Categories
News

KEATH pilot – call for applications

Hands typing on computer keyboard closeup, business man or student using laptop at home

This pilot is part of our wider AI in Assessment Pilot for 2025–26. If you haven’t already, we recommend starting with the AI in Assessment Pilot blog to understand the context of the overarching project.  

In this blog, you’ll find out: 

  • About KEATH
  • What the pilot will involve 
  • What’s expected of pilot participants 

About KEATH

KEATH is a generative AI platform designed to support educators in providing feedback on student work, including essays, reports, and dissertations. 

Educators can train their own version of KEATH’s model using as few as ten manually graded essays, making the platform responsive to specific assessment rubrics and institutional practices. This ensures that feedback aligns closely with how grading criteria are interpreted in practice. 

KEATH is built around a human-in-the-loop approach, meaning that educators remain central to the marking process. Users can review and amend feedback generated by the system, and human confirmation is required before marks are finalised. 

Pilot Overview 

The pilot will run from September 2025 to August 2026 and will include 10 institutions (colleges or universities).  

Each participating institution will choose a consecutive 6-month window within the 2025-26 academic year in which to actively use KEATH.

KEATH will be used as a standalone platform as part of this pilot, and so no integrations (i.e. with Virtual Learning Environments) will be needed. 

Accounts will be set up for participating staff, with no accounts required for students at this stage. Single Sign-On (SSO) functionality is not available as part of this pilot.

Who Will Do What? 

The pilot will involve four different types of staff participant: project leads, project sponsors, technical staff and staff users. 

Staff Users 

Staff users will use KEATH to help mark and/or give feedback on student work for at least one module.  

We suggest that around 10 staff users are selected for the pilot. You don’t need to specify who each of the staff users will be at this stage, however, we’d highly recommend you identify the modules or subject areas in which you intend to use KEATH.

Staff users’ key responsibilities will be to: 

  • Attend a one-hour kick-off session (for the whole pilot cohort) 
  • Attend a one-hour onboarding session (institution specific) 
  • Support the project lead in collating past papers and rubrics (which will be used to train the AI model) 
  • Use KEATH to help with marking and/or feedback for at least one module 
  • Attend an evaluation focus group, and complete an end of pilot survey 

Staff users are invited but not required to attend fortnightly vendor check-ins (which may be weekly during busier periods) and monthly community sessions.  

Project Leads 

The project leads will act as internal experts on KEATH. For the purposes of the pilot, KEATH will implement a ‘train the trainers’ model, whereby all staff users will be given initial training and a smaller group of staff (which will be the project leads) will receive more intensive training, so that they can further upskill their colleagues.  

Institutions may select multiple project leads if preferred. The project leads may also act as staff users. 

The project lead’s responsibilities will be to: 

  • Manage the process of getting the Pilot Participation Agreement signed
  • Complete an initial information form (which will include details of staff users for onboarding, along with information regarding how KEATH will be used) 
  • Attend a one-hour kick-off session 
  • Attend a one-hour onboarding session 
  • Attend 1:1 or small group training sessions 
  • Upskill colleagues and troubleshoot basic issues 
  • Provide material needed for the training of KEATH’s model. This is so that the model can be fine-tuned for the context of each module. 
  • Join fortnightly vendor check-ins (which may be weekly during busier periods), complete fortnightly progress updates, and attend monthly community sessions 
  • Coordinate end of pilot focus groups, and distribute evaluation surveys 

Project Sponsors 

Project Sponsors are there to provide strategic support and make sure the pilot runs smoothly. They’ll help by: 

  • Overseeing the signing of the Pilot Participation Agreement
  • Ensuring institutional capacity and engagement 
  • Supporting the Project Lead 
  • Supporting the evaluation activities 

Sponsors should attend the kick-off session. They are invited but not required to attend the other sessions. 

Technical staff 

While KEATH does not require platform-level integrations, participation in the pilot will benefit from engagement with technical staff during onboarding and early use. Technical staff are not expected to engage regularly during the pilot but play a useful role in ensuring a smooth setup and addressing any IT-related blockers quickly and effectively. 

