What should AI know about your childcare centre? A practical improvement plan

To improve what ChatGPT, Google AI Mode and Microsoft Copilot say about your childcare centre, fix the facts before the copy. Verify your hours, ages, fees and enquiry steps, correct every listing that is wrong, add real detail to your room pages and check search crawlers can reach them. Then send a test enquiry. Allow four weeks.

Imagine a family looking for toddler care in your suburb. They need Mondays and Fridays, want to understand how settling in works and are trying to fit drop-off around work.

They ask an AI tool for help. Your centre appears.

But the answer gives your old closing time. The page it links to says very little about your toddler room. And when the family follows the link, they cannot work out how to ask about the days they need.

Your centre was found, but that family still has no reason to enquire.

For childcare centres, AI visibility should lead to something useful: a family understanding whether your service could suit their child, then feeling confident enough to contact you.

This guide will help you review the information available about your centre, decide what needs attention and give your team a practical plan to improve it.

Be wary of anyone guaranteeing AI rankings

If someone promises to “rank your centre number one in ChatGPT” or guarantees that AI will recommend you, treat that as a serious red flag.

No marketing agency, consultant or website provider controls which centres an independent AI platform includes in its answers. That includes Enrolment Boost. We cannot guarantee your centre will appear, hold a particular position or be recommended to every family.

Google explicitly states that meeting its requirements and following best practices does not guarantee inclusion. [2] A screenshot of your centre appearing in one answer is evidence of that particular result, not proof of a reliable ranking or future recommendations.

There is legitimate work a provider can do: improve your website, correct public information, address technical access problems and make your content more useful. But a guarantee about an outcome they do not control is a misleading sales promise. Treat it as a reason to question the offer before you pay.

Ask what work will be delivered, how it will be checked and what the provider can actually commit to. At Enrolment Boost, our approach is to follow current best practice, make practical improvements and review the evidence. We can commit to the work. We cannot promise the platform's response.

Start with a check, then work through the findings

If you have not looked at what AI tools say about your centre, start with our founder's guide, Does ChatGPT recommend your childcare centre? Here's how to check.

It explains how to run a short discovery and accuracy check. This article picks up from there: what to change, who should do it and how to review the work.

1. Choose the places you want to fill

Before reviewing your website or searching for your centre, write down one enrolment priority.

For example:

We want enquiries from families seeking care for children aged two to three, particularly on Mondays and Fridays.

This gives the exercise a purpose. You can now ask whether a family with those needs would find the information required to take the next step.

Think beyond the age group. What does your team hear when families call or visit? Perhaps they ask about sleep routines, nappies, meals, orientation visits or collecting a child after work.

Choose three real questions. For a toddler room, they might be:

• Which childcare centres in [suburb, state] care for two-year-olds?
• How do childcare centres in [suburb, state] help toddlers settle in?
• Where can I ask about toddler care on Mondays and Fridays in [suburb, state]?

Use these to review both AI answers and your own website. An AI response is never confirmation that a particular place is available. Your team still needs to check the child's age, requested days and intended start date.

2. Agree on the facts before changing the copy

It is difficult to correct public information if your team has not first agreed on the current details.

Create one centre facts sheet. Have the centre manager verify it, and record the date checked.

Use this sheet to check your website, Google Business Profile and relevant directory listings. For a group, complete it for each location. A program offered at one service should not accidentally become a promise across every centre.

This sheet is an internal reference. It does not need to include children's information or individual family records.

3. Find the source of each problem

When an AI answer gets something wrong, follow the citations before deciding what to change.

Suppose it says your service closes at 6 pm, but you close at 6:30 pm. The cited source might be an old directory listing. Updating your homepage alone would leave that listing incorrect.

Open the source, compare it with your verified facts and record the action needed. OpenAI itself cautions that search answers and citations can be inaccurate or outdated, so a citation should be checked rather than accepted at face value. [1]

Prioritise problems that could stop a suitable family enquiring. Incorrect ages, opening hours or contact details deserve attention before minor wording preferences.

4. Turn broad claims into useful answers

Your educators may have an excellent way of helping children settle. If the website only says “a nurturing environment”, families have very little to assess.

The following are illustrative examples. Only use details that reflect your service's actual practices.

Ask your centre manager to describe what happens in practice. What would a family see? Who would they speak with? What happens before the first day?

Those answers give you material for useful website content. Keep important information on the relevant centre or room page, where families can find it easily. Do not bury every answer in a long, disconnected FAQ.

