Getting Started with AI
Half-day interactive work session:
Executive introduction to AI. In plain English we’ll describe the different types of AI and the strengths and weaknesses of each. Depending on backgrounds and interest we can go deeper into the risks of Generative AI, how to evaluate the quality of a solution, alternative business models, prompts as the new IP, and planning for user acceptance.
Following the introduction, we will conduct a guided ideation on ways your specific business can leverage AI. Optional follow up includes a technical assessment of each and guidance for planning for quality and demonstrable ROI.
Avoiding POC Purgatory
Most AI projects get stuck in the Proof of Concept stage, never moving to production. Key blockers include lack of alignment between business and technical stakeholders, lack of rigorous quality metrics and improvement plans, and poor user acceptance.
To ensure stakeholder alignment we conduct one-on-one and small group interviews with different stakeholders to learn what each wants out of the project and what they are willing to put into it. With this we identify lack of alignment and build a plan to get alignment.
Evaluating the accuracy of GenAI programs is complex because there isn’t only one correct answer, but relying on Domain SMEs to monitor quality can take those key assets off-line for unacceptably long periods. We use the latest approaches to develop robust automated assessment.
GenAI has great potential to make your most highly paid employees more efficient; but these professionals usually have great leeway in how they do their jobs. If you build it, they will come rarely works out. We ensure user adoption by mapping current processes, identifying blockers to adoption, finding the early adopters who can become evangelists and testing the system with the people you want to use it.
Building AI-powered software
Adding AI to your software is expensive. You need to know if new features will pay out in new subscriptions and trade-ups to premium subscriptions. We work with your team to define the customer and user value propositions and conduct interviews and surveys among your intended buyers and users to get an objective assessment of whether and how adding AI to your software will pay out.
When you need to choose among multiple new features we use discrete choice and max-diff to measure which will add most to your product’s appeal.