Enterprise Generative AI Solution Opportunity Identification Process

I thought I would share what we use at work to help identify opportunities for generative AI implementations. Here is the step by step process feel free to ask any questions or provide any feedback. This is less technical its what we put in front of decision makers before collaborating with the teams. So mainly product management, Directors, Architects go through this process with us and then we work with a tiger team for development following Lean, Agile, and DevOps Best Practices.

GenAI Alignment: Align generative AI initiatives with your overall business strategy to ensure they support key objectives. (We get the vision, strategy, and tactics back of the napkin)
Business Context: Understand the current business landscape and identify areas where AI can drive transformation. ( We use Strategyzers business model canvas here primarly with some other self made canvases)
Value Streams: Map out the value streams to pinpoint where AI can enhance efficiency and value delivery. (Use the basic Karin Martin Value Stream Mapping process here. You cant really improve what you can measure and following the "theory of constraints" every value stream is severely suppressed by some major bottleneck and we want to start with those major bottlenecks as improvement items first before smaller less valuable chagnes)
Objectives & Key Results (OKRs): Set clear objectives and key results to measure the success and impact of AI initiatives. (First we collect their current OKRs and KPIs at the company/portfolio level as well as any that are related to their teams.)
Opportunity Identification: Identify specific opportunities within the value streams where AI can provide the most significant benefits. ( Using the business context, OKRs and KPIs we then use the value stream maps to identify where a generative AI solution will give the biggest bang for our buck)
MVP Definition & Prioritization: Define minimum viable products (MVPs) for the identified opportunities and prioritize them based on potential impact and feasibility. (We focus here on leaning out the first iteration functionality based on value measurements and economic prioritization methods. You can also use any of the luke holeman innovation games here as well to get to the right scope for the MVP)
Action Plan: Develop a detailed action plan to implement the prioritized MVPs, ensuring a focused and structured approach to leveraging AI. ( Lastly we focus on the actual tactics of implementation like who will build it, what they need, who they need to collaborate with etc and put everything into the teams backlog. )

This is a very high level approach because we work with teams all over our company and everyone is a bit different so this process can be modified as you see fit. This has been excellent for us to cut through the noise and focus on extracting value quickly.

I hope you found this valuable.

5 comments