Notes on building an internal agent

llm (15), agents (6), internal-agent (2)

A few weeks ago in Facilitating AI adoption at Imprint, I mentioned our internal agent workflows that we are developing. This is not the core of Imprint–our core is powering co-branded credit card programs–and I wanted to document how a company like ours is developing these internal capabilities.

Building on that post’s ideas like a company-public prompt library for the prompts powering internal workflows, I wanted to write up some of the interesting problems and approaches we’ve taken as we’ve evolved our workflows, split into a series of shorter posts:

(Many of which are currently drafts!)

  1. DRAFT: Logging
  2. DRAFT: Eval support and integration
  3. Skill support
  4. Progressive disclosure and large files
  5. DRAFT: Triggers
  6. DRAFT: Context window compaction
  7. DRAFT: Iterative prompt and skill refinement
  8. DRAFT: Sandboxing Python
  9. DRAFT: Sub-agents

In the same spirit as the original post, I’m not writing these as an industry expert unveiling best practice, rather these are just the things that we’ve specifically learned along the way. If you’re developing internal framesworks as well, then opefully you’ll find something interesting in these posts.