Kainos AI CoE · Outcome Framework

The AI Outcome Framework

The measurement backbone that proves AI value at every stage — from first activation through to autonomous operation. Four levels, each with distinct outcomes, metrics, and advisory angles.

Foundation Quick Wins Advance Create

01 · Why measure outcomes?

What you measure is what you can defend.

Clients need a clear story at every stage — from first activation through to full AI programme maturity. The framework gives that story a consistent structure.

Activation is not adoption

An agent that is switched on but unused delivers no value. The framework measures actual usage, task deflection, and workflow change — not configuration completion.

Metrics justify the next investment

Every level of the framework produces evidence that funds the next. Without measurement, the AI programme stalls at Level 1 — configured but never growing.

Clients need a story not just a stat

The framework translates raw metrics into a narrative: where we are, where we are going, and what it takes to get there. That is the conversation that wins exec sponsorship.

02 · Foundation (Level 1)

AI is available — but not yet embedded.

Foundation is the baseline level. The agent is configured, security groups are in place, and users are aware it exists. Value has not yet been measured, but the infrastructure to measure it is live.

What it looks like

  • Agent configured in the production tenant
  • Security groups mapped and access policies confirmed
  • User awareness communication sent to target population
  • Baseline metrics recorded: activation rate, session starts, awareness survey
  • No measurable workflow change yet — this is the starting line

Key metrics

  • Activation rate — % of target population with agent access enabled
  • Security group configuration completeness
  • User awareness score from post-launch survey
  • Number of session starts in first 30 days

Kainos actions

  • Configure and validate agent in production
  • Deliver security group readiness checklist
  • Run awareness communications with the client change lead
  • Capture and baseline all Level 1 metrics before moving to Level 2

03 · Quick Wins (Level 2)

First measurable value delivered.

Quick Wins is where the ROI story begins. Task deflection, time-to-answer, and case reduction are measurable within weeks of activation — and defensible to exec sponsors.

What it looks like

  • Employees resolving queries via the agent rather than contacting HR
  • Policy Intelligence returning accurate answers to policy questions
  • Case volume in the help queue beginning to fall for targeted query types
  • Time-to-answer for common questions measurably shorter than pre-agent baseline

Key metrics

  • Task deflection rate — % of queries resolved by agent without HR intervention
  • Time-to-answer — average resolution time vs. pre-agent baseline
  • Case reduction — volume drop in targeted HR case categories
  • Agent satisfaction score from end-user feedback

Kainos actions

  • Run targeted adoption campaign for the Level 2 metric population
  • Activate Policy Intelligence if Workday Help is live
  • Deliver first productivity measurement report at 60 days
  • Package the deflection story for the exec sponsor

04 · Advance (Level 3)

AI embedded in daily workflows.

Advance is the maturity level where AI stops being a tool people remember to use and becomes part of how work gets done. Weekly active usage and process cycle time reductions are the signal.

What it looks like

  • A measurable share of the target audience using the agent every week
  • Managers using coaching and performance skills without HR prompting
  • Process cycle times (e.g. absence requests, onboarding tasks) measurably shorter
  • Agent usage patterns informing configuration iteration and skills expansion

Key metrics

  • Weekly active agent users as % of target audience
  • Process cycle time reduction for targeted workflow categories
  • Manager coaching skill uptake rate
  • Skills coverage breadth — number of active skill categories per user

Kainos actions

  • Iterate agent configuration based on real usage patterns
  • Expand skills coverage to the next priority workflow category
  • Establish governance cadence with the client HR ops lead
  • Present Level 3 metrics in quarterly business review

05 · Create (Level 4)

New capabilities built on Workday's platform.

Create is the frontier level — where clients move from consuming Workday's native agents to extending the platform with custom agents, automations, and integrated workflows.

What it looks like

  • Custom agents deployed using Workday Extend or Sana design patterns
  • Integration touchpoints automated end-to-end without manual intervention
  • Agent factory patterns in use — repeatable deployment of new skills
  • Client team building new skills independently with Kainos in an advisory role

Key metrics

  • Custom agents deployed — count of net-new agent capabilities in production
  • Integration touchpoints automated — end-to-end workflows without manual steps
  • Time-to-deploy new skills — from design to production (target: days not weeks)
  • Client self-sufficiency score — % of new skills deployed without Kainos build

Kainos actions

  • Workday Extend and Sana design advisory for custom agent architecture
  • Agent factory pattern design and documentation
  • Integration automation scoping with the client's technical team
  • Capability transfer programme so the client can build independently

06 · Using the framework with clients

How each role uses the framework in practice.

The framework is most effective when the whole team speaks the same level language with the client — from pre-sales through to AMS.

Role How they use the framework
Account Exec / Pre-Sales Uses the four levels to map where a prospect sits today and position the Kainos engagement as the path to the next level. Anchors commercial conversations to measurable outcomes rather than feature lists.
Delivery Lead Sets the target level for each DARE cycle and defines the metrics that gate progression. Presents level status at project reviews and uses gaps to prioritise the backlog.
AMS Lead Tracks level attainment across the client portfolio in steady state. Uses the framework to identify clients stuck at Level 1 or 2 and builds a case for additional engagement scope.
Exec Sponsor Receives level-framed reporting in QBRs. Uses the framework to communicate AI programme progress to the board and to justify investment in moving to the next level.