Put Claude to work on your Proclaim cases.
Proclaim holds years of documents, notes and history that your team re-reads by hand. We connect it to Claude, Anthropic's AI model, so the reading, summarising and first drafts take seconds, and a person still decides what goes back into the case.
Work your fee earners do by hand today.
Each one is built to your case types and your rules. Document summaries already run across twenty-five years of Proclaim cases in CaseViewer.
Summarise any document or history entry
A summary, the key points and why it may matter, beside the original. Running today in CaseViewer across twenty-five years of Proclaim cases.
Read incoming post and email
Work out what arrived, match it to the right case, pull out the dates and figures, and queue it for a person to confirm.
First drafts from the file
Letters, attendance notes and case reviews drafted from the case history and your templates into a review queue, never straight out of the door.
Ask the case a question
Plain-English questions over one case or a whole caseload, answered with a link to the entry the answer came from.
File reviews and checks
Missing documents, approaching deadlines and steps overdue for their stage, flagged in a list for supervisors.
Structured data from unstructured files
Medical reports, offers and correspondence turned into fields Proclaim can store, so nobody re-keys them.
Proclaim API integration, with AI in the middle.
We use routes into Proclaim that are already proven in production, so the AI work sits on a join that is known to hold. The same service can join Proclaim to your website, portals, payment providers and the OIC portal: software integration.
Proclaim calls out
Proclaim's own workflow calls a web address when a case reaches a stage. We already run integrations triggered this way, for review requests and client portal sign-ups.
Results come back as data
The service writes structured XML into a folder that Proclaim imports on a schedule, the method behind eight of our Proclaim integrations. Nothing is installed on the Proclaim side.
Reading the archive
For work across old cases, a one-way export gives the AI the history without touching live Proclaim. This is how CaseViewer works.
Figures from your data, decisions from your people.
The AI reads and suggests; a person approves before anything is written back to a case. Answers are forced into a fixed structure and checked in code, so a case reference or figure the model invents is rejected, not saved. Every call is logged with what was sent and what came back. How we use AI.
What people ask first.
Can Proclaim connect to AI?
Yes. Proclaim can call a web address from its own workflow when a case reaches a stage, and it imports structured files that arrive in a folder. Between them they cover most AI work: Proclaim says when something happens, and the results come back as data Proclaim imports. We use whichever routes your version and hosting support, and tell you which before you commit.
Is our client data sent to the AI provider?
Only the part a task needs, through business API terms under which the provider does not train its models on your data. The provider acts as a processor, and every call is logged with what was sent and what came back. A firm can use its own API key so the account and the bill are its own.
Can the AI change a case on its own?
Not unless you decide it should, for a narrow task with fixed rules. By default the AI reads and suggests, a person approves before anything is written back to Proclaim, and every write is logged.
Why Claude?
We build with Anthropic's Claude every day and run it in systems already in production, including our own platform. It follows structured instructions reliably, which is what case work needs. Where another model suits a task better we use that instead: the document summaries in CaseViewer run on OpenAI.
Tell us which part of the file your team re-reads most.
A short call and a straight answer on what AI would change, and what it wouldn't.
0800 246 1101 · hello@prolink.io