Novosus Framework
I.M.P.A.C.T.: the enterprise AI value realisation framework
Identify · Map · Prioritise · Architect · Capture · Transform — a six-phase methodology for moving from fragmented AI experimentation to disciplined, measurable value, grounded in research from BCG, McKinsey, Deloitte, PwC, Accenture, and MIT Sloan Management Review.
Why most AI programmes stall
Across analyst commentary, consultant surveys, and practitioner analysis, the same failure patterns recur:
- Starting with technology or tools instead of business problems and workflows.
- Poor data foundations and integration capabilities.
- Lack of workflow redesign and human-AI collaboration design.
- Weak governance and risk management, especially around GenAI.
- Underinvestment in change management, training, and support — leading to low adoption.
BCG's research finds that roughly 10% of AI value comes from algorithms, 20% from data and technology, and 70% from people and process. I.M.P.A.C.T. treats this as a design constraint: most effort in Architect, Capture, and Transform goes to workflow redesign, operating-model change, and governance — not model selection alone.
The six phases
Sequential in direction, iterative in practice — quick wins can move into Architect and Capture immediately, while longer-horizon bets proceed through fuller design and governance cycles.
Identify
Align AI efforts to three or four strategic priorities and assess readiness across data, platform, governance, talent, and change capacity — before committing to a large portfolio.
Map
Generate and structure AI use cases through decision and journey mapping, blending top-down strategy with bottom-up discovery from real workflows.
Prioritise
Score and gate use cases into a focused, sequenced portfolio of 5–12 near-term initiatives — quick wins, strategic bets, capability builders, and parked ideas.
Architect
Redesign workflows and roles, define KPIs and responsible-AI controls, and build pilot charters and production-aware technical plans.
Capture
Run time-boxed pilots against a defined evidence standard — value, quality, control, and operability — not just qualitative feedback.
Transform
Scale what works across three parallel substreams: platform & architecture, operating model & workforce, and governance & risk.
Frequently asked
What is the I.M.P.A.C.T. framework?
Novosus's six-phase enterprise AI methodology — Identify, Map, Prioritise, Architect, Capture, Transform — for moving organisations from fragmented AI experimentation to disciplined, measurable value realisation.
What research is it based on?
Published findings from BCG, McKinsey, Deloitte, PwC, Accenture, and MIT Sloan Management Review on what separates AI programmes that deliver sustained financial impact from stalled pilots and "innovation theatre."
What is the 10-20-70 principle?
BCG's finding that roughly 10% of AI value comes from algorithms, 20% from data and technology, and 70% from people and process — used here as a constraint that keeps effort focused on workflow and governance, not just technology choice.
How long does a full I.M.P.A.C.T. engagement take?
I.M.P.A.C.T. is a phased methodology, not a single fixed-length engagement. Most clients move through Identify, Map, and Prioritise — roughly 4–8 weeks — before deciding how many Architect–Capture pilot cycles to run, typically 6–12 weeks each. Transform, the scaling phase, continues for as long as it's delivering value.
Start with Identify
An AI diagnostic with Novosus is the Identify phase in practice — a short, structured assessment of where AI can create real value in your business, and whether your data, governance, and team are ready for it.