Build or Buy: Agents that run and fix your production software
This 45-minute on-demand session breaks down build vs. buy for AI agents that run and fix production software, with Resolve's Dave Lawson (Founding Solutions Engineer) and Ed Li (Technical Staff Engineer).
Where the 70% of engineering time actually goes
What breaks three months into a DIY agent
How agent complexity compounds from team to org
A live build-with demo inside Claude Code
Featuring
Dave Lawson
Founding Solutions Engineer, Resolve AI
Edward Li
Member of Technical Staff, Resolve AI
What you'll learn
Why running software now outweighs writing it
Roughly 70% of engineering time goes to operating production rather than shipping features — and coding agents have made that load heavier, not lighter.
The DIY iceberg
What looks straightforward with a single agent versus what breaks three to four months in: shallow reasoning, token explosion, and governance nobody designed for.
How complexity compounds
The jump from one engineer to a team to a whole org — on-call handoffs, knowledge silos, multi-agent architectures, and access governance.
A live build-with demo
Resolve running inside Claude Code over MCP: a root-cause investigation carried all the way through to a PR, a Linear ticket, and a Slack notification.
Watch the full session
Forty-five minutes on what it actually takes to build agents for production — and when buying is the cheaper answer.