Building AI Products
Chitrang Shah brings over 18 years of experience creating category-defining products and building large scale businesses. At Lattice, Chitrang was part of the executive team that built Forrester's highest-ranking B2B Customer Data Platform (CDP) and scaled the business from 0 to 30,000+ marketing & sales users resulting in the acquisition by Dun & Bradstreet in July 2019. Prior to Lattice Engines, Chitrang lead product management at Clearwell Systems, a market-leading big data application, where he brought three major innovations to market in less than two years, securing #1 position in the Gartner Magic Quadrant for eDiscovery Software and growing the company’s revenue from $50M to $100 million run-rate.
He recently spoke at a Products That Count hosted webinar and discussed lessons learned from building the first-generation B2B AI Products. Shah shared a lot of his personal experience in building AI products that any product manager can learn from.
Chitrang Shah ran through many lessons learned from building the first generation of AI products. It’s worth checking out the entire video, which is right here for you in this post. Otherwise, some of the highlights are detailed below:
On the challenges of building AI products
When it comes to building AI products, Shah took the webinar attendees through three types of challenges.
“I can give you three high-level themes. One is what I call heterogeneity which creates a lot of investment challenges on where you invest but also go to market challenges.
Then there is an adoption challenge. In a B2B software world, you all know that the adoption of the product is a challenge. It’s something you have to actively work on. But AI products have a specific sort of nuances. And if you don’t carefully manage it, and if you don’t know some small moves upfront, you might underestimate what it takes to get the solutions adopted.
Finally, this is my favorite, often the great becomes the enemy of good. It’s very much true in the world of AI. AI founders tend to be very much technology, back-end and come from physics, or a statistics background. Sometimes, in the hope of building the best things, you forget that sometimes that good is good enough.”
On the purpose of AI products
Just like every great product manager knows there has to be a purpose for building a product, there has to be a reason for building AI products.
“First, any AI product usually is trying to solve a workflow problem in the sense that you’re trying to reduce the friction. You’re trying to eliminate some manual work. Or you’re trying to provide some intelligence that helps the end-user make intelligent decisions better, faster.”
On adoption challenges with AI
When your team is building AI products, this is an important factor to consider.
“Think about all of these kinds of adoption. Challenges you’re going to face from the end-user. Before you start thinking about the experience you want to provide, don’t start from what AI can do. Think about how you’re going to blend AI into the end-users workflow day-to-day.”