A build log for one real Tamaton custom agent that triages the inbox, checks the calendar and docs, drafts replies, and files attachments — plus the three failure modes we fixed.
A concrete look at building RAG over your own email, docs, and drive — and why retrieval quality, not model size, decides whether AI knowledge work is actually useful.
Coding demos are graded by a test suite. Email is graded by your boss, your customer, and your calendar. Here's why the inbox is the hardest honest test of an AI agent.
Reasoning and planning demos are easy. Proving a multi-step agent actually finished the job — correctly, safely, once — is the unsolved part. Here's how to measure it.
Email was built for eyeballs, so agents waste context untangling HTML soup and quoted replies. Here's what an inbox designed for machines looks like — and why the human UI should be just one renderer.
A defensible framework for measuring whether an AI tool actually returns hours — including verification overhead, context-switching, and the trust tax nobody counts.
The most reliable multi-step agents aren't the smartest — they're the ones handed structured places to stash state, permissions, and checkpoints. Here's how to build those places.
A practical taxonomy of AI agents delegation built on reversibility, verifiability, and blast radius — five tiers from 'let it run' to 'don't, ever.'
Stop writing clever prompts and start writing employment contracts. A concrete walkthrough of scoping a Tamaton custom agent with permissions, escalation thresholds, and an audit trail.
Identical agents don't cooperate by default — they clash, collude, and coordinate in ways nobody specified. That's an architecture problem, and it has fixes.
Prompt and model tweaks that pass every smoke test can quietly wreck real outcomes. Here's how to evaluate agents on results instead of vibes.
Public benchmarks won't tell you if a model can triage your inbox or reconcile your spreadsheet. Here's how to build a private, brutally specific eval set in an afternoon.
Get started
Claim your address before someone else does — free to start, with an AI-native inbox built in.