Few key points to get the AI journey right (as I understood)

With AI specifically generative AI being stated as the most disruptive technology, to come in after the smartphone revolution, the IT organization in the large enterprises are looking for the use cases to apply these technologies.

1. AI is the last mileThere are multiple steps that need to be completed before an identified use case can evolve from POC stage to production. (a) Digitization of the business functions, to get both business data and process metrics from the systems.(b) Domain specific data models to have true representation of the state of the systems and tracking through its lifecycle with the required fidelity and accuracy(c) Eliminating the latency between source of data and decision point (inferencing), to be highly responsive in actions with both strong signals and weak signals. The emphasis is only on strong signals to reduce latency, but when there is data insufficiency, weak signals can help make informed decisions with ML models, instead of being in an indeterministic state.(d) Moving from a deterministic decision processing (rule based) which is fully explainable to non-deterministic like AI (with confidence score), needs change of mindset of people as still it is human in the center of the AI adoption.
2. AI use cases are best determined by the process owner and technology is only a facilitator.(a) AI technology specialist & data scientist brings in good understanding of technology, mathematical modeling and analytics skills, and the real business process cannot be fully understood by an external person with just telemetry data. The process owner can bring out the problem statement with considerations on efficiency, profitability, safety, and user experience.(b) Use cases from elsewhere including from the same industry sub-segment is just a starting point. The nuances of how value is created and how internal and external factors affect value creation varies across enterprises. [Again, look from inside](c) For the disruptions/ new ideas initiated from the inside, there is collaborative in building the prototype, working in validating the new concept and scaling from concept to product scale implementation. [I am involved in creating the new ways of working, dramatically reduces the effort in implementing the change]
3. AI champions are storytellers(a) Though AI starts as a technology initiative and early champions are from technology, but technology team need to go beyond just being implementors to becoming evangelist, getting business, finance, legal and risk aligned to create value with their articulation and engagement skills.(b) Bring in zeal into the systems, amplify early wins and energize the others to follow, create strong collaborative across the participants in ensuring a cascading effort, resulting in wider adoption.

Most of the time, it is puzzling to hear that the enterprises are identifying the AI use cases to be taken up, as they have asked by others as what they are doing with AI or more from the fear of missing out. However, to be successful with AI, it is important to put the business process value stream and business process owner in the center and create a support structure with AI technologist, data scientist and process change agents to ensure apt usage of the technology in enhancing the customer value & experience.