Start with your highest-volume or highest-risk processes: hiring, onboarding, contractual changes, leavers and performance cycles. Map how the work actually happens rather than how the policy says it should, find where handoffs break and manual effort concentrates, then fix a small number of high-impact steps. Depth on five processes beats a shallow pass across thirty.
Your team is not slow. The process is
The signal is familiar: a capable HR team working late, a hiring manager chasing an offer for the third time, and a leaver who still has system access a fortnight after their last day. Nobody in the chain is underperforming. The work is simply passing through too many hands, with too little clarity about who decides what.
Process debt accumulates quietly. Each workaround made sense when it was introduced. Together they add up to a function that spends its week on coordination rather than judgement.
Pick five processes, not thirty
Full-function process mapping projects almost always stall. They produce a beautiful document and no change. A better opening move is to choose the five processes that carry the most volume or the most risk, and go deep on those.
For most scaling businesses that means hiring and offer, onboarding and first ninety days, contractual and pay changes, leavers and offboarding, and the performance or review cycle. Between them they account for the bulk of HR transactional load and nearly all of the compliance exposure.
Map the work as it actually happens
The gap between the documented process and the real one is where the friction lives. Walk each process with the people who do it, capture every touchpoint, and count the handoffs, the re-keying, the chase emails and the approvals that nobody can justify.
Three questions expose most of it: who touches this and why, where does it wait, and what breaks when the person who normally does it is away.
- Count handoffs, not steps.
- Record where work waits, not just where it moves.
- Note every place data is entered twice.
- Ask who is accountable for the outcome, not the task.
Prioritise by friction removed, not by effort
Once mapped, resist the urge to fix everything. Score each opportunity on volume, effort released and risk reduced, then take the top three. A small number of well-chosen changes will free more capacity than a comprehensive programme that never finishes.
Typical high-yield fixes: removing an approval that has never once been refused, letting managers self-serve a change with validation, consolidating three intake channels into one, and closing the leaver loop with a single triggered workflow.
Redesign before you automate
Automating a broken process makes it faster and no better. The order matters: understand the work, redesign it, then use the technology you already own to carry it. That sequence is the difference between an automation that releases a day a week and one that adds another system to maintain.
Automating a broken process just gets you to the wrong answer faster.
What to do next
- Choose five high-volume or high-risk processes.
- Map the real workflow with the people who run it.
- Score opportunities by friction removed and risk reduced.
- Redesign the top three, then automate what remains.
If you would like an objective read on where your own function stands, take the free work design assessment, or book a diagnostic call. More questions like this are answered on our HR work design FAQs page.
Related questions
What does an HR process audit cover?
A focused audit covers how work actually flows through your highest-volume processes, where handoffs and approvals create delay, how much manual re-keying is involved, what your systems support natively, and where risk sits. The output should be a prioritised list of changes, not a theoretical process library.
How do we identify bottlenecks in our HR workflows?
Follow the work rather than the org chart. Count handoffs and waiting time in each process, look for steps where data is entered more than once, and identify tasks that stop entirely when one person is unavailable. Those three signals locate almost every bottleneck.