Resources
Better ways to work.
More to build on.
Practical guides for the work you repeat. Ideas for making AI work reliable. Notes on what we’re building.
The thinking behind Amphra · PCI
Good work should become
a capability you can keep.
Meet Process-Centric Intelligence: a framework for turning useful results into repeatable, governable work.
Read the introductionFrom the blog
All articlesPractical guides
How to automate data entry from Excel to web forms
Choose the right route, map spreadsheet fields, and handle missing data and duplicate submissions before automating a form.
Practical guides
How to automate competitor research with AI
Build a recurring competitor brief with dated sources, comparable pricing, visible unknowns, and a review step before decisions.
Reliable AI work
Why AI agents drift on the same task—and how to investigate it
Separate harmless variation from real failure, check what changed between runs, and evaluate repeated outcomes instead of one good answer.
Reliable AI work
Why agent skills need governance as work becomes shared
A portable skill can capture a useful method. Shared operational work also needs clear ownership, compatible tools, permissions, and checks.
Reliable AI work
When AI workflows need distributed governance
Put decisions and checks where work happens, while keeping clear authority and a shared record of what actually happened.
PCI foundations
PCI: turning useful AI work into a repeatable capability
An introduction to Process-Centric Intelligence: outcome first, the minimum necessary intelligence, and improvement governed by evidence.
Changelog
Follow the work as it takes shape.
Product progress, preview notes, and what’s available today.