How to Give AI Agents Enough Context to Be Useful
Summary
AI agents produce poor output not because of capability but because they lack business context—remote companies have an unfair advantage since their knowledge is documented in Slack, calls, and Notion rather than lost in in-person conversations. Sachin from Skyvern scaled to $2M ARR solo by giving agents access to customer recordings, email, and databases, then having them self-critique using frameworks like RICE.
Key Takeaways
- Implement multi-source context retrieval for agents: search call recordings, Slack messages, Notion docs, and customer communications simultaneously to ground AI outputs in evidence-based requirements.
- Use sub-agents for adversarial review and framework-based prioritization (e.g., RICE) to filter low-value outputs. This reduced 'slop' feedback and made team adoption real.
- Remote-first communication structures create AI-accessible knowledge bases automatically; in-person companies must intentionally record all context or lose it. Document everything in searchable tools.
- Build multi-step agent workflows: Skyvern's content marketing agent searches 20 recent calls, buckets topics (pain points, contrary observations, social-worthy), drafts 5 variations, filters sloppy language, adds memes, then sends for human review.
- Agent-generated PRDs that include specific linked evidence (e.g., timestamps from customer recordings) increase team adoption significantly versus generic spec summaries.
Related topics
Transcript Excerpt
I'm Sachin from Skyvern and I'm here to talk to you about how context is all you need or as for a tongue twister, stop the slop shop. So, just for context, Skyvern is a company that helps open source company that helps largely healthcare companies automate things they do manually in the browser. And we managed to scale past 2 million run rate with me basically being a PM, doing all the marketing, doing all the sales and customer support all on my own. And it wouldn't have been possible without all of these agents helping me every day. And so, we have agents that write PRDs, that manage SEO, do content marketing, customer support, and and even fixing small bugs. And I'm going to kind of go through my favorite ones today with you guys. One of the big reasons that agents produce slop, which e…