Pigment brings intent modeling to enterprise planning

Pigment wants to make enterprise planning feel like vibe coding, but production-ready.

Enterprise planning software has a dirty secret: most of the time companies spend on it isn't spent planning. It's spent building, configuring, and rebuilding the models that make planning possible. Change one assumption in a legacy system, and you might break the process. Need a new scenario? You typically have to call in a specialist and wait weeks.

Pigment, a six-year-old enterprise performance management platform approaching $100 million in ARR, thinks it has a fundamentally better approach. This week, the company launched what it calls the Modeler Agent, a new AI agent that lets teams describe what they want in natural language and spin up production-ready planning models and applications in real time, rather than taking months.

See how Pigment’s Modeler Agent works

The agent is built natively into Pigment, specializes in planning use cases, and has guardrails that large organizations can set.

"The modeler agent drastically cuts the time it takes to get value from your data. Instead of fragile spreadsheets or months‑long EPM implementations, you describe your intent in plain English and Pigment delivers a robust, governed application your business can rely on every week," Ben Previeux, Pigment's head of product and AI strategy, told The Deep View. "With the modeler agent, we’re turning vibe coding into something enterprises can trust. You describe what you want in natural language, and Pigment builds a production‑ready planning application on top of billions of data points in minutes instead of months."

That's the core tension Pigment is trying to resolve. Coding agents have made it remarkably easy to spin up applications from natural language prompts, but the output is still "fragile" for enterprise use, as Previeux said. There's no built-in governance, no role-based permissions, no connection to the production data environments that large companies rely on. 

Pigment's argument is that its platform provides the foundation (a real-time computation engine that handles billions of data points, multi-dimensional modeling, and enterprise-grade security) while the Modeler Agent provides the accessible, conversational interface on top.

The numbers suggest the market is receptive. Pigment says it has doubled ARR for three consecutive years, with 54% of new customers migrating from legacy vendors. The majority of new revenue now comes from enterprises, with customers including ServiceNow, Unilever, and Siemens.

And then there's the detail that may say the most about the vibe coding vs. SaaS debate. Anthropic, the company behind Claude and Claude Code, is itself a Pigment customer. If the builders of one of the most capable coding AI tools on the planet still need dedicated planning software, that's an interesting data point against the "AI agents will replace all SaaS" narrative. It means the platforms capable of adjusting to the needs of today can remain relevant

The Modeler Agent also exposes its capabilities via MCP (Model Context Protocol), meaning developers can interact with Pigment's modeling engine directly from tools like Cursor or Claude Code. It's a smart hedge. Rather than competing with vibe coding, Pigment is positioning itself as the enterprise back end that modelers should build on.

Carta's strategic finance team reported cutting model build times from a week to two days. Non-technical business users have been locked out of model planning for decades. If Pigment can give them an on-ramp, then they will expand the market for who gets to build with enterprise data.

Explore Pigment’s Modeler Agent