Function 01 / 07 · Enable

How your operation gets ready.

Enable is the work nobody costs in. It's the difference between an operation that ramps in 12 weeks and one that ramps in 6. Onboarding the team. Briefing the AI. The path from new hire to senior, and from raw system prompt to calibrated AI.

SAMPLE READING CALIBRATION-04 · APR 26
Function 01 / 07 · Enable
You're emerging on this one.
CALIBRATION-04
v 02 · live
52/100
Composite
Emerging
01
Reactive
Now
Emerging
03
Defined
04
Optimised
Next move Build a competency map. Most onboarding is a deck. The AI's onboarding is a system prompt nobody owns.
02

What Enable means.

Enable is the work of getting your operation ready to do the work. Both populations. Not the work of writing process docs. Not the work of running classroom training. The work of designing how humans ramp from "first day" to "trusted operator," and how the AI gets briefed on what good looks like, with named milestones and a path that shortens every quarter.

Most CX teams have onboarding that's a slide deck and three weeks of ride-alongs. New hires absorb the work through osmosis. The good ones figure it out fast. The rest take months. Onboarding cost is a line item nobody owns.

On the AI side, the parallel problem: the system prompt was written in a hurry. The vendor's CSM tweaks it on a schedule the team doesn't see. New product changes flow into the AI through a different channel than they flow into onboarding. Two different versions of "what we tell new agents" are running in parallel and nobody reconciles them.

That's the gap Enable names. Ramp time is real money. Six months of ramp on a team of 30 is a serious capacity loss every year. An AI running on a six-month-old prompt is a different operation than the one the team is being trained to. Most CX leaders know both and still don't have a competency map, a milestone schedule, or a way to keep the AI briefed alongside the team.

Haven's Enable module builds the competency map first. The capabilities every operator needs (human or AI), broken into named tiers, with assessment criteria. A milestone schedule replaces a slide deck. The AI's system prompt is treated as onboarding artifact, not vendor mystery. Both populations ramp against the same map.

Ramp velocity becomes a number you can measure. AI prompt drift becomes a number you can catch. Onboarding cost becomes a line item you can defend.

03

The progression. Four levels.

Level 01 You've passed
Reactive

Onboarding is a slide deck. The AI runs on a system prompt nobody reads. New hires shadow seniors. The path to competence is informal on the human side and invisible on the AI side. Ramp time varies wildly between hires.

  • Slide-deck training
  • Shadow-and-ask method
  • AI prompt owned by a vendor CSM
  • Ramp time is anyone's guess
Level 02 · Now You are here
Emerging

Some structure exists for humans, but it's lead-dependent. The AI is still on its own track. Specific seniors are great trainers; the rest aren't. Some milestones are tracked, most aren't. The AI's prompt drifts on a vendor cadence. Ramp time is improving but unpredictable; AI behavior is unread.

  • Lead-driven training
  • Some milestones tracked
  • AI prompt updates land without operator review
  • Senior agents and the vendor are the bottlenecks
Level 03 2-3 months out
Defined

The competency map is named, owned, and used across humans and AI. A milestone schedule replaces ride-alongs. The AI's prompt structure is treated as onboarding artifact. Every new hire moves through the same path; every prompt update is reviewed against the map. Ramp velocity is a measured number.

  • Competency map covers humans and AI behaviors
  • Milestone schedule
  • AI prompt versions tracked alongside ramp
  • Owner named
Level 04 12+ months out
Optimised

The map evolves with the work and updates both populations. Calibration findings update the competency tiers, the onboarding ladder, and the AI's prompt structure. Onboarding self-improves. Ramp velocity decreases each quarter and AI drift gets caught at the prompt line.

  • Self-improving map
  • Skill gaps on humans, prompt gaps on AI, same backlog
  • Continuous learning paths for both
  • Ramp velocity decreasing; AI drift contained
04

What Enable builds.

Artifact 01

The competency map

Capabilities every operator needs, broken into tiers with assessment criteria. Covers human capabilities and AI behaviors. Named, version-controlled, evolved quarterly.

  • Tier 01–04 levels with criteria
  • Mapped to Perform standard dimensions
  • Covers AI behaviors as well as human capabilities
  • Quarterly review & version history
~6 hours to first draft
Artifact 02

The milestone schedule

Six-week ramp where new hires move from Tier 01 to Tier 02. Named checkpoints, named owners, predictable cadence.

  • Week-by-week ramp checkpoints
  • Named milestone owners
  • Pass / hold / extend criteria
  • Linked to live competency map
6 weeks · per hire
Artifact 03

The ongoing learning cadence

A monthly 60-minute team learning ritual where one new capability is named, taught, and added to the map. Same backlog feeds the AI's prompt updates. Continuous improvement made routine.

  • One capability named, taught, added
  • Senior-led teach-back format
  • Same backlog feeds AI prompt updates
  • Recorded & indexed in Know
60 min · monthly
05

See it cascade.

A QA flag on the team often traces back to Enable. The same product knowledge gap is hitting the AI on the same intent, and the fix routes through here. One root cause, not two performance conversations. See how the cascade lands in Enable →