
  • It does not speed up Google. Nobody outside Google controls when Google re-crawls.
  • It does not promise a position, a rating or a result. Not in month one, not ever.

The score

Five areas, 100 points, and nothing hidden.

Each area starts at full marks and loses the weight of every open problem against it. The same business always scores the same, and every lost point can be pointed at. No model decides the number.

Business detailsentity25 pts

Sixteen details that identify you: address, phone, website, categories, hours, price range, service area, socials, FAQs. Machines answer questions from these.

Website wordingwebsite25 pts

What your own pages actually say. Page titles, description, headings, whether your suburb and category appear in plain words rather than only in a logo.

Structured dataschema20 pts

The structured data behind your site — the machine-readable version of your address, hours and type of business, in a format engines and assistants parse directly.

Crawler accesscrawl15 pts

Whether the crawlers are allowed in at all. A single line in a robots file, or one stray noindex tag, can make an otherwise perfect site invisible.

Reviews and reputationreviews15 pts

Your public rating and how many reviews stand behind it. This is the one area the plaques move directly, and for a walk-in business it is usually the heaviest lever.

Competitor positionnot scored0 pts

On the plan, not built. Worth nothing until it is, rather than quietly worth full marks.

AI answer visibilitynot scored0 pts

On the plan, not built. Worth nothing until it is, rather than quietly worth full marks.

The two zeroes, explained properly

Why two of the seven areas are worth nothing at all

If you have never touched this subject before, the short version is this. There are seven things worth knowing about how machines see your business. We can measure five of them honestly. We cannot yet measure the other two, so we give them nothing — rather than pretend. Here is what each one is, and why it is stuck at zero.

Competitor position · 0 pts

Where you sit against the other businesses near you

This means: of the businesses a customer would realistically choose between, are you the first one a machine mentions, or the fifth? To answer that we would have to go and look up other businesses’ listings — not once, but over and over, for every search phrase, every week. Each of those lookups costs us money, and the answer changes constantly: a rival gets ten reviews, someone updates their hours, and the order shuffles. A number we printed for you on Monday could be wrong by Tuesday, and you would have no way of knowing.

So it is on the plan and it is not built. When it is built, it will be built to a standard where the number is worth reading. Until then it is worth zero.

AI answer visibility · 0 pts

Whether ChatGPT and friends actually name you in an answer

This means: when somebody asks an AI assistant “where should I go for coffee in my suburb”, does your name come out of it? Everybody wants this measured. The honest position is that there is currently no legitimate, affordable way to measure it. The search services that used to make it possible have closed their doors: Brave ended its free tier in February 2026, and Microsoft retired the Bing search API in August 2025.

The only thing left is to sit there asking a chatbot the same question again and again — which gives a different answer each time, is not permitted under those services’ own rules, and would tell you nothing you could rely on. If a supplier tells you they will “rank you in ChatGPT”, ask them how they measure it. They cannot verify it either.

The principle underneath both. If we scored something we do not actually measure, we would have to guess — and a guess only goes two ways. Guess generously and every business on earth gets free marks it did not earn, which makes the score meaningless. Guess harshly and every business gets punished for a gap that is ours, not theirs. Neither is honest, so the areas sit at zero with a line saying why. When we can measure them, they will be worth points.

What the audit looks at

The things we inspect on your site, every week

These are not steps and they are not in an order — they all happen in the same automatic pass. They are simply the six things the audit goes and looks at.

Pages

Up to five pages read

Home, about, contact, services or menu, and one deeper page. Your own site only — we never crawl anyone else’s on your behalf.

Crawlers

Five crawlers checked by name

Google, Bing, ChatGPT search, Perplexity and Claude. We test what your robots file actually permits, one agent at a time.

Fairness

Blocking on purpose is not a fault

If you block the AI training crawler deliberately, that is a legitimate choice. It is reported as information and costs you no points.

Site checks

Sixteen named site checks

Missing title, no suburb in the title, no description, no H1, no sitemap, noindex, structured data missing or the wrong type, phone mismatch, and more.

Feedback

What customers keep saying

Themes pulled from your own private feedback, then checked against your website — if five people mention parking and your site never says it, that is a finding.

Searches

Up to twenty searches

The phrases you ought to be findable for, built from your category, suburb, services and those themes. Add or delete any of them; ours never overwrite yours.

“Why only five pages? I have forty.”

A fair question, and the answer is that everything about the audit is bounded on purpose: a fixed page budget, a short timeout, a cap on how much of a page we will read, and no following links further than one hop from your home page. An audit that wandered around a whole site would be slower, would risk getting our crawler blocked by your host, and would tell us nothing the first five pages do not.

That is because what we are looking for are site-wide signals — page titles, structured data, robots rules, whether your contact details match — and those live on the pages everyone lands on. If your home page has no title and no structured data, page thirty-one will not save you.

Three other things worth knowing. Our crawler identifies itself honestly as NudgeTagsBot/1.0 rather than pretending to be a browser. It only ever runs on a paying customer’s own site — never on anybody else’s. And it runs every single week, so it is a repeated sample rather than one snapshot: if a page was down the day we looked, next week catches it.

The rule that matters most

We cannot check it, so we do not score it.

This is the single most important rule in the whole feature, and it is the one most tools get wrong in their own favour.

62OF 75 ASSESSED

Suppose we hold no website for a business. Three of the five areas — website wording, structured data and crawler access — cannot be assessed at all, because there is nothing to read.

A tool that awards full marks for anything it did not look at would show that business 75 out of 100. That is not a generous score. It is a false one, and it would send an owner away believing three quarters of the job was already done.

