Questions about Promenaut

Promenaut is the operating system for a governed digital workforce.

It runs teams of AI agents in production, composed into teams and workflows that do real work across the business, with people deciding where the workflow says.

It ships modules that already do the work, and your teams build their own agents, workflows and modules on the same platform.

Regulated enterprises that need AI doing accountable work across the business, not only answering questions. Financial services is the focus, banks, insurers and asset managers, with capital markets first, because it is the hardest governance environment there is.

Yes. Promenaut is running in production with leading industry partners, including a global tier 1 bank.

HSBC is a strategic investor in Promenaut and a design partner.

Governance is produced by the work itself, not added afterwards.

An agent reaches knowledge as a deputy of the person who launched the work, within that person’s permissions.

Work can be held for a person’s decision, and in a domain that requires it, an agent’s change to knowledge waits for a checker who did not launch the work.

A run’s cost is estimated before it starts, against the budget of the business unit that runs it.

Each workflow run ends with a verdict on whether the agents that ran matched those it launched with, recorded in a tamper-evident ledger.

Promenaut can run in your own cloud account from a shipped kit, including an environment with no outbound internet access.

Agents run on models from Anthropic, OpenAI and Google directly, through Amazon Bedrock, or through your own AI gateway, and each agent has a fallback that can come from another provider.

It is designed not to. Promenaut is neutral to the model and to where it runs.

When it runs in your own cloud account, your data stays there, and your module code is yours to read and change.

Any agent can be shown as the code an engineer would write for it, whichever way it was built.

Yes. Your teams build agents, teams, workflows and whole modules, in the interface or in code with the SDK, and what they build runs under the same governance as what ships.

Engineers get one-command scaffolding, a test harness that needs no database, and a command to run one agent against the real platform.

Frameworks and runtimes are how you build and host agents yourself: the parts, and somewhere to run them.

Promenaut ships a workforce that already does the work, with approval, budgets and run evidence around it, and gives your engineers the tools to build their own on it.

Suite agents are governed inside one vendor’s suite, over that suite’s data.

Promenaut can run in your own cloud account, on the model providers you choose, and the modules your teams build stay yours.

It is not a model vendor: it runs on the models you choose.

It holds work for a person where the workflow says, rather than proceeding on assumption, and an operator can pause an agent so it is refused at its next dispatch, or stop a running workflow.

The workforce is delivered as modules, installable packages of agents, teams and workflows chosen per organisation, beside the ones your teams build.

Pricing is tailored to the deployment and the modules in scope. Get in touch for a conversation.