Digital HR: improve the service before choosing the system

The core idea

Digital HR applies technology to people processes, information and services. Its value depends on whether it improves the work, not on how many processes move online. Armstrong and Taylor connect digital tools with evidence and analytical capability. Automating a confusing process can increase the speed and scale of its mistakes.

Using it in practice

Define the problem in the user’s language and map the current journey. Remove unnecessary steps before specifying software. Agree data definitions, access roles and an accountable process owner. Test with employees who have different devices, accessibility needs and working patterns. Run a limited pilot with a manual recovery route, then compare completion, accuracy and user effort with the previous process.

What to watch for

Technology claims need evidence in the intended setting. A vendor’s accuracy figure may not describe your workforce or decision. Personal data, monitoring and automated decisions need separate legal and governance assessment; the 2023 handbook cannot establish the current rules. ICO employment guidance should be checked for its scope and update notices.

Write acceptance criteria before buying

Specify what a user must be able to do, what data may be used and how failures will be detected. Include export, correction, accessibility and audit requirements. A demonstration with ideal data is not evidence that the system will work with local exceptions.

Keep responsibility visible

Name the person who can override an error, investigate a complaint and stop an unsafe process. A human approval button is not meaningful oversight if the reviewer lacks time, information or authority to challenge the output.

Plan for exit as well as launch

Test a usable data export and document key definitions. Agree ownership of configuration and integrations. Review access when roles change and retire duplicate spreadsheets deliberately, with retention needs considered rather than indiscriminate deletion.

Design around the complete employee task

If the task is changing working hours, consider how an employee understands the options, submits a request, receives a decision and checks that systems reflect it. A polished form solves only part of that journey. Include users with limited digital access, assistive technology or different working patterns. Document where human judgement is required and what happens when the standard route does not fit. Keep the service understandable without requiring knowledge of the organisation’s internal HR structure.

Treat automation as a change to accountability

Identify the data used, the rule applied and who is responsible when the result is wrong. A notification, recommendation and decision are different functions. Test exceptions and provide a workable correction route before expanding use. Do not assume that a vendor’s confidence score establishes accuracy or fairness in your workforce. Data protection, employment and equality requirements need current assessment for the specific use; a technical feature does not settle those questions.

Evaluate burden that moves elsewhere

A self-service change may reduce HR processing while increasing manager effort or employee confusion. Compare the whole process, including unsuccessful attempts, repeat contact and manual workarounds. Khan and Millner’s emphasis on useful questions helps frame evaluation: did the change make the task easier and more reliable? Start with a contained pilot, review errors and access barriers, and retain support for people who cannot complete the digital route. Efficiency is meaningful only when the service still works.

Related terms