So the unassessed areas are removed from the total instead. The business sees 62 out of an available 75, and a plain sentence naming what was skipped and why — in this case, we have no website on file, or we could not read it.

Step by step

How the work actually happens.

Timings below are typical, not promised. The parts we control are fast; the parts search engines control are not, and no supplier can honestly tell you otherwise.

Who actually does what

We do the thinking. The only thing anyone on your side ever has to do is paste in what we hand you — or forward it to whoever runs your website.

We do this
  • All of the diagnosis, every week, automatically. Reading your pages, testing the crawlers, checking your details, pulling the themes out of your feedback, scoring it and ranking the findings. You are not asked for anything and you do not have to remember anything.
  • Fixing everything we host. Your review page, menu page and welcome page, and the structured data behind them, are ours — so we correct them directly.
  • Telling the engines. The moment one of our pages changes we ping Bing, Yandex, Seznam and Naver.
  • Writing your fix for you. For your own website we produce the exact wording and a paste-ready code snippet generated for your business — you never have to work out what to write.
Needs you
  • Pasting the snippet into your own website. We do not have access to your site and we will not ask for it, so we cannot make that edit for you. The instruction we give is: paste it into the <head> of your home page, or into your site builder’s custom-code area.
  • Or forwarding it on. If someone else looks after your website, send them the snippet. It is self-contained — there is nothing for them to figure out.
  • The odd detail only you know. A category, a service area, a phone number that changed. We can tell you it is missing; we cannot invent it.
Day 1

The first audit runs

We read up to five pages of your site, check what crawlers are allowed, look at the details we hold about you, and pull the themes out of your private feedback. Nothing is guessed and no AI decides anything — every check is a fixed rule with a name.

Automatic
Day 1

You get a score with its denominator, and a ranked list

Not a bare number. Every point you have lost is attached to a specific finding, and every finding says what to actually do about it. Anything we could not assess is named as unassessed rather than quietly marked correct.

Automatic
Week 1–2

We fix the parts we host, ourselves

Your review page, your menu page, your welcome page and the structured data behind them are ours, so we simply correct them. We also tell Bing, Yandex, Seznam and Naver the moment one of those pages changes. Nobody has to ask you for anything and nothing waits on you.

We do this — no involvement from you
Week 1–4

We hand you the fix for your own website; someone there pastes it in

We cannot log into your website and we do not want to. So for anything on your own site — a missing page title, no suburb in it, absent structured data — we write the exact fix and generate a paste-ready code snippet for your business. The instruction is literally “paste it into the <head> of your home page, or into your site builder’s custom-code area”. Do it yourself or forward it to whoever runs the site. These are the fixes that move the score most, and they move it when they are pasted, not when they are agreed to.

Needs you — about ten minutes
Week 2–12

The engines re-read you, on their schedule

This is the slow part and no one can shorten it. We tell Bing, Yandex, Seznam and Naver the moment a page of ours changes. Google does not accept those pings at all — it re-crawls when it re-crawls. A change made today may not show up in an answer for weeks.

Out of our hands
Every week

It re-checks, forever

Every location is re-audited weekly. This is the part that matters most over a year: a site redesign can wipe your structured data overnight, and a stray setting can block a crawler without anyone noticing. The weekly check is what catches it.

Automatic
Month 3+

Reviews start doing the heavy lifting

For a business people walk into, review volume and how recent those reviews are is still the strongest signal there is. That is the part the plaques drive, and it is why this works better at month six than at month one.

Compounding

The shape of it

Fast at first, then slow and permanent.

The structural problems are fixable in weeks and they move the number quickly. After that the gains come from crawling, indexing and review volume, which arrive over months. The line below is the shape of the process, not a forecast of your result.

100 — NEVER REACHEDSTRUCTURAL FIXES DONERE-INDEXEDMONTH 1MONTH 3MONTH 6MONTH 12
Month 1

The steep part. Missing details, page titles, structured data, crawler access. Most of a first score is won here, and most of it is quick.

Months 2–3

The quiet part. The fixes are live but the engines have not all been back yet. Little visibly changes. This is normal and it is where people give up.

Months 3–6

The compounding part. Re-indexing lands and review volume from the plaques starts to carry more weight than anything technical we did.

Month 6 onward

The maintenance part. The weekly check earns its place — catching the redesign, the expired page, the setting somebody changed.

The honest bit

Why it gets close to 100 and never arrives.

Two areas score zero on purpose

Competitor position and AI answer visibility are not built. Until they are, they are worth nothing and cannot be earned.

A finding only clears when it is confirmed

Marking something as done does not return the points. The next audit has to actually observe the fix before the score moves.

Your website keeps changing

New pages, new plugins, a redesign. Each change is a fresh chance for something to break, which is exactly why the check repeats weekly rather than once.

The finish line moves

Engines change what they read and competitors improve too. A perfect score against last year’s rules would not be worth having.

Nothing on this page is a guarantee. Search engines and AI assistants decide for themselves what they show and when, and no supplier — us included — can commit to a ranking, a rating or a position on your behalf. What we commit to is this: we check what machines can read about your business, we tell you plainly what is wrong and what was not assessed, we fix what is ours to fix, and we keep checking every week for as long as you are with us. AI Nudge is part of the software: included in every plan, whichever feature you started with, and never charged for separately. It is the software doing this work, so a plaques-only purchase does not include it — there is no subscription behind it to run the weekly check.

Want to see it against your own listing?

We will walk through it in person, on your business, with your real pages on the screen. No obligation and nothing to install to find out.

Part of the Nudge platform

One plaque. Point it at any of these.

The tier price buys one feature and you pick which. Add another later and the plaque on the counter does not change — we repoint it from the dashboard.