Vision & strategy

Raise the operating intelligence of every business

Noviz delivers business-domain AI agents as Software as a Service: shared, secure, and continuously evolving across business domains, while working within the workflows, processes and platforms organisations already run, such as SAP, Microsoft Dynamics, ERPNext, and more.

Watch the demonstration

Every layer of the enterprise stack has become shared

The history of business technology follows one consistent pattern. Enterprises once wrote their own applications, until packaged software made them common. They ran their own data centres, until the cloud made infrastructure shared and elastic. They installed and maintained their own software, until Software as a Service made it a subscription that improves for every customer. Each time, what was built in isolation became a shared service, and the shared model prevailed.

Intelligence is the next layer to make that transition.

Applications, infrastructure and software each moved from built alone to shared. AI Agent SaaS is the next step.

AI agents remain in their custom-build era

Today, each enterprise engineers its own AI agent, against its own systems, with its own team. These agents learn from a single organisation and never from the thousands solving the same problems. Their capability is fixed at deployment while markets, regulation and models continue to advance. Every platform upgrade and every new model generation triggers another round of re-engineering, and the same work is funded again and again, company after company.

Enterprise AI agents remain in their custom-build era: isolated, static, fragile and duplicated.

An isolated agent is at its best on the day it launches. A shared agent behaves differently: every deployment, every reviewed workflow and every model generation advances it for all. The measure of an agent is not its launch capability, but its rate of improvement.

Illustrative capability curve: company-built agents plateau at launch while the shared Noviz agent compounds.

Our thesis: decouple the intelligence from the system of record

Platform vendors such as SAP, Oracle, Microsoft Dynamics and ERPNext should continue to do what they do best: hold the data, the transactions and the processes. Noviz supplies the intelligence that acts upon them: reasoning, workflow knowledge, automation and governance, delivered as SaaS and evolving across business domains.

Intel became the intelligence inside every personal computer, whatever the brand outside. Noviz intends to play that role in business software.

The intelligence layer, Noviz, above the systems of record held by platform vendors.

A business-domain AI agent, delivered as SaaS

Platform-native. The agent operates within existing workflows, processes and platforms, under each user’s own identity and permissions. There is no migration and no disruption to established operations.

Domain-evolving. It learns from business workflows across domains. Every advance in reasoning, coverage, business rules and model capability is released once and reaches every customer.

Delivered as SaaS. Shared, secure and fully managed. Organisations gain enterprise-grade AI capability without in-house AI engineering, model training or maintenance, through a single predictable subscription.

A business-domain AI agent, delivered as SaaS: platform-native, domain-evolving and delivered as SaaS.

Each deployment compounds the intelligence available to all

Every organisation contributes governed, reviewed execution. Recurring patterns across industries become new tools, rules and automations, and those capabilities make the agent more valuable to the next organisation. Scale produces intelligence; intelligence produces scale.

One principle is non-negotiable: patterns are shared, data never is. Tenant data remains isolated, and only reviewed, anonymised operating patterns inform shared capability, with customer consent.

The network effect: new deployment, governed execution, reviewed patterns and new capability for all.

How Noviz is delivered

Noviz operates one central agent as a managed service. Each platform connects to it through a lightweight connector, so the agent works inside the systems an organisation already runs, and every call executes under the user’s own authenticated session. Every improvement is delivered centrally and inherited by every customer, with nothing to install or upgrade on their side.

Governed by architecture, not by prompt

Systems of record demand certainty, so governance is built into the architecture rather than left to instructions given to a model. Every call executes under the user’s own session, so the platform’s native permissions apply exactly as if that person had acted themselves; the agent can never exceed their access. Roles, workflow states and numerical integrity are validated in code before any transaction is committed. Analytics are computed server-side, and bulk records never reach the language model, which sees only what a question requires. Every query, tool invocation and action is recorded per tenant for audit and review.

In production today

Our first production deployment is the Noviz agent within ERPNext, operating on the central Noviz Relay: natural-language access to live business data, reports generated on demand, and transactions executed under the user’s own permissions.

Noviz agent within ERPNext

The same agent now operates within Microsoft Dynamics 365 Business Central. The demonstration below carries out one workflow in ERPNext and then in Business Central, through the same Noviz agent.

One workflow, ERPNext to Business Central

Organisations running ERPNext can add the agent today through Noviz for ERPNext, also available on the Frappe Cloud Marketplace. Organisations on Business Central can request the Business Central Extension.

From assisted operations to autonomous execution

Autonomy is never switched on. It is earned through trusted execution, governance and accumulated operating intelligence. In the first stage, Implement, the agent retrieves, analyses and executes on instruction, and every decision is recorded and reviewed. In the second, Authorize, it recommends actions and executes routine operations within defined approval limits. In the third, Autonomize, industry-trained agents orchestrate end-to-end processes, while people retain ownership of strategy and exceptions. Because the agent is shared, trust earned in one industry shortens the path to autonomy for the next.

The trajectory: Implement, Authorize, Autonomize.

Neutral by design

Noviz does not compete with platform vendors; it makes their platforms more valuable. Any organisation running SAP, Microsoft Dynamics, ERPNext or Odoo can adopt Noviz without altering its platform. Implementation partners extend every engagement with an AI agent and share in the recurring revenue, and Noviz is distributed through the marketplaces where customers already procure extensions for the platforms they run.

Systems of record remain. Intelligence becomes a service.

No organisation should have to replace the systems it trusts in order to become intelligent. That is what Noviz is building.

Read more about the company behind Noviz on our About page, or talk to us.