Processes
Operating model processes describe how AI initiatives and AI products move from incoming demand to adoption, operations, and scaling.
If entities answer "what we manage" and roles answer "who is accountable," processes answer how everything moves.
Core process
Demand
↓
Evaluation
↓
Routing
↓
Delivery
↓
Adoption
↓
Impact expectation and confirmation
↓
Operations / scaling / closure
1. Demand intake and formalization
The business brings a problem, idea, or repeated pain. The AI function helps turn it into an initial AI initiative: as-is process, owner, users, expected impact, and constraints.
2. Evaluation and prioritization
The initiative is evaluated for value, AI applicability, data readiness, risks, duplicates, urgency, and strategic fit.
3. Routing to an AI product
After evaluation, the initiative is linked to the right AI product or capability: RAG, LLM, ML platform, AI agent, automation, code agent, and so on.
4. Delivery track
The initiative enters the delivery track of the selected product. Different solution classes require different checks, artifacts, quality criteria, and security requirements.
5. Stage gates
Control points confirm readiness to move forward: owner, data, selected product, success criteria, risks, and adoption plan.
6. Adoption
Technical launch is not adoption. Users need training, first-use support, feedback loops, business owners, and usage metrics.
7. Impact confirmation
After adoption, expected value must become confirmed impact or a management decision: scale, refine, stop, or hand over to operations.
8. Operations and evolution
If the solution remains in use, it needs a support model: product owner, SLA/SLO, monitoring, updates, incident handling, and roadmap.