New York City based startup Emergence AI, founded by former IBM researchers, previously made headlines for its impressive automated system that allows enterprises to type in a requested task in plain natural language and automatically create a fleet of agents to help complete it.

But that’s not all the company has up its sleeve when it comes to automation and AI: today it is launching CRAFT, a new self-serve platform designed to automate enterprise data pipelines, that is, all the enterprise data a user could ever want, organized and made searchable behind the scenes, and retrievable in seconds with the help of Emergence’s AI agents.

Using only plain English, CRAFT allows business users — not just developers — to construct intelligent agent systems that handle tasks traditionally managed by teams of engineers.

Emergence AI designed its platform to be interoperable with a range of leading AI models and agent frameworks. The system integrates with foundation models such as OpenAI’s GPT-4o and GPT-4.5, Anthropic’s Claude 3.7 Sonnet, and Meta’s Llama 3.3. It also supports orchestration frameworks including LangChain, Crew AI, and Microsoft Autogen, allowing enterprises to bring their own models and tools into CRAFT’s agentic workflows with minimal friction.

CRAFT, which stands for Create, Remember, Assemble, Fine-tune, Trust, is positioned as a solution for a global challenge that represents over $200 billion in annual enterprise spend. It replaces brittle, developer-intensive workflows with swarms of self-governing agents capable of building, testing, and running data workflows from a simple prompt.

“This is a big moment,” said Satya Nitta, Co-founder and CEO of Emergence AI. “I’m very excited about the refinement of what we’re trying to do.”

What is Emergence AI and who is behind it?

Emergence AI was founded by veterans of IBM Research and came out of stealth in late 2024 with over $97 million in funding. The company first drew industry attention with the launch of its cross-platform multi-agent orchestrator, designed to operate across enterprise systems from various vendors. Unlike offerings from Microsoft, Salesforce, or Amazon, Emergence emphasized cross-vendor compatibility as a core advantage. Read the full story at VentureBeat.

According to Nitta, “The action space, as I call it, is very hard to generalize. That’s one of the core challenges we’ve unpacked.” Emergence’s orchestrator functions as a meta-agent, dynamically planning and executing tasks across disparate systems using a blend of web automation and secure API integrations.

In early 2025, Emergence followed up with a real-time, no-code AI agent builder capable of generating recursive workflows and domain-specific agents on the fly. Nitta described this as a new phase in enterprise intelligence, where agent systems can not only execute tasks but also create new agents to expand their capabilities. Read more in VentureBeat’s April 2025 article.

Exchange weeks of engineering for a few minutes of automation

With CRAFT, tasks that once required weeks of coding—such as writing ETLs, cleaning datasets, and integrating tools—can now be completed in minutes.

The system’s capabilities go beyond traditional RPA (robotic process automation), introducing features such as self-improvement, planning and reasoning, and long-term memory.

CRAFT is purpose-built for data-heavy environments where rapid decision-making is critical and data is often siloed or fragmented. Key industries include:

• Financial services
• Supply chain and logistics
• SaaS and tech platforms
• Oil and gas
• E-commerce and research organizations

It is particularly suited for organizations requiring complex, real-time orchestration without engineering bottlenecks.

Nitta emphasized the platform’s clarity and intentionality: “What makes this interesting is how focused it is. It’s a very clear story—no messing around.”

He also highlighted the technical depth behind CRAFT’s capabilities. “Orchestration in its entirety is an AI problem. Most people don’t realize how complex it is to ask a system to perform any action in the physical or digital world.”

Early rave reviews from enterprises and industry

CRAFT is already in use by design partners across sectors such as semiconductors, oil and gas, telecom, healthcare, and financial services.

One notable deployment is with NI/Emerson, where the platform is integrated with the NI O+ solution.

Eran Rousseau, VP and Fellow at NI/Emerson, reported that CRAFT enables real-time insights aligned with complex production workflows, improving yields and accelerating issue detection.

“Previously, by the time manual analysis was complete, the next production batch was underway, limiting timely action,” Rosseau said in a statement emailed to VentureBeat by Emergence reps. “With CRAFT, we gain real-time insights aligned with complex test workflows, enabling smart automation to improve yields and detect issues quickly. Integrated with NI’s O+ solution, CRAFT AI generates insights that translate into immediate actions, allowing customers to focus on solving critical challenges rather than on lengthy reports. This represents the future of semiconductor operations.”

Carrol Chang, CEO of Andela, emphasized the importance of mastering data-centric workflows, dynamic data integration, and governance in the AI age.

“Our partnership with Emergence will empower Andela engineers to build the skills needed to thrive in the next chapter: steering intelligent agents, designing resilient systems, and working creatively alongside AI,” Chang wrote in another emailed state from Emergence. “Critically, this includes mastering data-centric workflow, from integrating dynamic data sources to ensuring data integrity and governance, which are foundational to modern enterprise applications. It’s about preparing talent to lead the future, not just adapt to it.”

Other early results include significant productivity gains:

• In semiconductor fabrication, CRAFT has helped identify millions in weekly savings
• A telecom firm saw a 70% reduction in data governance time
• A large online forum reduced 60 million unsafe images per month using AI agents

The tech behind the platform: Agents Creating Agents (ACA)

CRAFT’s engine includes a novel framework called Agents Creating Agents (ACA). This system allows agents to generate, improve, and coordinate other agents in real time—enabling complex multi-agent workflows with minimal input.

The platform’s interface supports both no-code and developer-customized workflows. It features an agent registry for reuse, memory capabilities across sessions, and customizable behavior through feedback or direct editing. Enterprise features include deployment flexibility (on-prem, cloud, or VPC) and built-in compliance controls such as SOC 2, GDPR, and audit logging.

Nitta explained why Emergence has been ahead of the curve on agentic systems: “Everybody else is now talking about agent orchestrations, but we’ve already been thinking deeply about it.”

CRAFT also has implications for workforce transformation. Through a partnership with Emergence AI, Andela is training engineers to work with agentic systems.

Availability and pricing

CRAFT is currently available in private preview as part of a phased rollout. The platform will be offered in three pricing tiers:

• Free – suitable for individual experimentation
• Pro – a cost-based tier with additional capabilities
• Enterprise – custom pricing with advanced governance and infrastructure support

Pro and Enterprise pricing will evolve with user adoption as Emergence AI focuses on accessibility and learning in this early rollout.

Moving from functional user to system architect

According to Nitta, CRAFT represents a step toward democratizing the development of intelligent systems. By enabling non-developers to define and deploy agentic workflows, the platform reimagines how enterprises interact with AI. The goal is to lower the barrier to entry while expanding what’s possible—turning business intent into executable systems without requiring deep technical expertise.

Emergence AI, headquartered in New York with offices in California, Spain, and India, aims to lead the evolution of enterprise AI by building platforms that not only perform tasks—but also create, adapt, and scale autonomously.