AI Founder Series: Solving Problems Unique to AI-Enabled Startups

During nearly every tech industry event I attend, at least one person asks me what it means to invest in ‘AI startups.’ Inevitably I’ll hear, “But isn’t every company destined to be an AI company?”

To be fair, they’re right — every company will soon have some aspect of their business that relies on machine intelligence. But the way I see it, it’s not about the fact that every company uses AI; it’s about the fact that a certain set of companies have value props rooted in machine learning, and that results in entirely unique challenges.

Building an AI startup means struggling to find and hire top machine learning engineers. It means collecting large quantities of data for long periods of time without revenue. It means figuring out how to execute a human-in-the-loop strategy such that neither the human nor the machine ends up becoming a bottleneck. And it means dealing with people who think AI startups are just disoriented stats majors throwing regressions at things (I kid — mostly).

To help shed light on challenges like these and many others, we’re launching AI Founder Series: Solving Problems Unique to AI-Enabled Startups. Each video will highlight a key challenge one of the AI startups we work with has been able to overcome. We hope these stories are relatable and the case studies valuable to the broader community of founders working to commercialize machine learning.

Our first episode, Selling a Black Box, explores what it really means to democratize AI. Alex Weidauer, CEO of Rasa, which builds tools to help developer and product teams, explains that democratizing AI necessitates understanding the dynamics of enterprise teams.

Selling NLP tools to enterprises means convincing customers that they are capable of deploying conversational AI on their own, with limited resources. By prioritizing simplicity and simultaneously integrating with both product owner and developer workflows, Rasa has been able to successfully get customers on board with this idea.

Moving forward, our AI Founder Series will cover topics including:

• Optimizing a human-in-the-loop workflow
• Scaling operational teams
• Iterating hardware and ML/software simultaneously
• Finding the right balance between vertical and horizontal GTMs

Post authored by John Mannes, investor at BSV