Applications & AI screening

Every application arrives with the household, their documents and an AI review already prepared — a plain-language summary, an affordability check, and a document-by-document cross-check of what the applicant stated against what their payslips and statements actually show. The AI never decides; it points a reviewer straight at what needs a look.

1. 1

   Every application for the company in one list, filterable by status and searchable by applicant or property.

   

2. 2

   Filtered to the applications currently under review.

   

3. 3

   The application header shows the property, the unit and the household at a glance, with a tab for each occupant and for documents, emails, history and the AI review.

   

4. 4

   A complete, internally consistent file: the recommended next step, a plain-language summary, and the affordability check — all within the landlord guideline, on a file where every document cross-check comes back consistent.

   

5. 5

   Rent-to-income is computed from the stated income and the landlord threshold, with scenarios around the current rent and a note of what the payslips actually evidence — worked out without the AI, so it shows instantly.

   

6. 6

   The AI reviews every uploaded file: a status dot and a one-line finding per document, grouped by occupant. It reads a payslip and a bank statement, checks a landlord reference — and where a document is in another language it detects that, translates the key content, and still reconciles the figures.

   

7. 7

   A check against the RTB tenancy-dispute register and the Irish bankruptcy (CSOL) records. Because both are searched by name, a hit is often a namesake — so the AI triages each match as likely the same person, likely a different person, or unclear, with its reasoning shown. Staff record the outcome.

   

8. 8

   The same review on a thinner file: income stated but not evidenced, rent over the guideline, a stale bank statement and a missing landlord reference — flagged as discrepancies, recommended next step "request documents".

   

9. 9

   Composing a reply: a saved email template — scoped to the company, the client or the property — drops a ready subject and body into the editor, with the applicant’s and property’s details already merged in.

   

10. 10

    Or let the AI draft it: an approve / decline / request-documents intent writes a first draft for the agent to edit and send — a decline draft is barred from citing anything like a protected ground.

    

Every screen above is captured from a real, running copy of Real Enquiries, last refreshed September 2026. They update automatically as the product changes — this isn’t a mockup. All tours.