Their responsibilities are likely to include: 

  • Checking that institutional networks do not block access to the KEATH.ai domain 
  • Addressing any firewall or proxy issues that might affect login or platform performance 
  • Verifying browser compatibility and troubleshoot access problems if they arise 

Where KEATH Works Best 

KEATH is designed for a range of text-based assessment tasks, including: 

  • Formative and reflective assessment activities 
  • Essays, literature reviews, and research reports 
  • Lab write-ups, project reflections, and short-answer questions 

It works particularly well in Humanities, Social Sciences, and Business-related disciplines, where in-depth, narrative feedback is particularly valuable.  

KEATH is effective across both undergraduate and postgraduate levels, including masters dissertations and doctoral draft reviews. It is also suitable for use in college settings. 

That said, it is not recommended for: 

  • Highly structured, quantitative assessments (e.g. multiple-choice or numeric problem-sets) 
  • Non-text formats, such as audio, video, or visual media 

It should also be noted that institutional policies may limit how AI can be used within the marking of high-stakes summative assessments. 

As with most AI in assessment tools, KEATH is not designed to replace the professional judgement of an educator. All feedback and grading suggestions should be reviewed—and where necessary, adapted—by a human before use. 

KEATH scales to any number of users, and benefits from larger sample sizes to enhance the accuracy and quality of feedback. 

What to Expect During the Pilot 

Participants will be selected before the end of the 2024-25 academic year.

Selected participants will then be issued a Pilot Participant Agreement to sign before the end of September 2025.

Next, we’ll hold a kick-off webinar in September/October 2025 for the whole pilot cohort.

After the kick-off session, we’ll ask all participants to complete an information form, which will gather details on: which members of staff will participate, in which modules KEATH will be used, key documents (e.g. rubrics and past papers) for the relevant assessments.

The information form will also ask participants to specify the 6-month window in which they intend to use KEATH.

Following these steps, the key stages of the pilot will be as follows. The timing of these stages will depend upon the chosen 6-month window: 

Account Provisioning & Platform Setup. KEATH will create accounts and activate platform access 

Live Onboarding Session. An institution-specific session introducing key platform features and workflows 

Sample Submission & Model Training. Educators will submit assessment samples and rubrics, and  choose between the following training routes: 

  • Custom Request (upload 10+ samples for a 2-week training cycle) 
  • Train-As-You-Go (TAYGo) (feedback improves continuously with each use) 

 Confirmation & Go-Live. Once the model is trained and confirmed, institutions will begin full pilot use 

 Ongoing engagement

  • Project leads will attend fortnightly/weekly vendor sessions, complete fortnightly progress update reports, and attend monthly community sessions 
  • Project leads will be able to request 1:1 sessions for further support from KEATH.
  • KEATH will provide resources and guides to supplement training, and they will be available to address any issues that should arise 

Evaluation 

  • Project leads and Jisc will work together to organise a series of focus groups with staff and students who were involved in the pilot 
  • Project leads will work with Jisc to distribute evaluation surveys 

Setting Up 

Getting started with KEATH is straightforward and does not require VLE integrations. 

To set up: 

  • Project leads provide usernames and emails (these will be collected as part of the information form) 
  • Project leads upload sample assignments and associated rubrics 
  • Choose your model training path:
  • Custom Request: Upload a batch of at least 10 papers; KEATH will use these to fine-tune the model
  • Train-As-You-Go: Grade an initial assignment directly in KEATH; feedback is refined over time

Training takes approximately 2 weeks for Custom Request. TAYGo users receive immediate results that improve iteratively.

Technical requirements are minimal: a modern browser, stable internet, and consistent file formatting are all that’s needed. 

