HealthCentre.Ai A KwickBlocks product

July 2026 · Care homes

Is an AI phone agent safe for a CQC-regulated care home?

Yes, if it stays inside clear boundaries and logs every call and message. The record it keeps turns out to be the part managers value most.

The short answer: yes, if it stays inside clear boundaries and logs every call and message.

An AI agent in a CQC-regulated care home should not deliver care, give clinical advice, or decide who gets a bed. It answers questions from the home's own documents, hands anything clinical or distressed to a named member of staff, and logs every call and message with a timestamp and the document the answer came from.

That last part is what managers do not expect. It also creates a useful record if a family, a complaint, or a regulator later asks what was said.

The three questions a compliance lead will ask

1. Can it do anything clinical? No. It has no authority over care. It will not advise on medication or symptoms, it will not assess anyone, and it will not decide an admission. A clinical question is passed to a named nurse or manager, and logged either way so it cannot be quietly dropped.

2. Where do its answers come from? Your own documents: the fee schedule, what is included and what is extra, the funding routes you accept, the care types you are registered for, the visiting policy, the inspection report. It answers from source, and it can tell you which document it read. If the answer is not in your paperwork, it says it does not know and takes a message.

3. What is the record? Every conversation is logged: who called, when, what they asked, what was said, and the document the answer came from. It is searchable.

How an AI agent logs enquiries and answers

Care homes already have to evidence how they communicate with families, and how enquiries and complaints are handled and resolved. Today that evidence usually lives in three places at once: a notepad by the phone, an inbox, and somebody's memory.

Six months after a family was told the hairdressing was included, none of those three will settle the argument.

A logged conversation will. You have the call, the date, the exact wording, and the page of the fee schedule the answer was read from. The same log tells you which enquiries came in, which were answered, and which turned into residents, which is the operational question most homes cannot currently answer either.

We are not going to claim it inspects well on your behalf. It is not a compliance product and anyone selling it as one is overreaching. It is a communication tool that happens to keep an excellent record, and in a regulated setting that is worth a great deal.

What "grounded" actually means

A general chatbot may give a plausible answer about care home fees even though it has never seen your fee schedule.

A grounded AI agent reads your documents, names the one it used, and refuses when they do not cover the question. That is how the same document gets used on the phone, in chat and by email, and it is why a wrong answer is traceable to a document you can correct rather than a mystery you cannot.

The honest risks

  • Your documents have to be right. If your fee schedule is out of date, the agent will confidently quote an out-of-date fee. Loading the knowledge base is the real work of setting this up.
  • Boundaries need deciding, not defaulting. What it answers, what it books, and where it hands over should be a decision your manager and clinical lead make together, and revisit.
  • It is not a safeguarding system. A distressed or safeguarding call goes to a person, fast. That is the design, but it is your people who respond.

Frequently asked questions

Is AI allowed in a CQC-regulated care home?

There is nothing that prevents a home using software to answer its phone and inbox, as long as care and clinical decisions stay with the people qualified to make them.

Could an AI agent get us into trouble with a regulator?

The risk in any regulated setting is an unrecorded, unaccountable answer. The design is aimed at that: answers come from your documents, clinical questions go to staff, and everything is logged.

What about resident data and GDPR?

An AI agent handles enquiries and communication, not care records. What it holds is what a family told you and what your own published documents say, and every conversation is logged and retrievable.

Can we see and change what an AI agent is allowed to answer?

Yes. You set it, and you can change it whenever you want.

Sources

Hear it answer for your own home

Enter your home's website address and hear HealthCentre answer live, built around your own fees, care types and rooms.

Build your home's agent, live