01
Studio
The author's surface. Where the person closest to the problem builds the Workblock: with real data, against a real schema, in-room with the client.
Read morePlatform
The execution layer for enterprise work.
Beyond Work is where intelligence meets the systems that run the enterprise. Workblocks are the unit. Studio is where they're authored. Flow is where they run. Library is where they accumulate.
01The four surfaces, and the framework
02
Workblocks
The unit of execution. A typed, versioned, schema-bound artefact that takes inputs, runs intelligence over them, and produces outputs downstream systems can use.
Read more03
Flow
The runtime. Where Workblocks execute in production, with human-in-the-loop as a native state and full audit captured by default.
Read more04
Library
The accumulation. Workblocks authored at one client become reusable across the network. The Library is where the compounding happens.
Read more05
GroundWorks
The framework. Source-to-pay built on the platform and delivered with a partner: five stages, nine domain Workblocks and one canonical record model across every spend type.
Read more02Composer & Conductor
Two specific surfaces inside the platform sit alongside Studio:
01
Composer
The document surface: where contracts, memos, legal positions, and structured documents are authored against typed schemas. It is to Word what Studio is to bespoke engineering: a way for the person doing the work to capture it in a form the system understands.
Read more02
Conductor
The spreadsheet surface: where models, analyses, and structured calculations live with the same schema rigour. Composer and Conductor will be externally announceable at v1.5. A short note for the curious, not the headline of this page.
Read more03Properties
Fully model-agnostic. Workblocks are bound to schemas, not to models. Anthropic, OpenAI, and open-weight models like Mistral and Llama are interchangeable at the Flow layer, and a Workblock can run one model for reasoning and a cheaper one for extraction in the same job. Deploy on AWS, Azure, or the customer's own enterprise licenses: the model choice and the data both stay theirs.
Single-tenant by default. Customer data does not leave the customer's environment. The platform deploys into the customer's cloud, not ours.
Human-in-the-loop as a native runtime state. Not a fallback. Not an exception. A first-class state that Workblocks declare and Flow enforces.
Versioning and channels. Workblocks have semantic versions. Channels (draft, staging, production) gate which version runs where. Reverting is a query.
Audit by default. Every input, every output, every model call, every human approval, recorded as an event-sourced timeline. The audit is the runtime, not a sidecar.
04Principles
01
How we think.
The platform is shaped by how we believe enterprise intelligence has to work. Five principles, written down.
Methodology02
How we operate.
We sell into industries that got there by being careful. We carry the same posture. SOC2, ISO, EU AI Act registration.
Trust03
How it fits.
One layer above the tools you already run. Beyond Work does not replace your landscape: it is the orchestration, document intelligence and workflow layer that connects it.
Fit