Kainos AI CoE · Delivery Framework

The DARE Framework

The delivery cycle Kainos uses to turn AI intent into customer value — Design, Activate, Realise, and Evolve. A practical operating model for every AI engagement.

Design Activate Realise Evolve

01 · What is DARE?

A delivery cycle, not a methodology.

DARE gives every Kainos AI engagement a shared operating rhythm — repeatable enough to scale, flexible enough to fit any client context.

A delivery cycle not a methodology

DARE is an operational loop, not a waterfall process. Each phase produces a concrete output that gates the next — and the cycle repeats as AI scope expands.

Forward-deployed execution

Kainos consultants work alongside client teams in each phase, not as remote advisors. The goal is capability transfer as well as delivery.

Outcome-obsessed by design

Success metrics are defined before build begins and measured at each phase gate. AI that cannot demonstrate value does not advance.

02 · Design

What gets designed before a line of configuration is written.

Design is the phase that prevents rework. Discovery scope, readiness, and success criteria are locked here — not assumed at go-live.

Best for

  • Scoping the first agentic capability with the client
  • Establishing AI readiness across people, process, and data
  • Running a structured data quality assessment
  • Baselining the security and governance posture
  • Locking success metrics before any configuration begins

Watch out for

  • Scope creep before readiness is confirmed
  • Success metrics defined post-hoc rather than upfront
  • Security and data quality treated as Activate-phase concerns
  • Skipping stakeholder alignment on what "good" looks like

Start here

  • AI Readiness Assessment (people, process, data, governance)
  • Success metric definition workshop with the client exec sponsor
  • Data quality review against the target agent's required inputs
  • Security group mapping for the first agent population

03 · Activate

Turning design into running AI.

Activate is where configuration happens — and where baseline metrics are captured so Realise has something to measure against.

Best for

  • Configuring available Workday agents against the agreed scope
  • Deploying the first agentic capability into a non-prod environment
  • Wiring data sources, security groups, and governance controls
  • Capturing baseline metrics before end-user access is granted
  • Running a structured UAT cycle with the client HR team

Watch out for

  • Going live without baseline metrics captured — you lose the comparison point
  • Security configuration left until the end of the phase
  • UAT treated as a sign-off step rather than an iteration cycle
  • Skipping the non-prod validation step under time pressure

Start here

  • Configure ASOR and security groups in non-prod first
  • Deploy agent to a pilot population before full rollout
  • Capture time-to-answer and task-deflection baselines
  • Confirm credit consumption and governance reporting are live

04 · Realise

From running to value.

Realise is the adoption phase. The agent is live — the work now is making sure people use it, measuring what changes, and iterating on configuration based on real usage data.

Best for

  • Running targeted adoption campaigns for the agent population
  • Delivering structured change management and manager enablement
  • Measuring usage against the baselines captured in Activate
  • Iterating on agent configuration based on real usage patterns
  • Building the first ROI narrative for the exec sponsor

Watch out for

  • Treating go-live as project completion — adoption is the real work
  • Measuring too early before usage has had time to ramp
  • Change management that is comms-only with no manager involvement
  • Configuration changes driven by anecdote rather than usage data

Start here

  • 30/60/90-day adoption plan with named owner from the client side
  • Weekly usage review against baseline metrics for first 8 weeks
  • Manager briefing pack and talking points for team leads
  • First ROI data point packaged for exec sponsor at 60 days

05 · Evolve

Sustained and expanding value.

Evolve is the phase that turns a project into a programme. Periodic health checks, new agents or skills, and a clear ROI story build the case for the next DARE cycle.

Best for

  • Quarterly health checks against the original success metrics
  • Identifying the next agent or skill to add to the programme
  • Measuring return on investment across the full agent portfolio
  • Building the business case for the next DARE cycle
  • AMS-led ongoing optimisation and governance

Watch out for

  • Adding new agents before the first deployment is fully adopted
  • ROI measurement that relies on self-reported estimates rather than system data
  • Losing exec sponsorship between cycles — keep the story visible
  • Treating Evolve as maintenance rather than growth

Start here

  • Quarterly business review template tied to DARE phase outcomes
  • Agent roadmap review with the solution architect and AMS lead
  • ROI deck for exec sponsor anchored to system-sourced metrics
  • Next Design phase scoping initiated before Evolve concludes

06 · How Kainos uses DARE

Who owns what across the delivery cycle.

DARE is a team sport. Each role has a primary accountability in the cycle — and a handoff point to the next.

Role Primary DARE Responsibilities
Delivery Lead Owns the end-to-end DARE cycle for the engagement. Locks success metrics in Design, gates progression through each phase, and escalates adoption blockers in Realise.
Change Lead Leads stakeholder alignment and change impact assessment in Design. Owns the adoption campaign in Realise. Monitors manager engagement and tracks usage uplift against plan.
Solution Architect Drives technical scope in Design, leads configuration and UAT in Activate, and identifies the next capability in Evolve. Owns the agent architecture roadmap.
AMS Lead Joins from Realise onward. Owns ongoing configuration optimisation, quarterly health checks in Evolve, and the AMS engagement scope for the next cycle.