Users desire AI in SaaS products to quietly remove or automate entire workflows, rather than adding new steps, screens, or decisions. The goal is to reduce friction and simplify processes, even if it means fewer visible 'features.'
What's the right way for SaaS companies to add AI to their product and who is doing it well? On our most recent Move the Needle podcast episode, Rodrigo Fernandez from ProductLed shared a story about how a SaaS company's activation rate dropped by 20% when they added AI to their product. Hear him tell the story -> https://lnkd.in/ez5zsU7G Every SaaS company is racing to call themselves “AI-first” or "AI-native," present company not excluded. It’s expected. It’s 2026. A Databox investor recently told me, "Software companies aren't raising funds right now unless they have an AI story where customers are getting value out of it." It's not about slapping a name or making an announcement. The AI needs to provide value to the customer that they couldn't get without AI. The mistake Rodrigo sees over and over: • AI added as a feature, not as a value accelerator • Cheap models & shallow prompts • “AI onboarding” that asks questions… and does nothing meaningful with the answers This shouldn’t come as a news flash, but users aren’t impressed by AI existing. What impresses them is AI removing work or doing something they couldn't do before. Here’s a better pattern: • Start with the outcome the customer wants • Use AI behind the scenes to configure, personalize, and guide • Reduce time-to-value instead of adding another layer to learn Rodrigo made a great point, "Your users already use great AI tools every day. Their expectations are high, and getting higher.” In our product, we've rolled out a few AI-powered capabilities last year that deliver value and have happy users. 1️⃣ AI performance summaries: They summarize performance in words --> https://lnkd.in/exaGPvYf 2️⃣ AI for building metrics from advanced datasets. Most companies struggle to turn their data into metrics. It requires an understanding of how the data is stored (tables, common identifiers between tables, field names, etc), how and if it can be aggregated, how it should be filtered, etc. AI can assist in this process. 3️⃣ Analysis by AI. Our MCP server allows LLMs (like Claude, chatgpt) to do analysis that would take a human a lot of time to do: https://databox.com/mcp We have much more AI-powered capabilities launching imminently that will make it even easier for our customers to generate insights from their data, as well as setup and use the rest of our product more easily to visualize, monitor and report performance. But, as we roll these things out, I'm very curious: which SaaS companies are doing this well?