Updates
Choose when a Primer change calls for new analysis
Primer records what changed and lets workspace admins decide whether earlier reports need to be marked for a rerun.
A publication from LiteLedger
Field Notes.
Experiments, ideas, and lessons from building
and understanding AI products.
For people curious about how AI products work,
what they cost, and how they feel to use.
The notes
RSSUpdates
Customer signals can travel through MCP
Export customer signals over time, with coverage and source evidence, for use alongside other customer-experience data.
Updates
Customer frustration, with the evidence in view
A new report brings frustration, confusion, success and delight together with message counts, trends and analysis coverage.
Updates
Backup checks that read the data
Backup verification now checks the full selected archive and reports its scope separately from restore rehearsals.
Engineering
Finding the right job for OpenAI’s Decisions API
We traced the outputs of an analysis pipeline and inspected 217 saved annotations to find a useful role for OpenAI’s Decisions API.
No notes of this type have been published yet. Browse all notes.
An open notebook
Follow the questions.
Show the work.
A useful finding. A decision worth explaining. A question we’re still working through. Field Notes is a place for all three.
Some pieces come from building LiteLedger. Others explore the wider world of AI products. Each starts with something worth understanding.
01
Studies
Measured investigations, methods, and findings.
02
Engineering
Decisions and lessons from building AI systems.
03
Perspectives
Ideas and thought experiments about AI products.
04
Updates
News and developments from LiteLedger.
Published when there’s something worth sharing.
Field Notes / LiteLedger