01

What is Service as Software?

Service as Software is software that completes a unit of work and hands back the result, rather than presenting a person with the information required to do it themselves. The term is doing a lot of marketing work it has not earned — a chatbot on top of a dashboard is not it, and neither is a recommendation a person still has to carry out.

  • SaaS leaves the labour with you. You buy the platform; your people still read, correlate, decide, type and follow up.
  • Service as Software hands back the result. The approval granted, the record updated, the incident resolved, the report delivered — confirmed and logged.
  • People move from doing to governing. Engineers and managers set the permissions, approve the exceptions and audit the record, instead of running the queue.
The simple test

At the end of the loop, is there work no human had to do?

If the answer is no, it is a tool with a new label. If the answer is yes, the next question is governance.

02

What actually changes

The shapes of the two models. The unit of value moves from seats to outcomes, and the failure mode moves from shelfware to ungoverned action — which is why governance is most of this page.

How SaaS compares with Service as Software across five dimensions.
Dimension
SaaS
A tool your team operates
Service as Software x101
Work the system completes, governed
What you buy
A platform to operate Work the platform completes
Who does the work
Your people, assisted Agents execute; people govern and verify
Unit of value
Seats and features Completed actions, evidenced answers
What scales
Licences; the queue still scales with people Capacity to absorb repetitive work
The failure mode
Shelfware nobody adopts Designed out: every action scoped, approved, audited

03

Every capability is a complete service stack

Ask what you are actually buying. With x101, each capability arrives as the entire stack that gets that work done — interface, intelligence, integrations and administration together — not a toolkit your team has to assemble.

Knowledge

Company Knowledge

The entire knowledge service: RAG retrieval, document management, conversational access, and every answer cited to its source — with administrative control over what is indexed and who sees it.

Analytics

Instant Analytics

Plain-language questions answered from live databases: query understanding, permission-checked connectors, and the result charted and explained — delivered, not exported for someone to finish.

Actions

Act

Completed transactions across your systems: connectors, policy gates, approvals where required, and every step confirmed back from the system of record.

Research

Deep Research

The full analyst workflow as one service: gathering across internal sources and the web, verification, synthesis, and a citable, boardroom-ready report at the end.

Orchestration

Intelligent Flow Orchestrator

The coordination layer as a service: intent understood, the work broken down, specialist agents dispatched across apps, data, documents and know-how in parallel — and the results assembled into one governed outcome.

Build your own

Agent Studio

New service stacks, built by your teams: a visual builder with testing, memory and knowledge included — and governance inherited from the platform, not re-implemented per agent.

Every stack ships on the same foundation: governance, immutable audit, explainability and administrative control, built in. That is the difference between a feature and a service.

04

Governance, audit, explainability: part of every service

These are not separate products on a price list. Every service stack above runs on them as integral parts — they are how each service works, enforced in code on every request, not options added to it.

  • Permission checks before every action

    Every system call and agent handoff is cleared against the asker's entitlements before it runs — not logged after it ran.

  • A held decision above the limit

    Autonomous limits are set by you, per action and per amount. Anything above them is held and raised for human sign-off, with the evidence pack prepared.

  • Guardrails on the way in and out

    Requests are screened before they run, and sensitive data is protected in what comes back.

  • Audit and explainability, on the record

    Each decision explained in words anyone can read, cited to its source, and written to an immutable audit trail — per user, per agent, per action, with cost tracked.

None of this can be unbundled or switched off — it ships inside every service, not beside it. Autonomy is a set of permissions your organisation grants deliberately — action by action, limit by limit — and expands as the evidence record justifies it. Governance is not the brake on the model; it is the reason the model can be approved at all.