EU AI Act Technical File: Which Parts Can Be Generated?
What the EU AI Act technical file is, why a hand-written copy can drift, and which half a build can generate from each release while named people own the rest.
AI agents, explained from zero
An AI agent reads, decides and acts for you. These guides show how to trust one with real money.
What this topic covers
An AI agent is a program that reads, decides and acts for you, a bit like a new assistant who never sleeps. It can sort invoices, match bank lines or draft a report. The hard part is not making it work once. The hard part is trusting it every day, with real money and real customers.
The articles here follow that path. First, what an agent actually does in a company and where it saves time. Then how you know it is ready: tests that run before every release, answers that point back to their sources, and an agent that says "I do not know" instead of guessing. Then how you keep it safe: give it the smallest set of permissions it needs, keep a record of what it did, and make sure a person can check and stop it.
Some articles cover the rules too, like the EU AI Act, and what they mean for the people who build these systems. Every article starts with a 30-second answer and a few words to know, so you can stop early and still leave with the main idea. No maths and no coding background needed for the first stages.
Start here: three episodes, in order
- Read 1Bank Reconciliation With AI: From 8 Hours to 23 MinutesSee what an AI agent actually does at a bank, on one real task.3 stages · 16 min
- Read 2Testing an AI Agent Before Launch: Go, Wait or BlockLearn how you know an agent is ready before it goes live.3 stages · 16 min
- Read 3AI Agents That Move Money: How to Limit What They Can DoLearn how to limit what an agent can do with money.3 stages · 18 min
All articles on this topic
AI Agent Tests: How to Set the Score Needed to Launch
An eval threshold is the score a new AI agent version must reach before launch. Why a copied one fails, what the record holds, and the loop that earns it.
Know Your Agent: Checking an AI Bot’s Permission to Trade
An investor lets an AI agent trade her security tokens. ERC-3643 checks the holder, not the bot. Here is how an integration following the ERC-8226 draft can check a mandate on each trade.
Running 46 AI Agents: An Illustrative Monthly Cost Model
I deployed about 46 AI agents for Integra’s internal operations. Here is an illustrative monthly cost model: tokens, platform, review and maintenance.
AI Agent Audit Trail: A 14-Field Record for Auditors
A proposed audit-record design for AI agents: 14 fields that capture source evidence, checks, decisions and review, so one output can be investigated later.
Bank Reconciliation With AI: From 8 Hours to 23 Minutes
What bank reconciliation is, why it took one real-estate fund a whole day, and where those 8 hours went before an AI agent brought them down to 23 minutes.
EU AI Act for High-Risk AI Agents: A Build Checklist
AI Act high-risk obligations moved to 2 December 2027 and 2 August 2028; Article 72 monitoring did not. An engineering checklist, not a legal summary.
x402 Agent Payments: Linking Each Payment to Its Approval
x402 handles payment authorization and settlement. Agent apps still need records of delegation, spending policy and outcomes, which alone do not decide liability.
Testing an AI Agent Before Launch: Go, Wait or Block
How a release gate turns an AI agent’s test results into one verdict, ship, hold or block, and why a clean test run can still say “not yet”.
AI Agents That Move Money: How to Limit What They Can Do
Prompt injection is not solved, so contain it. Why the permission check for a money-moving AI agent belongs in the payment system, not in the agent.
Verified Citations for AI Answers: When to Say I Don't Know
Verified citations and evidence-based abstention: the RAG design I shipped inside a regulated fund, where a RAG that abstains beats one that is confidently wrong.
11 AI Agents Inside a $3B Fund: The Base and the Guardrails
I delivered 11 production AI agents to a US real-estate fund managing roughly $3B in assets. Here is the one foundation they shared and the checks built into it.