I’m thinking about experimenting with AI, and I don’t want to go it alone. Where do I start?
Step 6: Develop a Roadmap
Roadmaps can be a helpful tool, but they can also be an annoying waste of time. How do you balance the helpfulness with the need to create a roadmap that can change necessarily and rapidly?
Much can be learned from how we build and use product roadmaps. What uses does a product roadmap serve? When communicating about the future of a product, it can be helpful to share a roadmap, including a timeline, to ensure that all stakeholders are on the same page.
But how about an experimental AI project? How do the roadmaps differ?
For the AI Experimentation Roadmap, you’ll want to talk about the purpose of the project and how you will be exploring, testing, and implementing new AI technologies to enhance workforce productivity, efficiency, and innovation.
• Focus Areas: In the case of an experimental AI roadmap, you will want to focus primarily on internal process improvements, pilot projects, and the gradual integration of AI tools within the organization. If you were building a product roadmap, you’d probably want to focus on features, market demands, user needs, and competitive positioning instead.
• Project Phases: For an experimental AI roadmap, you might want to establish a Preparation Phase, a Pilot Project Identification Phase, an Evaluation and Iteration Phase, and a Scaling Phase. In the case of a product roadmap, you’d probably include more traditional phases like Discovery, Planning, Development, Launch, Post-Launch, and the Growth/Maturity Phase.
• Project Scopes: An experimental AI roadmap would probably include multiple small-scale projects or pilots to test various AI applications whereas a product roadmap might include an entire product lifecycle, from ideation to launch and beyond.
• Stakeholders and Collaboration: Because experimental AI roadmap projects are internally focused, the stakeholders will be a bit different. They would probably include department heads, IT teams, data scientists, and the internal end-users. Collaboration between internal teams plus outside vendors who are helping with the projects (i.e. Humaxa — we can help). However, for a product roadmap, your stakeholders will be different. They might include product managers, marketing teams, sales teams, customers, and external stakeholders such as investors. Collaboration will be different too, with an emphasis on communication between Marketing Teams, Product Teams, and Customers.
• Metrics and KPIs: Assuming you have solidified your success criterial and KPIs and are sure they will accurately measure the impact of your experimental AI projects, the metrics you choose for an experimental AI roadmap and a product roadmap will differ. In general, product roadmap KPIs will have more an external, market-orientation. Experimental AI Roadmap metrics will generally have more of an internal focus. For example, Experimental AI Roadmap metrics might include efficiency improvement, cost reduction, employee adoption rate, and compliance success metrics. You might choose to measure a reduction in downtime or an improvement in customer service. On the other hand, metrics for a Product Roadmap might include client engagement metrics, user acquisition rates, market share, revenue growth, customer satisfaction, customer retention rates, or feature adoption rates.
Of course, there are other items you may want to include on an Experimental AI Roadmap. You might want to include a section about training or a section on change management. If the AI experiment will alter people’s workflow at all (hopefully for the better!) I’ve seen it work well to walk through the upsides and downsides of the new workflows. It might require a slight change in the workflow, but usually the time and energy saved makes it more than worth it.
As you strike out and begin (or continue) your experiment with AI, I hope these steps have been helpful. There’s so much more to share and I’d love to share more details with you — and help — if you’d like. Just let me know.
Carolyn Peer
CEO/Co-founder, Humaxa
carolyn.peer@humaxa.com
