Founding partner offer: First 20 providers get a 3-year price lock. Book your demo today. · See pricing →
AI built for supported accommodation

AI that does the work you don't have time for.

Auto-drafted commissioner reports. Weekly financial briefings written in plain English. Candidate risk scoring. Variance anomaly detection. All inside TIFA Connect. All anonymised before any external call. All Zero Data Retention.

← Back to TIFA Connect home
The shape of AI in TIFA Connect

An operational layer. Not a chatbot.

Most "AI in care software" is a chatbot bolted onto a settings page. TIFA Connect treats AI as an operational layer. It writes the things you'd otherwise spend hours on, then hands you a draft to approve. It flags the things you'd otherwise miss. It does not replace clinical or safeguarding judgment. It does the work that doesn't require judgment, so your team can spend their time on the work that does.

Weekly financial briefing

Your finance week, written in plain English. Every Monday.

Every Monday morning your finance dashboard auto-generates a plain-English briefing summarising the week. Revenue, costs, margin, variance against the prior week, agency spend, voids, EBITDA, cashflow runway. One paragraph. Approved with one click. Sent to your board, your accountant, or your senior team.

  • Live data: generated from your live finance data, not a static template.
  • Your voice: editable before send, so the final voice is yours.
  • Same format, every week: trend reading is instant.
Commissioner report drafting

The narrative writes itself. You review and send.

The narrative section of the monthly commissioner report writes itself from your daily logs, incidents, key working sessions, and welfare checks. You get a first draft, in your provider voice, ready to review. Welsh Reg 80 reports auto-generate end to end. English Reg 45 reports follow the same pattern.

  • Welsh Reg 80: narrative auto-populated.
  • English Reg 45: supported.
  • LA-specific tone: configurable per commissioner.
  • Source citations: every claim links to the underlying log entry so the reviewer can verify in one click.
Candidate scoring and document extraction

The recruiter sees the work, not the paper-handling.

Inside the ATS, every candidate is automatically scored for placement fit, every reference letter is parsed for risk signals, every Right-to-Work document is extracted and validated, every DBS update is matched against the DBS Update Service. The recruiter sees the work, not the paper-handling.

  • Right-to-Work extraction: passport, BRP, share-code support.
  • DBS Update Service: auto-checking.
  • Reference parsing: risk-signal flagging on every reference letter.
  • Placement-fit scoring: tuned to your service profile.
Variance and anomaly detection

The system finds it. The system asks you a question.

The finance module watches your week-on-week metrics and flags anything unusual. Unusual staff cost on a property. A YP whose hours suddenly doubled. Revenue against an LA that's drifted. You don't have to find it. The system finds it and asks you a question.

  • Weekly variance scoring: per property, per LA, per YP.
  • Anomaly flagging: staff cost, hours, sleep-in spend, agency mix.
  • Pattern detection: across placements, early warning before a bad month closes.
  • Plain-English explanations: attached to every flag.
The security model

Not optional. Not configurable away. Every AI feature runs under the same three rules.

Every AI feature in TIFA Connect runs under the same three rules. No exceptions.

100%

of personally identifying information is removed before any external API call. Names, addresses, dates of birth, NHS numbers, social work case IDs. None of it leaves the platform.

Zero

Data Retention. Every call to Anthropic goes with the Zero Data Retention header set. Your inputs are not stored. Your outputs are not used for training. Confirmable in writing.

UK

data residency. The platform runs on Supabase in EU West (Ireland) with UK GDPR data residency. ICO registered. Data Processing Agreement available on request.

What you control

AI is an option, not a requirement.

AI features are configurable per role and per module. Your CFO can switch off the weekly briefing if they prefer to write it themselves. Your registered manager can require human approval on every AI-generated commissioner report before send. Your safeguarding lead can require human approval on every AI-flagged anomaly before it gets actioned. AI is an option, not a requirement, and the audit log records every approval, override, and edit.

See the AI in action.

Book a 30-minute demo. We'll walk you through every AI surface in the platform on a live tenant.