For fees, explain your inclusions and how families obtain current pricing. Where useful, link to a customised CCS calculator as an estimate tool, alongside a clear way to ask your team about their circumstances. An estimate should not be presented as a confirmed fee or entitlement.

5. Give your website provider a specific brief

“Make us show up in AI” is too vague to be a useful task.

Send the affected pages, the issues you found and the approved information. Ask your provider to confirm that important pages can be accessed and understood by search systems.

Google says its existing SEO fundamentals apply to AI Overviews and AI Mode. It does not require special AI files or special schema markup. Relevant checks include indexing eligibility, readable page text, internal links and accurate business information. These do not guarantee inclusion. [2]

For ChatGPT search, OpenAI identifies OAI-SearchBot as its search crawler. Its controls are separate from GPTBot, which relates to model training. Your provider should check search access specifically rather than treating all AI crawlers as one setting. [3]

You can send this:

Please review the centre and room pages listed below. We have attached the verified information and the issues we found. Please check Google indexing eligibility, access for OpenAI's search crawler, and whether the important information is available as readable page text. Please also check that families can reach these pages through our navigation or relevant internal links. Implement the approved corrections and confirm what changed, which URLs were affected and whether anything remains unresolved. Please keep any search-access recommendations separate from model-training permissions.

6. Can a family get from the answer to an enquiry?

Open the page a family would reach from the answer. Use your phone.

Can you find the relevant age group, understand the service and see how to ask about care? Does the tour button work? Does the form make it clear what happens after submission?

Arrange a labelled test enquiry with your team. Check that it reaches the correct person and that someone knows who is responsible for responding.

For a toddler enquiry, collecting the child's age, preferred days and intended start date can help the team have a useful first conversation. Keep the initial form proportionate; detailed enrolment paperwork can come later.

If your page asks families to “book a tour”, be clear whether they are selecting a confirmed appointment or requesting a time that your team will confirm.

A worked example: from an incorrect answer to a completed task

Imagine a centre seeking toddler enquiries for Mondays and Fridays. The team runs a small set of searches and follows the links families would see.

They identify three issues:

• One answer gives an outdated closing time, citing a directory that still has the old hours.
• The toddler page gives no explanation of orientation or settling in.
• The enquiry form does not ask which days the family needs.

They turn those findings into a plan:

At the next review, they check the sources and repeat the original questions. If the listing is corrected but an AI answer remains wrong, they record that separately.

The completed website and listing changes are evidence of work done. They are not proof of improved AI visibility or additional enrolments. Those outcomes need their own observations over time.

Your 30-day improvement plan

Use this as a work schedule. It is not a promise that search results will change within 30 days.

Keep the review manageable. One centre, one enrolment priority and a handful of useful questions are enough to begin.

Measure three things separately

Work completed: Which facts were corrected? Which pages improved? Did the enquiry test work?

What the checks show: Was the centre mentioned? Was the information accurate? Which sources appeared? Keep the question, tool, date and relevant settings with each observation. Two platform reports can help. Google's Search Console added generative AI performance reports in 2026, showing when your pages appear in its AI features. Bing Webmaster Tools has an AI Performance report in public preview showing citations in Microsoft Copilot. Both show visibility, not enquiries.

What families do: Ask how they heard about you and what helped them decide to enquire. Where they mention an AI tool, record it and follow the enquiry through tours, enrolments and starts.

Avoid turning these into one “AI readiness score”. A checklist measures whether tasks have been completed. It cannot reliably tell you the likelihood of a recommendation.

Search checks are snapshots. ChatGPT can use location and saved memories when searching, so a fresh conversation does not necessarily remove account-specific context. A temporary chat gives a cleaner test. Keep your method consistent and note the conditions. [1]

Family feedback is also imperfect attribution. Someone might hear about you from a friend, ask an AI tool a question and then visit your website. Record everything they mention, even if it is a friend's tip followed by a ChatGPT search.

Start with one improvement your team can finish

Choose the places you want to fill. Check the information those families need. Pick the most important issue, give it an owner and agree on what “finished” looks like.

Your next useful task might be correcting a closing time, explaining orientation properly or fixing an enquiry form. Each should make it easier for a family to understand your centre and contact the right person.

If you would like help deciding where to focus, take the free Occupancy Health Audit. It shows where families drop out between finding your centre and starting, with suggested actions for your team.