Nuggets

Check out this new Substack post from our Founder & CEO, Alastair Johnson...

We've been selected for FinTech Scotland's Responsible AI Innovation Challenge, joining other fintechs working with financial and professional services partners on responsible AI. We're making sure that when an AI agent acts on sensitive data, every action is tied to who authorised it and on whose authority, so you can trust and control what AI does. The programme puts us alongside the institutions actually setting these requirements, including Barclays, HSBC, Lloyds Banking Group, NatWest, BlackRock, Baillie Gifford, M&G , Equifax, EY, PwC, CMS, Atos and Sopra Steria. Working with them directly is how we build what these organisations actually need. Thanks to FinTech Scotland, the University of Strathclyde and Scottish Enterprise for putting the programme together.

Nuggets has launched the Authority Control Plane, an enterprise platform for defining and enforcing what autonomous AI agents are authorized to do. Sitting between agents and the systems they interact with, the platform evaluates identity, delegated authority, organizational policy and runtime context before permitting, denying or referring an action for human approval. Each decision produces a signed Action Receipt that provides cryptographic proof of the agent, its authority, the requested action and the outcome. #AIAgents #DigitalIdentity #AIGovernance

We're taking on a select group of founding design partners for the Authority Control Plane. If you're already running AI agents and can't yet prove what they're authorised to do, this is for you. It works with your existing stack, LangChain and LangGraph, so there's nothing to rebuild. Founding partners get hands-on access first: register your agents, set what each can do, see the Action Receipts they produce, and work directly with us to shape what we build next. We start with a 30-minute Agent Authority Exposure Review: where your agents can act today, whether you could prove what they were authorised to do, and what closing that looks like for you. If you know anyone who would find this interesting, please tag /reshare 🙏🏽 Link in comments to book.

We've launched the Authority Control Plane. Agent control planes deploy and monitor your agents. Identity says who an agent is; access says what it can reach. None of them tells you whether a specific action was authorised. The Authority Control Plane does. It gives organisations one place to define what their AI agents can do, enforce it the moment they act, and prove every decision. Behind every action is a verified chain: the person, the business and role, and the agent acting for them. Revoke the authority, and the agent's very next action is blocked. Every decision produces a portable Action Receipt, built on open standards: cryptographic proof anyone can verify independently, across any cloud or vendor, without exposing the personal data behind it. No authority, no action. Every action proven. Trust that travels. DM us to find out more. Link to the blog in comments.

New from Nuggets: Right to Work checks. No contract, no minimum spend, no setup fee. £12 a check. Whatever your volume, no commitment. Once your organisation is set up, a check takes seconds. The result, with verification images and passport detail, lands in your inbox as a clear report. Sits alongside our Know Your Employee checks, same platform, same setup. Details and setup link in the comments 👇

Most enterprises can tell you which systems their AI agents can reach. Very few can control, at the moment an agent acts, whether that action is actually allowed, or prove it afterwards. As agents start doing more than respond, that is where projects stall on the way to production. Introducing the Authority Control Plane. It decides what an agent is allowed to do, and proves it. New blog from Seema Khinda Johnson, link in comments.

Thank you for covering the launch of Nuggets Langchain Biometric Update 🙏🏽 If you or your team are using LangChain and LangGraph this is for you ⬇️

The conversation around autonomous AI is shifting from capability to governance. UK-based Nuggets has launched an open-source package that applies authorization and cryptographic proof to every action taken by AI agents, while Innovation Labs has assembled an advisory council of leaders in internet infrastructure, cybersecurity and standards to help shape trusted AI ecosystems. As AI agents begin handling payments, accessing sensitive data and acting on behalf of users, questions around authority, accountability and auditability are becoming just as important as the models themselves. #AgenticAI #DigitalTrust #AI

Don't just give your AI agents access. Prove they have authority. We just shipped langchain-nuggets, for LangChain and LangGraph developers. Your AI agents can call every tool you give them, for any user, as many times as the model decides. The moment one of those tools moves money, changes records, or acts for a partner, access is not enough. langchain-nuggets checks every tool call before it runs: who's acting, on whose authority, and whether it's allowed. Anything out of scope is blocked before it executes. Every allowed action carries a receipt a partner or regulator on a different cloud can verify for themselves, without access to your systems. Proofs reference hashes, not raw data. Open source. Python and TypeScript. pip install langchain-nuggets npm install @nuggetslife/langchain-nuggets Full write-up in the comments. Enjoy ❤️