ibl.ai Agentic AI Blog

Insights on building and deploying agentic AI systems. Our blog covers AI agent architectures, LLM infrastructure, MCP servers, enterprise deployment strategies, and real-world implementation guides. Whether you are a developer building AI agents, a CTO evaluating agentic platforms, or a technical leader driving AI adoption, you will find practical guidance here.

Topics We Cover

Featured Research and Reports

We analyze key research from leading institutions and labs including Google DeepMind, Anthropic, OpenAI, Meta AI, McKinsey, and the World Economic Forum. Our content includes detailed analysis of reports on AI agents, foundation models, and enterprise AI strategy.

For Technical Leaders

CTOs, engineering leads, and AI architects turn to our blog for guidance on agent orchestration, model evaluation, infrastructure planning, and building production-ready AI systems. We provide frameworks for responsible AI deployment that balance capability with safety and reliability.

Back to Blog

Anne Boden's Lockwood: A Bank for AI Agents, Not Yet Licensed

Miguel AmigotNovember 19, 2026
Premium

Starling founder Anne Boden is building Lockwood, a planned bank whose current account lets customers' AI agents pay and shop within limits they set. It is not licensed yet: it targets an e-money licence in 2027 and launch in 2028. Here is what was announced and what banks should take from it.

The Short Answer

Anne Boden, who founded Starling Bank, is building Lockwood, a planned bank whose current account lets customers' AI agents pay and shop within limits they set; it is not licensed yet and targets an e-money licence in 2027 and launch in 2028. Incumbent banks need the same per-agent controls, and ibl.ai deploys them where you own all the code and the data.

The announcement is real and the idea is serious. The shorthand circulating about it, "Starling's founder has launched a bank for AI agents", runs ahead of the facts. Lockwood has no licence, no customers and no launch date earlier than 2028.

This post separates what Lockwood has said from what it has not done, then turns to the part a bank can act on now: the controls an AI agent needs before it touches a customer's money.

What is Lockwood, Anne Boden's bank for AI agents?

Lockwood is a banking venture Anne Boden unveiled in an interview with The Sunday Times, reported by the trade and financial press on November 16, 2026. Boden founded Starling Bank in 2014, saw it licensed on July 14, 2016, and stepped down as CEO on June 30, 2023.

As FinTech Futures relayed the interview, Boden describes Lockwood as a bank built by agents, for customers who want to use agents, and defended by agents.

The product is narrow by design. Börsen-Zeitung reports that Lockwood will start with a current account only, with no investment accounts and no loans.

The feature that matters is delegation with limits. Customers would set parameters for each agent, for example a shopping agent restricted to groceries with a spending cap.

The intended uses reported from the interview are routine ones, such as paying bills and managing subscriptions, all within spending limits the customer sets.

Is Lockwood a licensed bank yet?

No. Lockwood has announced a plan and a timeline, not an operating bank. According to Börsen-Zeitung's account of the interview, Lockwood will apply for an e-money licence as soon as possible and then convert into a bank.

Question What has been announced Status on November 19, 2026
Licence Apply for an e-money (EMI) licence, then convert to a full bank Not licensed; EMI targeted for 2027
Launch Open the year after the licence Targeted for 2028; no customers yet
Scale goal One million customers in the first year A target, not a result
Products Current account with per-agent limits No loans or investment accounts planned at first
Funding Self-funded by Boden for now Outside investors sought next year

Börsen-Zeitung reports the licence next year, the launch the year after, the one-million goal and the self-funding; FinTech Futures gives the years as 2027 and 2028. All of them are Boden's stated plans, and plans for a licence depend on the regulator.

The money context is also public. On October 20, 2026, City AM reported, citing Sky News, that Boden had sold her remaining 2.7% stake in Starling at a valuation of £1.7 billion.

What is the difference between an e-money licence and a banking licence for Lockwood's customers?

The difference is how a customer's money is protected. The UK Financial Conduct Authority states that if a non-bank payment provider fails, "your money won't be protected by the Financial Services Compensation Scheme (FSCS)".

Instead, the FCA explains, an e-money institution must protect customer money "by a process known as safeguarding", and it groups e-money firms with non-bank payment service providers.

So during the e-money stage Lockwood would be a regulated payments firm, not a deposit-taking bank. That is a normal path for a UK fintech, and it is also why "bank for AI agents" should be read as a destination, not a description of today.

Why are so few banks running AI agents at scale?

Because the controls are harder than the pilots. Capgemini's World Cloud Report in Financial Services 2026 surveyed 1,100 leaders at financial institutions, published November 12, 2025.

It found 80% of firms in the ideation or pilot stage with AI agents and only 10% with agents implemented at scale. In the same survey, 96% of executives cited a regulatory and compliance burden as a critical roadblock.

The firms that are moving are building supervision, not just agents: 48% of financial institutions said they were creating new roles for employees to supervise agents, and 33% of banks said they were developing their own agents in-house.

That gap between pilot and production is the space Lockwood is betting on. A new entrant can design agent permissions in from the first line of code. An incumbent has to add them to systems that assumed a human at every payment.

What controls does a bank need before an AI agent can move customer money?

