Developer support, AI agents and building for technical teams
Practical writing on support operations, ticket deflection, AI agent deployment and what we are learning building Inkeep.
Answering Developer Questions at Scale Without Growing Your Support Team
How AI-assisted Tier-1 deflection changes the math on developer support staffing and what the numbers look like after two months.
From Docs Site to Working Agent: What the First Eight Weeks Actually Look Like
A practical timeline for teams moving from interesting to it handled 200 questions last week.
Async-First Coverage: Why Developer Support Cannot Live in a 9-to-5 Schedule
Developers ship at all hours. How small teams use async patterns and AI agents to cover the gaps.
Why Developer Support Is Fundamentally Different From Consumer Support
Developers do not want empathy scripts. They want accurate technical answers with a clear path to the source.
Keeping Your Support Agent Fresh When Your Docs Change Weekly
Stale knowledge causes confident wrong answers. How Inkeep handles continuous doc updates.
Why Developers Do Not Trust AI Answers (and What Changes That)
Source citations are not a nice-to-have. They are the entire trust mechanism.
How to Actually Measure Ticket Deflection Rate (Not the Metric You Are Probably Using)
Most teams measure deflection rate wrong, leading to inflated numbers and bad decisions.
We Closed Our Angel Round: What We Learned Building Inkeep to This Point
We closed $1M in angel funding. This is what we learned about the problem and what we are building toward.
Integrating an AI Support Agent into Discord and Slack: What We Got Wrong First
Bot integrations for developer communities have a specific set of failure modes.
Your Codebase Is a Documentation Source You Are Ignoring
READMEs, inline comments and example files contain answers your docs do not.
The Practical Guide to Automating Tier-1 Developer Support
Which question categories are safe to automate, how to set escalation thresholds, and how to measure whether it is working.
Training an AI Agent on Private Documentation: What We Learned in the First Six Months
Private docs have different challenges: access control, stale content, inconsistent formatting.