Managers across adult social care are recording meetings all over the place.
Not as part of a formal AI strategy. Not because someone told them to. They're reaching for whatever recording tool is available — Teams, Word, voice memos on their phones, Zoom — and using them to capture conversations that matter. Supervisions. Initial assessments. Safeguarding discussions. Care plan reviews. Handovers.
They're doing it because it works. It reduces the note-taking burden. It means less reconstructing from memory afterwards. It keeps them more present in the conversation.
They just don't think of it as AI. It's just useful.
You've been here before
Think about how far the sector has already come without really noticing.
Zoom and Teams didn't just move meetings online. They made it possible to record, transcribe, and search conversations that previously existed only in someone's memory and a hastily typed document. Covid forced the sector to rethink everything it had always done, and most managers navigated that transition in a matter of weeks. What felt impossible became standard practice almost overnight.
The managers who said "I'm not a technology person" in 2019 were running Teams supervisions by 2021. The same shift is happening now with AI transcription, just more quietly and more organically. Nobody announced it. People just started finding it useful.
That's how good technology actually spreads. Not through mandates or training programmes. Through people discovering that something makes their job easier and telling their colleagues.
What's already happening
The picture that's emerging is one of widespread, informal AI adoption across social care. Managers recording supervision sessions. Care coordinators using Word transcription for assessment notes. Team leaders capturing handover discussions on their phones.
It's happening at scale. It's largely unmanaged, and most of it is happening with consumer tools that were never designed for the sensitivity of these conversations.
A supervision session contains personal data about a member of staff, potentially including health disclosures, performance concerns, and personal circumstances. A safeguarding discussion contains information about a vulnerable adult that carries the highest level of GDPR protection. An initial assessment contains detailed, sensitive information about someone's care needs and living situation.
When that data gets processed through a consumer transcription app or pasted into a free AI tool, it leaves the organisation's control. There may be no Data Processing Agreement. The data may be used to train the provider's models. There is no audit trail.
The adoption is already there. The governance isn't.
The nuance worth designing for
Something else came up in those conversations that's worth naming. People can feel more comfortable being recorded without the camera, audio and transcript, rather than video. A record of what was said feels fair. Being watched feels different.
That distinction matters for how you design and position a recording tool for care. The transcript is documentation. The video is surveillance. Most managers and staff intuitively understand the difference, even if they couldn't articulate it.
Where 1two.one fits
The sector isn't waiting to be convinced about AI recording. It's already convinced. The question is whether the tools being used were built for this environment.
1two.one was. UK data residency. A clear Data Processing Agreement. No model training on your data. A workflow designed specifically for supervision, not bolted on from another context.
If you're already using Teams or Zoom for supervision, you don't need to change anything. Upload your transcript and 1two.one will scan it, pulling out the themes, the agreed actions, the development points, the things that need following up, mapping them against CQC Quality Statements, and producing a structured record ready for sign-off by both parties.
That's not a disruption to how you work. It's a safer, more compliant version of something you're probably already doing.
The managers who are "wary of AI" are often already using it. They just don't think of it that way. The question isn't whether to use technology in supervision. It's whether the technology you're using was built for the job.
The reason recording works, cognitively speaking, is that it frees managers from doing two jobs at once. Read the research on why you can't take notes and listen at the same time, or read what actually changes when 1two.one is in the room.
If you'd like a plain-English map of what supervision needs to be — with or without AI — download The Supervision Conversation guide.
Or, if you're ready to move from consumer tools to something built for the sensitivity of the conversation, join the waitlist.
Part of a short series on supervision in UK adult social care. See also: Competence and safety to practice don't change when the regulations do, What is supervision in social care?, London boroughs asked what AI could do for social care, and Supervision 2.0: what it actually looks like.
Found this useful?
Share it with your network — help another registered manager save an hour today.