Lockwood's own examples describe the minimum: each agent known by name, a scope of what it may buy, and a ceiling on how much. Three more come from supervisors rather than startups.

Federal Reserve Governor Michael Barr said on November 16, 2026 that agents acting in customer accounts need strict authentication, explicit consent for specific actions and technical limits on what they can do. We covered that speech in Fed's Barr on AI agents and bank runs.

The Hong Kong Monetary Authority is exploring "Know Your Agent" practices for verifying an agent's identity and delegated authority, explained in Know Your Customer, Meet Know Your Agent.

Control Who has raised it The record a bank must hold
Agent identity HKMA (Know Your Agent); Barr (strict authentication) Which agent acted, for which customer
Merchant scope Lockwood (groceries-only shopping agent) What the customer allowed the agent to buy
Spending cap Lockwood (spending cap); Barr (technical limits) The limit in force at the moment of payment
Explicit consent Barr (consent for specific actions) When and how the customer granted it
Audit trail All three, implicitly A log the bank can produce without asking a vendor

Every row is a setting someone holds and a log someone keeps. The design question for a bank is whether it holds them, or a vendor does.

Should an incumbent bank build agent controls itself or wait for a bank like Lockwood?

An incumbent does not need a new charter to give customers per-agent limits. It needs the agent layer on infrastructure it controls, so the permissions, consent records and logs sit with the bank that answers for them.

That is how ibl.ai is deployed in financial services: the bank self-hosts the agent platform in its own cloud, on-premise or air-gapped, and it is model-agnostic, so the bank can change models without rebuilding its controls.

Because you own all the code and the data, each agent's scope, cap and log stays in the bank's perimeter, and pricing is usage-based with no per-seat pricing. More detail is on the financial services page and the Agentic OS page.

Lockwood may well prove that customers want to hand agents a spending limit.

The banks that already hold those customers' accounts can offer it first, if they build the controls before Lockwood's planned 2028 launch. For the wider infrastructure argument, see why banks building AI workforces should own the infrastructure.

Why does owning the AI stack matter?

ibl.ai is the agentic AI platform where you own all the code and the data. You self-host the entire stack inside your own perimeter, run it model-agnostic across any LLM and switch anytime, and pay by usage with no per-seat pricing — so you can deploy anywhere: your cloud, on-premise, GovCloud, or fully air-gapped.

  • You own all the code and the data

    Full source code under a perpetual license, running on your infrastructure. Not API access to someone else's platform — the stack itself is yours.

  • Model-agnostic

    Run any LLM — Claude, GPT, Gemini, Llama, Command, or your own fine-tune — and switch providers without rewriting the platform.

  • No per-seat pricing

    Usage-based billing against a budget cap you set. Cost tracks what your organization actually uses, not how many people you employ.

  • Deploy anywhere

    Your cloud, your VPC, on-premise, GovCloud, or a fully air-gapped network with no outbound connectivity.

1.6M+ users across 400+ organizations run the platform this way, including NVIDIA, MIT, and Syracuse University.

ibl.ai is family-owned and operated from New York, NY — a U.S.-headquartered, domestically-owned long-term partner, not a vendor that sells licenses and moves on.

See the ibl.ai AI Operating System in Action

Discover how leading universities and organizations are transforming education with the ibl.ai AI Operating System. Explore real-world implementations from Harvard, MIT, Stanford, and users from 400+ institutions worldwide.

View Case Studies
Work with our team

Pilots, deployment, and full ownership

Every engagement is scoped to the work: you integrate ibl.ai with your own data, deploy it on your own infrastructure, and the engineering hours scale with the scope — so the price tracks the scope, not your headcount. Most customers keep us on afterwards for platform updates, new agents and integrations, and ongoing services.

Start here

Pilot

from $15K

fixed scope · fixed timeline

A time-boxed proof of value on your real data — not a slide deck.

Best for: Teams that want to see ibl.ai working before committing.

  • Deployed on your infrastructure or our cloud
  • 1–2 production agents wired to a slice of your data
  • One integration (LMS / SIS / SSO / data source)
  • Weekly working sessions with our engineers
  • Pilot fee credits toward a full engagement
Scope a pilot
Most common

Integration & Deployment

$25K – $80K

one-time fee · scoped to the work

Full deployment integrated with your data and systems. Engineering hours scale with scope.

Best for: Organizations rolling ibl.ai out across a department, campus, or business unit.

  • Platform deployed in your VPC, on-prem, or air-gapped
  • Integrated with your data + identity (SSO / SAML)
  • Multiple custom agents built to your workflows
  • Engineering hours proportional to scope
  • You own the data · run any LLM you choose
Plan a deployment
Full ownership

Codebase Transfer + Custom AI Engineering

Custom quote

perpetual license · you own the stack

We transfer the full source code. You own and self-host the entire platform — outright.

Best for: Organizations and enterprises that benefit from perpetual ownership and sovereignty.

  • Complete source-code transfer + perpetual license
  • Dedicated AI engineering team on your roadmap
  • Custom agents, models, and integrations to spec
  • Air-gapped capable · zero vendor lock-in
  • Family-owned, New York–based long-term partner
Talk about ownership
You own the code and data Run any LLM — Claude, GPT, Gemini, Llama Family-owned & operated from New York, NY