Claude AI training for hr & people teams in Singapore
Claude AI training for hr teams in Singapore: one hands-on day, onsite, on your own files. Recruitment preparation. Onboarding and policy. People reporting.
Claude Cowork for hr & people teams
HR sits on long documents: handbooks, employment agreements, policy sets, investigation files. This programme uses Claude for careful reading of that material, with the reasoning visible enough to check.
Where the work happens
- Long documents
- Contract review
- Structured extraction
- Multi-source synthesis
This programme is built around long, dense documents — the contracts, reports and policy sets where the work is reading carefully rather than writing quickly.
Who this programme is for
HR, talent and people operations teams. Knowledge, legal and research teams.
- HR business partners who answer the same policy questions every week
- Talent acquisition teams preparing role profiles, interview questions and assessment rubrics
- People operations teams running onboarding and employee communications
- HR leaders deciding where AI may support people decisions and where it must not
Not the right fit
- Teams looking for a vendor demo rather than hands-on practice
- Anyone expecting AI to screen out, rank or decide on candidates or employees without human judgement
- HRIS implementation or system integration projects — this is applied assistant training
What the day covers
Module 1: Recruitment preparation
Employment agreements and policy sets are analysed into structured summaries with the governing clause named.
Exercise. Draft role profiles, structured interview questions and competency rubrics for human-led assessment.
What you leave with. A reusable hiring preparation pack.
Module 2: Onboarding and policy
Onboarding and policy answers are grounded in the handbook, with the exact source passage cited.
Exercise. Build an onboarding plan and draft policy answers grounded in approved documents.
What you leave with. An onboarding checklist and source-backed policy response template.
Module 3: People reporting
Long-form feedback is synthesised into themes, with the omissions stated as plainly as the findings.
Exercise. Summarise anonymised feedback and prepare workforce reporting with explicit checks for omissions and bias.
What you leave with. A reviewed people report and quality checklist.
A worked example: Employment agreement comparison
One exercise from the day, as it runs on your own material in Claude Cowork.
- What goes in
- Your current employment agreement template and three older versions still in force.
- What is asked
- Compare notice periods, restraint clauses, leave entitlements and benefits across all four, naming the governing clause each time.
- What comes out
- A comparison table showing which staff are on which terms.
- What the reviewer checks
- HR confirms each difference against the signed originals before anyone is told their terms differ.
Singapore rules the exercises respect
HR material is the most sensitive data most organisations hold, and Singapore law sets clear limits on how it is used.
- Employee personal data is covered by the PDPA, including purpose limitation: data collected for payroll is not automatically available for an AI exercise.
- Hiring content is reviewed against the Tripartite Guidelines on Fair Employment Practices and the Workplace Fairness Act passed in 2025.
- Survey and feedback analysis is aggregated so that small teams cannot be identified.
- Decisions about individuals stay with people. The programme covers preparation and review, not automated screening.
What this day does not cover
Scope is agreed before the session so the day stays hands-on.
- Coding, scripting or developer workflows, even though the tool is capable of them. This is a day on document and analysis work.
- API integration or automated pipelines.
- Legal advice. We cover how to review a document carefully and what to escalate, not what position to take on it.
Before the session
- A licensed Claude Cowork seat for each participant
- Working knowledge of your own HR processes, handbook and policy library
- Anonymised or sample working files — a role description, a policy extract, an onboarding checklist
- No coding background required
Format and schedule
- Delivery
- On site in Singapore or remote, delivered live. Cohorts are kept small enough that every participant works on their own files rather than watching a screen share.
- Duration
- Typically one full day, or two half-days scheduled around hiring and review cycles. Module order is adjusted to your calendar and tool stack.
- How each module runs
- Each module follows the same pattern: a short walkthrough of the technique, a hands-on exercise on your own data, then a review of what the assistant got wrong and how to catch it.
- Afterwards
- Participants leave with the prompt patterns, templates and review checklists produced during the session, so the work continues after the training day.
How your data is handled
- Exercises run inside your own tenant and your own licences, so your data does not leave your environment for the training.
- Files shared with the trainer in advance are anonymised extracts; employee records, candidate data and case files stay inside your own tenant and licences.
- Every module includes verification: checking output against the source, confirming what the assistant left out, and marking what still needs a human read.
- Decisions about candidates and employees stay with your people. The programme teaches AI-assisted preparation and review, never automated hiring or employment decisions.
Who runs the session
Abhijit Ghosh
Founder, Eagentix · Data & Generative AI Practitioner · Singapore
The programme is delivered by Abhijit Ghosh, who has spent 25 years in data and AI, including roles at Accenture, IBM, Deloitte, Cloudera, Tibco and Wipro. Sessions are run by the practitioner, not handed to a junior facilitator.
Questions about Claude Cowork for hr & people
Can it read our employment agreements and handbook?
Yes, and the exercise is structured summaries with the governing clause named, so a reviewer can check the answer rather than trust it. Confidentiality boundaries are agreed before anything is shared.
Why choose Claude for this work rather than the assistant bundled with our office suite?
Because the bottleneck here is long, dense reading rather than quick drafting. The programme is built around documents that take a person an afternoon, and around making the reasoning visible enough that a reviewer can check it clause by clause.
How long can the documents be?
Long enough that the constraint in practice is your review capacity, not the tool. We work with your real documents on the day, and part of the exercise is deciding what still needs a full human read.
How does the programme handle PDPA for employee data?
Exercises use anonymised or synthetic records unless your data protection officer has cleared real data for the purpose. Purpose limitation is covered explicitly, because data collected for payroll or performance is not automatically available for an AI exercise.
Can the assistant screen candidates?
The programme does not teach automated screening. It covers drafting role profiles and interview guides and reviewing them against fair employment guidelines. Decisions about candidates stay with people.