AI agents for medical research institutes and CRCs

Your agreements, funder rules and SOPs, answered in plain English

We build AI agents inside your Microsoft 365 that answer questions from your own documents, check grant applications before they go out and sort the shared inbox. Small teams carrying big rulebooks get time back. Your people stay in charge.

Book a discovery call

A research operations manager checks a referenced answer on her laptop, with a printed research agreement beside her and a laboratory behind the glass.

Where research support hours go

Every application counts.

The 2025 NHMRC Ideas Grants round funded 8.1% of applications.1 Small grants teams need every one to be eligible, complete and on time.

Reporting never stops.

CRC grant agreements set out milestones, reporting and payment schedules, and payments depend on satisfactory progress.2 The same few people prepare the reports and answer the questions.

The AI rules changed in April.

ARC and NHMRC policies from 28 April 2026 let applicants use generative AI as a support tool but keep them responsible for accuracy.3 Researchers want a straight answer on what's allowed.

Policy and Domain Guru

Answers from your own rulebook, with the source attached

The Policy and Domain Guru answers questions from the documents your institute or CRC already runs on: grant and participants agreements, funder rules, ethics and governance procedures, and SOPs. Every answer shows where it came from, so your team can check it.

  • Lookups that meant digging through agreements and SOPs become one question and a referenced answer.
  • Understands the question in context, not just the keywords, and fits the answer to the task in front of you.
  • Every answer traces to the clause, page or section it came from.
  • Explore scenarios with your own documents, such as a delayed milestone or a change of participant, before you take them to the board.
  • One consistent answer for researchers, program staff and partners.
  • Your experts stop answering the same question twice and keep their time for the calls that need them.
A CRC program team reviews milestone plans around a meeting table.

Answers inform. People decide.The agent finds and explains what your documents say. Legal, ethics and funding decisions stay with the people accountable for them.

How a question gets answered

  1. ArrivesA researcher or staff member asks a question
  2. AgentYour agreements, rules and SOPs searched
  3. AgentAn answer prepared with the clause attached
  4. Your teamYour staff check the source and decide what to do

Two more ways to get time back

Application Analyzer

Grant applications checked before they go out

  • Checks each application against scheme rules and your internal sign-off requirements, and lists what's missing or doesn't fit.
  • Reviews a whole round in one pass, not one at a time.
  • Gives your grants team a short review note on each application.
  • Your grants team decides every fix and every sign-off.

Email Triage Agent

One shared inbox, sorted

  • Sorts grant, ethics, contract and researcher questions and routes each one to the right person.
  • Flags deadline-critical emails first.
  • Drafts replies from your own guidance for staff to approve.
  • A dashboard of volumes, turnaround and what's waiting.

How we build it with you

  1. Map the work

    On the discovery call we look at where your team's hours go and which emails, questions or applications repeat.

  2. Build inside your Microsoft 365

    We build the agent in your Microsoft 365 environment and agree up front what it can see and who can use it.

  3. Check everything in a pilot

    Your team reviews every output during the pilot. Nothing goes out without a person.

  4. Measure and tune

    We track what the agent gets right and where it needs a person, then tune it with your team.

Built by people who know research administration

Our founder spent two years as a product manager for university research administration software, before starting 25 Hours AI.

Clients using our agents report less manual sorting, less time processing by hand and faster turnaround.

Rules and dates to plan around

28 April 2026
ARC and NHMRC generative AI policies took effect. Applicants may use AI as a support tool and remain responsible for accuracy.3
10 December 2026
Privacy policies must explain decisions that software helps make about people. A person reviewing the output doesn't take the tool out of scope.4
2027
CRC Grants Round 28 is planned as an AI Accelerator round of about $50 million, opening in 2027.2

Questions we hear

Will this replace our research support staff?

No. The agents take on looking things up, checking and sorting. Your people review every output and make every decision.

Our agreements are confidential. Where does the information go?

We build inside your Microsoft 365 environment and connect to the apps your team already uses. Before we build, we agree what the agent can see and who can use it.

Does it give legal advice on our agreements?

No. It shows what your documents say and where. Legal and contract decisions stay with your legal and contracts team.

What happens when it gets something wrong?

A person checks every output before it's used. During the pilot we track what the agent gets right and where it needs help, then tune it with your team.

How long does it take to set up?

It depends on your inboxes, documents and systems. We map that on the discovery call and give you a plain answer, including whether an agent is the right fix at all.

Put AI to work.

Find out what your team could stop looking up

Book a discovery call. We'll look at your agreements, funder rules and the questions that repeat, then tell you plainly whether an agent will help.

Book a discovery call

Bookings are made through Microsoft Bookings. Read how we handle your information in our privacy policy.