Post by Lukas Mehnert

AI in B2B SaaS: Real talk from the trenches After 3 months of in-depth AI conversations with my B2B SaaS clients, here’s what’s actually working in B2B SaaS, and what’s not. 1. The “Master Prompt” is your Company Brain Shout-out to Matej Stefanik from Kentico for a great session on building a structured Company Brain prompt. What to include? – Boilerplate, positioning, ICP, personas – Battle cards & key objections – Messaging do’s/don’ts, tone, compliance notes – ”How would our CEO/CTO/CMO react to X?” When every prompt includes this structured context, you stop getting generic fluff and start getting on-brand, usable content. Still, you should do a double-check. A different approach, or manually copying the master prompt? Building your company's brain is to add a company-wide persistent memory over MCP (with versioning and time stamps). Jakub Dubec is seeking individuals who wish to exchange experiences with him. They've built a second brain, based on the Knowledge Graph Memory Server, to observe daily operations and update the context on the go for Claude. 2. Basic automation still matters I’ve been running marketing automation since 2013. The core job remains: to move clean data between systems, enable timely actions, and keep humans in the loop from the outset. It’s not about the platform. It’s about whether your team can actually run it. Remember, many AI automations fail to fulfill their goals. 3. SDR assist that actually helps One of the best AI tools for SDRs (from Kentico) I saw used this data: – Put in target account details (industry, tech, maturity) – The actual people (decision group, LinkedIn profiles) – Your case-study library and product capabilities Then, it drafted tailored first-pass outreach and personalized showcases. Caveat: Hallucinations still happen. Always review before sending. AI works best when it’s in the hands of people who understand the buyer, the process, and the real work. What’s actually working for you with AI right now? What blew up? What saved time?