Four Reminders for AI Startups: Focus, Customer Willingness to Pay, Sales Channels, and Big Tech’s Roadmaps

Points from yesterday’s talk: 1. Focus your business model. You are unlikely to make money from everything. 2. Sometimes users do not feel much difference in the quality of AI technology. Recognition or accuracy rates of 93%, 94%, and 95% may not produce much difference in actual use. A system might even achieve 95% recognition in one setting and fall to 90% in another. Each provider excels in different settings. Rather than pursuing a 1–2% improvement in recognition, examine the user’s entire workflow. Identify the few most critical points and address urgent needs. That is what may make users willing to pay. 3. Technology quality is one thing. Which channels and partners you use to sell is what matters. Sometimes the problem is not the product or technology, but that you have built something you cannot sell. The point is to make users willing to pay for it. 4. Many technologies still require thinking about major companies’ roadmaps. Apple’s direction, for example, appears to be adding medical-device functions to wearables. We therefore see functions such as electrocardiograms on Apple Watch and hearing-aid functions on AirPods Pro. Startups need to understand that if major technology companies target the broad consumer market (B2C), while established specialist companies focus on professional channels (B2B), they must think much more carefully about what space remains for them.