Loading...
Sector · HR and people teams
AI for HR, with decisions about people left to people
We build AI for HR teams that drafts, answers and summarises, while a named person makes every decision about hiring, discipline, dismissal and pay.
In focus →
Live routingHover to pause
Incoming
Checking…
AI handles0
A person decides0
Bosses uncomfortable letting AI dismiss underperformers0.0%Net figure where performance criteria were clear. CIPD survey of 814 UK bosses.
Source: CIPD ↗Where the line sits
HR runs on volume, and on judgement
Most HR teams spend their week on the same questions, the same letters and the same case files.
That volume is where AI for HR earns its keep.
It can draft a policy, answer a holiday query from your handbook and turn a long grievance file into a clear summary.
What it should not do is decide what happens to a person’s job, pay or career.
We design HR AI tools around that line, so your team gets time back and your people keep a human decision-maker.
What the AI handles
- First drafts of policies, letters and contracts from your templates
- Answers to routine questions, sourced from your own handbook
- Summaries of case files, meeting notes and investigation interviews
- Onboarding packs, checklists and reminders
- Tidying and tagging HR data for reporting
What stays with a person
- Hiring, promotion and redundancy selection
- Disciplinary and grievance outcomes
- Dismissal, in every case
- Pay, bonus and benefits decisions
- Reasonable adjustments and sensitive welfare conversations
Where it helps
AI for HR teams starts with the routine work
The best first uses of AI for HR are the jobs your team already finds repetitive.
Policy drafting, handbook questions, template letters and case summaries all fit that description.
The CIPD’s practical guide for HR professionals covers the same ground, from shaping an AI use policy to applying AI across the employee lifecycle.
We connect the AI to your own policies, so answers come from your handbook rather than the open internet.
Every answer can show its source, and anything outside the handbook is passed to a named HR adviser.
If you do not yet have rules for staff using AI, start with our AI policy template.
- Handbook assistant for managers and staff
- Policy and letter drafting from your templates
- Case file and interview summaries
- Onboarding and offboarding checklists
85%
AI acts alone
130
65% of 200 cases
Goes to a person
70
35% of 200 cases
Illustrative batch with a typical confidence spread. Your real split comes from a pilot on your own work. Some decisions go to a person whatever the score.
The line we hold
Decisions about people stay with a person
UK data protection law puts extra safeguards around solely automated decisions with legal or similarly significant effects.
The ICO’s guidance on automated decisions about workers gives cutting pay after an automated performance flag as an example of getting this wrong.
It also says a human reviewer needs the authority and competence to go against the system’s recommendation.
Those rules were changed by the Data (Use and Access) Act, and the ICO is reviewing its guidance, so check the current position before you build.
Discrimination risk sits alongside this, because the Equality Act 2010 applies whether a person or a tool applied the rule.
So we build AI for HR that recommends and summarises, and we route every decision about a person to someone who can overrule the machine.
AI drafts
Check
Person decides
Learns
01 / 04
Rollout
Bring your team and your workforce with you
AI tools that touch employee data work better when staff know what they do and why.
The ICO expects employers to seek the views of workers or their representatives before introducing monitoring, unless there is a good reason not to.
We help you run a data protection impact assessment and explain the tool in plain English to staff.
Your HR team then gets hands-on training on reviewing AI output, spotting errors and escalating.
Our AI literacy training covers the basics for managers who will use the tools day to day.
AI tools that touch employee data work better when staff know what they do and why.
How it works
Four steps, one owner
Each system ships with a named person accountable for what it does.
01
01
Map your HR workload
We sit with your team, list the recurring tasks and mark which ones affect a person’s job, pay or rights.
02
02
Draw the decision line
We agree which tasks the AI drafts, which it answers, and which always go to a named person.
03
03
Build and connect
We connect the AI to your handbook and templates, with escalation routes and an audit trail.
04
04
Train, launch and review
We train your HR team, launch with a small group and review the escalations and errors together.
Questions
Asked and answered
Further reading
From the source
CIPDAI use in the workplace: practical advice for HR professionals ↗Shaping an AI use policy and applying AI across the employee lifecycle.CIPDUsing AI responsibly in people management ↗Survey of 814 UK bosses on which people tasks they would hand to AI.Information Commissioner’s OfficeRights related to automated decision-making including profiling ↗When UK GDPR requires a human to be able to step in.Information Commissioner’s OfficeWhat do we need to do if we use monitoring tools that use solely automated processes? ↗What meaningful human involvement in decisions about workers looks like.legislation.gov.ukEquality Act 2010 ↗Direct and indirect discrimination, and the duty to make adjustments.
Keep reading
Related pages
SectorsAI for recruitment →AI for recruitment, with a recruiter behind every noGovernanceAI policy template →An AI policy template your team will actually followTrainingAI literacy training →AI literacy training for everyone who touches AI at workTrainingAI governance course →An AI governance course for the people who design the controls
Next step
Tell us the work. We’ll show you where the line goes.
A 30-minute call. We map one workflow, what the AI could take, and where a person must stay in the loop.