Applied Case Study 001
When AI Adoption Still Depends on Human Scaffolding
How the Autonomy Index™ was used to evaluate a real customer adoption workflow inside a VC-backed B2B AI startup, and the post sale operating architecture that was built around it.
Based on firsthand operational experience, 33 account reviews, 30 operational reviews or incident briefs, and a new operating architecture implemented across a portfolio of 36 customer and pilot accounts. The product could automate outbound activity. Handover, readiness, configuration, assurance, recovery and value evidence still depended on people.
- VC-backed B2B AI startup
- 36 customer and pilot accounts
- More than £350k in tracked value
- 33 account reviews
- 30 operational reviews or incident briefs
- February to May 2026
The problem
The AI completed tasks. People carried the workflow.
The product could automate outbound activity, but the operating model around it had not matured at the same speed. Customer Success, customers and Engineering were still holding the workflow together manually.
- 01
Handover
Human-led
- 02
Readiness
Human-led
- 03
Configure
Shared
- 04
Execute
AI-enabled
- 05
Assure and Recover
Human-led
- 06
Prove Value
Human-led
The AI-enabled execution stage was more mature than the handover, readiness, assurance, recovery and value evidence stages surrounding it. That made Workflow Integration the primary constraint.
What was implemented
Diagnose. Structure. Govern. Measure.
A structured post sale operating architecture covering onboarding, activation, health, issue assurance, adoption, value evidence and commercial continuation.
Diagnose
- Mapped the complete customer workflow
- Reviewed 33 accounts and 30 operational briefs
- Separated product capability from workflow maturity
Structure
- Six stage post sale lifecycle
- Five day activation target
- 25 page onboarding handbook and eight supporting assets
Govern
- RAG customer health scoring
- Clear issue categorisation, escalation and closure evidence
- Recurring problems grouped into portfolio level patterns
Measure
- Written success criteria and success plans
- QBR value evidence
- Renewal, recovery and pilot to paid pathways
The operational result
Before, and the implemented operating state.
Handover
Before
Context scattered across conversations and individual knowledge
Implemented
Defined lifecycle, inputs, owners and customer readiness conditions
Onboarding
Before
Open ended and account dependent
Implemented
Reusable handbook, readiness checks, account plans and activation target
Customer health
Before
Risk identified through individual judgement
Implemented
RAG view supported by adoption, delivery and commercial indicators
Issue management
Before
Ad hoc reporting and recurring symptoms
Implemented
Categorised issues, reproduction evidence, escalation and validated closure
Value evidence
Before
Reconstructed late in the lifecycle
Implemented
Success criteria, success plans and QBR evidence per account
Commercial risk
Before
Renewal and budget exposure not consistently visible
Implemented
Dates, values, dependencies and recovery actions documented portfolio wide
- One defect heavy account reached a verified stable QBR baseline.
- Referral and expansion discussions began.
- Written success criteria and recovery plans were introduced.
- Customer validated issue closure was introduced.
- Material five figure commercial risks were surfaced earlier than existing reporting exposed them.
Assessed baseline
Autonomy Maturity Band: Level 2, Fragmented Pilots
Genuine AI capability, fragmented surrounding workflow.
Implemented state
Operating architecture designed to support movement toward Autonomy Maturity Level 3, Functional Adoption
Defined lifecycle stages, standardised operating assets, structured risk visibility, clearer ownership, stronger issue assurance and customer level value evidence.
Modelled commercial potential
What this operating model was modelled to be worth.
The operational changes were implemented and observed. The longer term commercial effect has been modelled using the assessed portfolio and clearly stated assumptions.
- Modelled annualised value
- £35k to £70k
- Modelled net contribution
- £10k to £45k
- Modelled net ROI
- 40% to 180%
- Modelled payback
- 4 to 9 months
Modelling assumptions
- Assessed portfolio of more than £350k in tracked annual, pilot and contra value
- Assumed implementation and maintenance envelope of £25k
- Protected commercial value of £17.5k to £28k
- Reduced manual rework of £12k to £24k
- Improved activation and pilot conversion of £7k to £18k
How the conclusion was reached
Evidence, dimensions, constraint, controls, reassessment.
- 01
Evidence collected across 33 account reviews and 30 operational reviews or incident briefs
- 02
Eight formal dimensions evaluated across the four weighted pillars
- 03
Workflow Integration identified as the primary constraint
- 04
Operating controls implemented across the post sale lifecycle
- 05
Reassessment recommended once a full measurement cycle is available
The workflow was scored against the same four weighted pillars and eight formal dimensions used in every Autonomy Index™ assessment.
Leadership & Strategy
20%Leadership Alignment
Strategic Integration
Workflow & Operations
30%Workflow Integration
Operational Dependency
Behaviour & Trust
30%Behavioural Adoption
Operational Trust & Override
Governance & Risk
20%Governance Maturity
Risk & Output Assurance
Supporting detail
Methodology and evidence
This applied retrospective case study draws on firsthand operational experience, 33 account reviews, 30 detailed operational reviews or incident briefs, and contemporaneous customer, commercial and product evidence. The assessment reports a maturity band rather than a numerical score because the company ceased trading before a formal post-implementation reassessment could be completed. The company and its customers have been de-identified.
Four weighted pillars: Leadership & Strategy 20%, Workflow & Operations 30%, Behaviour & Trust 30%, Governance & Risk 20%. Autonomy means earning the right to delegate: the system must be embedded enough to rely on and governed enough to rely on safely.
- Leadership AlignmentDoes the stated AI narrative match operational reality?
- Strategic IntegrationIs AI connected to objectives, budget and decision making?
- Workflow IntegrationIs AI embedded in the workflow or bolted alongside it?
- Operational DependencyWould the workflow degrade if the AI were removed?
- Behavioural AdoptionDo people use it by default or only when prompted?
- Operational Trust & OverrideHow often is output overridden, and on what basis?
- Governance MaturityWho owns the output, and how is it controlled?
- Risk & Output AssuranceHow are errors detected, escalated and corrected?
Commercial modelling assumptions
The modelled range applies a £25k implementation and maintenance envelope to the assessed portfolio. Components: protected commercial value £17.5k to £28k, reduced manual rework £12k to £24k, improved activation and pilot conversion £7k to £18k. Together these give a modelled gross range of approximately £35k to £70k, a net contribution of £10k to £45k and a net ROI of 40% to 180%.
AI adoption is not the same as operational autonomy.
The Autonomy Index™ identifies where AI is genuinely carrying a workflow, where people are still holding it together, and what must change before greater autonomy can create dependable business value.