Architecting Insights, Features and Initiatives for AI backlog management
The basic architecture we'll need for SupaPM so that AI can take some of the stress of Product Management away. Insights, features and initiatives are the foundation on which we'll build.
In this video I share how I'm thinking of architecting three different entities in my AI product management SaaS supapm.com.
3 minute video on the architecture for AI product backlog management
Insights
Fundamentally we'll need to capture "insights" - which can be customer quotes from sales calls, feedback sent in by existing customers, or ideas from internal staff etc. These insights have an "importance" (something that's nice to have is a lot less important than something that's a must have for a customer). Ideally these insights are also linked to a monetary amount for a given customer or prospect. So if that insight is coming from a CRM like Hubspot, we can ingest the deal size too of course (more on that later).
Features
We then need a way to pair these insights with features. The Product team would create a feature and "link" relevant insights to it. If the insights were captured well with supporting data, we'll then have direct customer quotes and linked prospects that expressed interest in that feature. This makes future prioritisation a lot easier, as well as crafting an effective go-to-market strategy for the feature.
Initiatives
Finally we have initiatives. In my head these are things that the business wants to achieve (maybe 5% market penetration in a new geography, for instance). The product team can then link features to initiatives, again providing more context for better prioritisation and clarity for all stakeholders.
AI Product Backlog Management
The goal, of course, is to ingest insights from all the different potential channels let AI analyse them to find what features they're relevant for. Eventually we can even let the AI suggest features that would solve the biggest problems for customers, but babysteps for now...
Follow along to see what's next - some integrations to make the gathering and analysis of insights automatic.