Training & Support 

Live Sessions 

  • A kick-off webinar for all participating institutions
  • A 60-minute onboarding demo for each institution
  • 1:1 sessions for project leads

Asynchronous Resources 

  • Screencasts covering core workflows 
  • A comprehensive onboarding guidebook 
  • Documentation and templates for setup and training 

Ongoing Support 

  • Fortnightly/weekly check-ins for project leads with the project teams at KEATH and Jisc 

Evaluation: April 2026 

All participating institutions will contribute to the pilot evaluation, which includes: 

  • Surveys and group interviews with staff and students
  • Input from Project Leads and Sponsors to help coordinate participation 

The aims of the evaluation are to: 

  • Understand KEATH’s impact on marking and feedback workload 
  • Capture different stakeholder views on AI in assessment 
  • Assess KEATH’s effectiveness and suitability 
  • Help Jisc improve future pilots 

Post-Pilot: What Happens Next? 

There’s no obligation to subscribe after the pilot, but we know many participants want an idea of costs before committing to a pilot.

KEATH’s pricing model includes: 

  • A per-student licensing fee (with discounts at scale)
  • An implementation fee if VLE integrations are pursued post-pilot

The following is an example of how pricing would work out for a university wanting to use KEATH to augment the marking and feedback provision for 5,000 students

  • £20,000 + VAT for 5,000 student licences for one year 
  • £5,000 + VAT for a one-off integration with the university’s VLE (this would not be paid again in subsequent years) 

Summary of Pilot Commitments 

Project Leads 

Before the pilot: 
  • Complete the information form – this will include inputting details of participants, modules involved in the pilots, and examples of rubrics and past submissions for model training 
  • Attend the kick-off webinar
  • Attend a one-hour onboarding session (institution-specific)
  • (Optional) Attend 1:1 or small-group training sessions with the team at KEATH
During the pilot: 
  • Act as internal trainer under a “train the trainer” model 
  • Upskill staff users and troubleshoot basic issues 
  • Join fortnightly vendor-led check-in sessions (weekly during busier periods, i.e. the set up phase)
  • Submit fortnightly progress reports
  • Attend monthly community sessions
  • Request additional 1:1 support sessions as needed 
During evaluation: 
  • Help organise and coordinate staff and student focus groups 
  • Distribute and promote completion of evaluation surveys 

Project Sponsors 

Before the pilot: 
  • Ensure the Pilot Participation Agreement is signed
  • Confirm institutional capacity and readiness to participate
  • Attend the kick-off webinar
During the pilot: 
  • Provide strategic oversight and senior-level support 
  • Support the Project Lead with resourcing and engagement 
During evaluation: 
  • Encourage institutional participation in evaluation activities 

Staff Users 

Before the pilot: 
  • Attend the kick-off webinar
  • Attend the onboarding session for their institution
  • Support the project lead in gathering past papers and rubrics 
During the pilot: 
  • Use KEATH to support marking and/or feedback for at least one module 
  • (Optional) attend fortnightly/weekly vendor check-ins and monthly community sessions 
During evaluation: 
  • Participate in a focus group 
  • Complete an end-of-pilot survey 

Technical Staff 

Before the pilot: 
  • Confirm that institutional networks allow access to the KEATH.ai domain 
  • Address any firewall, proxy, or access-related issues 
  • Ensure participating staff are using compatible browsers 
During the pilot: 
  • Troubleshoot basic platform access/compatibility issues if they arise 

How to Apply 

Applications are open until 20th June 2025. To express your interest, complete the application form linked in the main AI in Assessment Pilot blog 

If you’re not selected for a product-based pilot, you’ll be invited to join our pilot of general-purpose productivity tools.

If your institution would like to apply to the pilot but the submission timeline is unfeasible then please get in touch with us at ai@jisc.ac.uk 

Get in Touch 

For questions relating to the KEATH platform, please get in touch with abby@keath.ai 

For anything else, please reach out to ai@jisc.ac.uk  


Find out more by visiting our Artificial Intelligence page to view publications and resources, join us for events and discover what AI has to offer through our range of interactive online demos.

Join our AI in Education communities to stay up to date and engage with other members.

Get in touch with the team directly at AI@jisc.ac.uk

By Tom Moule

Senior AI Specialist at The National Centre for AI in Tertiary Education

Leave a Reply

Your email address will not be published. Required fields are marked *