Practical guides to building AI agents with Claude Code, Codex, and Runtime. From payment investigations to production operations.
An agent harness is the software around an AI model that gives it a computer, tools, permissions, approvals, and an audit trail. A plain guide for ops, risk, and payments leaders.
A system of intelligence sits above your systems of record and turns every case into memory the next one can use. Origin, examples, and how it works.
A system of record stores what happened. A system of intelligence learns why, and what to do next. A side-by-side comparison with payment examples.
Systems of action do the work. Systems of intelligence decide which work to do and remember how it went. How they fit together, with payment ops examples.
Know Your Agent (KYA) is how you verify an AI agent, the person or business behind it, and what it may do. How KYA differs from KYC and KYB, and how payment teams apply it.
How payment ops teams use AI agents on Stripe: assemble dispute evidence, reconcile payouts to the bank and ledger, and triage Radar reviews, with read-only keys and human approval.
How payment ops teams use AI agents on Adyen: prepare dispute defenses, reconcile settlement batches from Adyen reports, and triage case management reviews, with scoped API roles and approvals.
How platforms on Finix use AI agents for payment ops: review settlements before approval, prepare dispute evidence, and investigate failed payouts, with the Finix API, webhooks, and approvals.
How card program teams use AI agents on Marqeta: dispute case prep, JIT Funding and authorization investigations, card lifecycle requests, KYC follow-up, and DiVA report checks, with approvals.
How card program teams use AI agents on Lithic: dispute prep, authorization and Auth Rules investigations, KYB reviews, ACH return checks, and settlement reconciliation, with human approvals.
How fintech ops teams use AI agents on Unit: KYC and KYB application reviews, ACH returns and payment reviews, card authorizations, disputes, and account freezes, with read-only tokens and approvals.
A plain-English guide for ops managers and analysts: how AI coding agents take over the weekly report, the ERP exports, the trackers, and the reconciliations, with you still holding the sign-off.
Give a coding agent safe, read-only access to your logs, data, and repos, and the technical tickets that used to wait days on engineering close the same day.
Use Jev, TypeSafe's System One model, as a pre-trigger that triages, ranks, and routes support tickets and alerts before a Claude Code or Codex agent investigates in a Runtime sandbox. Covers payment ops, support, and incident response.
Use Jev, TypeSafe's System One model, to triage onboarding applications, KYB and KYC reviews, sanctions and adverse media hits, and AML alerts, then hand the right cases to a Claude Code or Codex agent in a Runtime sandbox.
How compliance, risk, and IT leaders at payment companies, fintechs, and banks manage AI agent risk: agent inventory, guardrails, audit trails, risk scoring, evals, and controls mapped to US, UK, EU, and global frameworks like NIST AI RMF, SR 26-2, NYDFS, the FCA Consumer Duty, the EU AI Act, DORA, and ISO/IEC 42001.
Build a payment support agent with Claude Code or Codex. Trace declined transactions, check account data, and draft evidence-backed replies with Runtime.
AI agent compliance in the US for fintechs and banks: NIST AI RMF, SR 26-2, sponsor bank oversight, GLBA, CFPB, NYDFS, and state AI laws, mapped to controls.
Build AI agents for partner-bank RFIs, hold-harmless requests, and returns of funds. Parse requests, gather evidence, and draft reviewed replies with Runtime.
Which UK rules apply to AI agents at payment firms, e-money institutions, fintechs, and banks: FCA Consumer Duty, SM&CR, SYSC, PRA SS1/23, UK GDPR, and more.
Use Claude Code or Codex to build KYB and KYC agents that check documents, investigate screening matches, and prepare evidence for human onboarding reviews.
Which EU rules apply to AI agents at banks, payment and e-money institutions: AI Act dates after the AI Omnibus, DORA, GDPR Art. 22, EBA, and AML. Plain guide.
Build a compliance investigation agent with Claude Code or Codex. Gather transaction evidence, draft case narratives, and preserve human review in Runtime.
Build a fraud investigation agent with Codex or Claude Code. Trace transaction patterns, gather account context, and prepare evidence for risk review in Runtime.
Build a squad of cybersecurity agents with Claude Code or Codex that monitor Datadog and Google Cloud logs on a schedule, investigate account abuse and threats, and recommend a response for human approval.
Build a payment reconciliation agent with Claude Code or Codex. Explain bank, processor, and ledger differences, then prepare reviewed adjustments in Runtime.
What's actually possible when you point a coding agent at your underwriting playbook: KYB checks, sanctions screening, portfolio reviews, and alert triage, running on a schedule with a human still holding the approval.
A prompt-based guide for GTM engineering and RevOps teams: write the campaign brief yourself, then let a scheduled AI agent source, send, and report on outbound, with a human approving every message.
Build admin dashboards, approval workflows, and data tools in hours instead of weeks. A complete guide using Runtime with Claude Code, Codex, or Gemini CLI.
