People Practice
PART II · BThe HR-Enabled Leader · Where the cycle breaks

The rules on tap — and a clear line to stop at.

This is the move that actually breaks the rescue cycle. Most of what leaders escalate to HR is routine: “what does our policy say, what am I allowed to do here?” Answer that instantly and reliably — with an unmissable stop-line for the matters that genuinely need HR — and the leader self-serves the simple, while HR keeps the serious.

2·A gave the leader the records. 2·B gives them the rules — and, just as importantly, the boundary that keeps everyone safe.

Self-serve the routineThe stop-lineGrounded answersHR keeps the serious
Where the rescue cycle actually breaks

Most escalations were never the hard ones

Picture the queue of questions that lands on HR's desk. A large share are not sensitive judgement calls at all — they are “where do I find it / what does the policy say / am I allowed to do this” questions a confident leader could answer in a minute, if only they had a fast, trustworthy way to check.

Today, with no such way, the safest move a leader can see is “ask HR.” And so the simplest query becomes an escalation, the escalation becomes a dependency, and the dependency keeps the cycle turning. Give the leader the rules on tap, and that whole bottom tier of the queue simply dissolves.

Today · everything routes through HR

The routine question becomes a dependency

  1. Leader hits a simple policy question mid-task
  2. No fast, trusted way to check — so they email HR
  3. They wait; the task stalls; momentum is lost
  4. HR answers something a leader could have self-served
  5. The lesson lands again: “check with HR first”
Enabled · the routine self-serves

The leader acts, and HR keeps the serious

  1. Leader asks the question in plain words, in the moment
  2. Gets a grounded answer from the org's own policy
  3. Acts straight away — no wait, no stall
  4. The stop-line flags anything that genuinely needs HR
  5. HR's queue is now only the matters that need judgement
The stop-line · a built-in boundary

Three lights: what a leader can self-serve, and where they must stop

Enablement without a boundary is just risk. The stop-line makes the boundary visible and consistent, built into every answer. The principle is simple: the clearer the rule, the greener the light; the more it turns on judgement, sensitivity or law, the redder it gets.

Green · self-serve

Clear, rule-based, low-risk

Routine questions with a clear, published answer. The leader gets it instantly and acts.
Typically
  • “How much notice for annual leave?”
  • “What's our process for approving a flexible-work request?”
  • “Which form records a return-to-work conversation?”
AI does: answers from the org's own policy, cites where it found it, and lets the leader proceed.
Amber · proceed with care

Allowed, but log it & check

The leader can act, but the matter carries risk or precedent. Proceed carefully, document it, and consider a quick HR check.
Typically
  • A first informal performance conversation
  • Declining a request that could feel unfair
  • A pattern that's starting to look formal
AI does: gives the grounded answer, prompts a file note, and flags “consider raising with HR if it escalates.”
Red · stop & escalate

Sensitive, contested or high-stakes

The matter needs human HR judgement, specialist advice, or carries real legal exposure. The leader stops and brings HR in.
Typically
  • Suspected misconduct, bullying or a grievance
  • Anything touching dismissal or discipline
  • Disability, discrimination or psychosocial risk
AI does: does not advise the leader through it — it names the stop-line and routes them straight to HR.

The red light is the most important feature, not a limitation. A tool that knows what it must not answer is exactly what makes it safe to let leaders self-serve everything else.

The stop-line in practice

The same leader, three questions, three lights

Watch how one tool handles a green, an amber and a red — grounding the easy ones and refusing the hard one. This is the boundary working in real time.

Leader asks

“One of my team wants to take annual leave next month — how much notice are they meant to give, and can I say no if it clashes with a deadline?”

AI returns
  • The notice period from the org's own leave policy, cited
  • The grounds on which leave can reasonably be declined
  • A suggested fair, plain-English reply to the employee
GREEN · self-serve
Leader asks

“Someone's quality has slipped for a few weeks. I want to raise it informally — what's fair, and what should I be careful of?”

AI returns
  • How to frame a fair, specific, behaviour-focused conversation
  • A prompt to make a file note straight afterwards
  • A clear marker: “if this becomes formal, loop in HR before next steps”
AMBER · with care
Leader asks

“A team member has accused another of bullying. How do I investigate it and what disciplinary action can I take?”

AI returns
  • Does not walk the leader through an investigation
  • Names why this needs HR: fairness, legal exposure, wellbeing
  • Routes them to HR now, and notes what to preserve in the meantime
RED · stop & escalate
For the executive · why this is safe, not risky

The stop-line is a governance feature, not just a help feature

An exec's first instinct is rightly cautious: won't letting leaders self-serve policy create risk? Done this way, it does the opposite — it makes a currently invisible, inconsistent process visible, consistent and governed.

Consistency, not improvisation

Every leader gets the same grounded answer from the same source — replacing the lottery of who they happen to ask and what they half-remember.

A defined, auditable boundary

The red line is explicit and built in. You can see exactly what is self-served and what is escalated — and tune it. Governance you can point to.

HR focused where it matters

HR's time shifts from answering routine queries to handling the genuinely sensitive — and maintaining the policy base the tool draws on.

The discipline that keeps it trustworthy

Grounded in your policy. Advisory, never autonomous.

This only works if the answers are right and the boundary holds. Two disciplines make that true.

Grounded, not guessed

The tool answers from the organisation's own current policies and instruments — loaded into a secure, maintained knowledge base — and cites where each answer comes from, so the leader can verify. It is only ever as good as the content behind it, which makes HR's role as curator essential, not optional.

Advisory, with a hard stop

It is support, not authority. For anything green it informs; for amber it informs and prompts a check; for red it refuses to advise and routes to HR. The leader still owns the decision, and HR still owns the serious matters and the policy itself. Enablement and governance move together — that is what makes it safe to scale.

The break

When the routine answers itself and the serious is clearly flagged, the leader stops asking permission — and HR stops being a permission desk.

This is the hinge of the whole model: dependence dissolves on the simple, judgement is protected on the serious, and HR finally becomes the architect and guardian it was always meant to be.

Continue Part II
Part II · C

A Safe Place to Prepare

The rules handle “what am I allowed to do.” But the hardest part is often the human moment itself. Next: a private space for the leader to rehearse the difficult conversation, test the tone, and walk in prepared — coaching, not scripting.