With $20M, Thread AI Wants to Make AI Workflows Actually Work

In a business environment where AI promises to unlock productivity, many enterprises still find themselves stuck with one of two unsatisfying options: adapt to rigid, off-the-shelf AI tools that don’t reflect their workflows, or spend heavily to build custom systems from scratch. Thread AI, a company founded by Palantir veterans Angela McNeal and Mayada Gonimah, believes there’s a third path, and they aim to prove it with their recent $20 million Series A raise it.

Bridging the AI Infrastructure Gap

“There’s a frustrating dilemma in the market,” Thread AI wrote in its blog post announcing the funding. “Settle for rigid, pre-built AI applications… or invest heavily in talent, infrastructure, and resources to build custom AI workflows from scratch.”

Thread AI’s solution is Lemma, a platform designed to help enterprises build and deploy AI-powered workflows without discarding their existing tech stacks. The platform is composable: users can mix and match models, APIs, and business logic into secure, scalable pipelines that actually map to the complexity of their organizations.

“Thread AI offers a flexible platform that lets organizations build AI workflows suited to their operations, without compromising scalability or security,” said CEO Angela McNeal in the company’s press release.

Thread AI emerged from stealth in late 2024 with $6 million in seed funding and a product shaped by the co-founders’ experience leading machine learning products at Palantir. Their background isn’t just technical—it’s logistical. In one formative example, Gonimah built infrastructure for Department of Defense projects she wasn’t legally cleared to access due to her citizenship status. “She couldn’t see, she couldn’t access, and had to trust me to communicate back the requirements and the product vision,” McNeal recalled to Fortune.

That experience helped shape Thread AI’s ambition: to create infrastructure that can securely bridge data, tools, and organizational silos. Now, the company serves customers in agriculture, manufacturing, hospitality, and financial services, including Fortune 500s.

Differentiators in a Crowded Space

While many AI infrastructure startups focus on developer-first tools or single-purpose agents, Thread AI’s approach is intentionally broad and layered. Lemma supports both deterministic and non-deterministic workflows, making it usable for everything from robotic process automation to agentic AI systems. This choice enables cross-functional collaboration: “AI workflows are fundamentally cross-functional,” noted Head of Engineering Martin McRoy in a blog post, “involving data scientists, integrators, business users, and more.”

A core part of the platform is its Function Registry, which allows secure, stateless execution of business logic using gRPC, OpenAPI, or REST protocols. Each function includes credential configuration and version control, minimizing security risks and operational complexity. This registry becomes especially important in industries with strict regulatory requirements.

Thread AI hasn’t completely solved the fundamental challenge of AI in the enterprise: unpredictability. While its architecture allows for what McRoy calls “bounded non-determinism,” even controlled AI workflows can produce surprising results. That’s why Lemma emphasizes traceability and auditability, particularly with primitives like human review points and rollback options.

The company also acknowledges broader concerns about AI and labor. “If AI is going to change the way we work, we must use this opportunity to elevate human capabilities,” the team wrote. The inclusion of human-in-the-loop features suggests a deliberate move away from full automation in favor of augmented workflows.

Backed by Greycroft, Index Ventures, Scale Venture Partners, Meritech Capital, Plug and Play Tech Center, and Homebrew, Thread AI plans to use the Series A funding to grow its commercial team and push deeper into enterprise sales. “The problem that we’re trying to solve is top of mind for every company right now, how to securely and safely modernize your business with AI,” McNeal told Fortune..

Mark Terbeek of Greycroft put it plainly: “Enterprises need a secure, scalable infrastructure that evolves with their business. Thread AI provides a powerful, adaptable platform that helps organizations implement AI at scale.”

Thread AI’s idea seems simple: They want to make AI work within the messy realities of legacy systems and fragmented data.