The challenge
A UK care home group records meaningful activity for every resident. After a group session or a one-to-one, staff complete an enrichment chart describing how engaged the resident was, their mood, whether they took part as usual and whether the activity still suits them.
The group wanted AI activity summaries for care homes that would turn those entries into something a busy team could read.
Each entry is rich in detail, and the sheer volume of that detail is the obstacle. Over weeks they show what a resident enjoys and how they are changing, but nobody on a busy floor has time to read them all.
Left unsummarized, a resident's month is scattered across dozens of short records that nobody reads end to end.
What we built: AI activity summaries for care homes
We also scoped and built the enrichment chart itself, a web and mobile module running on React, Node.js, Express.js and MongoDB.
Two ways in. Staff open the chart from the activities schedule for a group session, or from the resident's profile for a one-to-one. Group sessions pre-fill the duration.
Photos and video. Staff can add photos or videos for unscheduled activities.
A 48-hour alert. The system alerts staff if a resident's enrichment chart has not been completed in 48 hours.
Above the entries sits a Generate Summary button. Pick any stretch of dates, and the AI goes through the entries in it to answer a set list of questions:
Which activities did the resident take part in, and how engaged were they overall?
What was their mood, and which activities lifted it most?
Did they engage more with staff or with other residents, and did any new interests appear?
The summary also compares the period with the week before, covering participation, mood shifts and types of activity. It groups engagement as physical, creative, social, cognitive, sensory or outdoors and, where the care plan sets activity goals, reports progress toward them.
How it works day to day
Take a member of staff preparing for a resident's care plan review. Instead of scrolling through a month of enrichment entries, they choose the date range and generate a summary. It opens in a pop-up and downloads as a PDF.
Every summary closes with a notice that AI wrote it, and that staff must review the charts before any change to the resident's care or care plan. So the staff member reads the summary against the entries behind it before anything changes.
The software shortens the reading. Deciding what the pattern means for the resident stays with staff.
The result
Staff and managers can see, for any period, what a resident took part in, how engaged they were and what lifted their mood, along with how that compares with the previous week.
The 48-hour alert keeps the entries coming, so there is a record to summarize. Each summary is labeled as AI generated and points staff back to the charts. We don't have outcome data for the summaries, so none is quoted here.
What the research says
Care records have gone digital. In the Department of Health and Social Care's 2025 survey of 1,085 adult social care providers in England, 73% used a digital social care record, and national adoption rose from 41% in December 2021 to 80% by July 2025 (DHSC technology survey). Once enrichment charts are typed in rather than written on paper, a summary can be built from them.
Staff bring their own AI if none is provided. Microsoft's 2024 Work Trend Index, a survey of 31,000 knowledge workers in 31 markets, found 75% use generative AI at work and 78% of those users bring their own AI tools (Microsoft Work Trend Index 2024). Building summaries into the group's own system gives staff an approved route, instead of pasting resident notes into a personal chatbot.
Label what the AI wrote. Among the measures the UK Information Commissioner's Office lists for communicating the risk of incorrect output is labeling it as generated by AI (ICO consultation response). The enrichment summaries follow that approach on every page they produce.
Where AI fits next
These are future options, not current features. The system could suggest new activities that match interests a summary picks up. A summary could also go to a resident's named carer each week, ready for them to check before a review.
Planning something similar?
Before you add AI summaries to activity records, ask:
What should staff record after every session, and how will you know when entries are missing?
Which questions should a summary answer, and how do they map to care plan goals?
What warning should appear on every summary, and who reads the charts behind it?
How will you group activities so trends are easy to compare?
The enrichment chart sits within the care and clinical platform for this group and looks at one resident at a time. For a view across sessions, see our AI breakdown of offered versus refused activities; for distress patterns, see AI behavior chart analysis.
Our AI automation page explains how we work, and our healthcare software page lists related projects.
