Databricks Observes Enterprise AI Transition Toward Agentic Systems

According to Databricks, enterprise AI adoption is moving toward agentic systems as organizations implement intelligent workflows.

The initial wave of generative AI promised business transformation but often resulted in isolated chatbots and stalled pilot programs. Technology leaders faced high expectations with limited practical utility. However, recent data from Databricks indicates the market has reached a turning point.

Information from over 20,000 organizations – including 60% of the Fortune 500 – shows a rapid transition to "agentic" architectures where models not only retrieve information but also autonomously plan and execute workflows.

This shift represents a fundamental reallocation of engineering resources. Between June and October 2025, usage of multi-agent workflows on the Databricks platform increased by 327%. This growth signals AI's evolution into a core component of system architecture.

The 'Supervisor Agent' drives enterprise adoption of agentic AI

The 'Supervisor Agent' is driving this expansion. Instead of depending on a single model for every request, a supervisor acts as an orchestrator, breaking down complex queries and assigning tasks to specialized sub-agents or tools.

Since its introduction in July 2025, the Supervisor Agent has become the leading agent application, accounting for 37% of usage by October. This approach reflects human organizational structures: a manager doesn't perform every task but ensures the team completes them. Similarly, a supervisor agent handles intent detection and compliance checks before directing work to domain-specific tools.

Technology companies currently lead this adoption, building nearly four times more multi-agent systems than any other industry. Yet the benefits extend across sectors. For example, a financial services firm might use a multi-agent system to manage document retrieval and regulatory compliance simultaneously, delivering verified client responses without human intervention.

Traditional infrastructure under pressure

As agents progress from answering questions to performing tasks, underlying data infrastructure faces new challenges. Traditional Online Transaction Processing (OLTP) databases were designed for human-speed interactions with predictable transactions and infrequent schema changes. Agentic workflows reverse these assumptions.

AI agents now generate continuous, high-frequency read and write patterns, often creating and dismantling environments programmatically to test code or run scenarios. The scale of this automation is evident in the data. Two years ago, AI agents created just 0.1% of databases; today, that figure stands at 80%.

Furthermore, 97% of database testing and development environments are now built by AI agents. This capability enables developers and "vibe coders" to create temporary environments in seconds rather than hours. Over 50,000 data and AI applications have been developed since the Public Preview of Databricks Apps, with a 250% growth rate over the past six months.

The multi-model standard

Vendor lock-in remains a persistent concern for enterprise leaders as they work to increase agentic AI adoption. The data shows that organizations are actively addressing this by implementing multi-model strategies. As of October 2025, 78% of companies used two or more Large Language Model (LLM) families, such as ChatGPT, Claude, Llama, and Gemini.

This approach is becoming more sophisticated. The percentage of companies using three or more model families increased from 36% to 59% between August and October 2025. This diversity allows engineering teams to assign simpler tasks to smaller, more cost-effective models while reserving advanced models for complex reasoning.

Retail companies are leading this trend, with 83% using two or more model families to balance performance and cost. A unified platform capable of integrating various proprietary and open-source models is quickly becoming essential for the modern enterprise AI stack.

Unlike the legacy big data approach of batch processing, agentic AI operates primarily in real-time. The report notes that 96% of all inference requests are processed immediately.

This is particularly evident in sectors where latency directly affects value. The technology sector processes 32 real-time requests for every single batch request. In healthcare and life sciences, where applications may involve patient monitoring or clinical decision support, the ratio is 13 to one. For IT leaders, this underscores the need for inference serving infrastructure that can handle traffic spikes without compromising user experience.

Governance accelerates enterprise AI deployments

Perhaps the most counter-intuitive finding for many executives is the relationship between governance and speed. Often seen as a bottleneck, rigorous governance and evaluation frameworks actually accelerate production deployment.

Organizations using AI governance tools deploy over 12 times more AI projects into production compared to those that don't. Similarly, companies using evaluation tools to systematically test model quality achieve nearly six times more production deployments.

The reasoning is straightforward. Governance provides essential guardrails – such as defining data usage and setting rate limits – which gives stakeholders the confidence to approve deployment. Without these controls, pilots often remain stuck in the proof-of-concept phase due to unquantified safety or compliance risks.

The value of 'boring' enterprise automation from agentic AI

While autonomous agents might evoke futuristic capabilities, current enterprise value from agentic AI comes from automating routine, mundane, yet essential tasks. The top AI applications vary by sector but focus on solving specific business challenges:

• Manufacturing and automotive: 35% of use cases focus on predictive maintenance.

• Health and life sciences: 23% of use cases involve medical literature synthesis.

• Retail and consumer goods: 14% of use cases are dedicated to market intelligence.

Furthermore, 40% of the top AI use cases address practical customer concerns such as customer support, advocacy, and onboarding. These applications deliver measurable efficiency and build the organizational capability needed for more advanced agentic workflows.

For the C-suite, the path forward involves less focus on the "magic" of AI and more on the engineering discipline surrounding it. Dael Williamson, EMEA CTO at Databricks, notes that the conversation has evolved.

"For businesses across EMEA, the conversation has moved from AI experimentation to operational reality," says Williamson. "AI agents are already running critical parts of enterprise infrastructure, but the organizations seeing real value are those treating governance and evaluation as foundations, not afterthoughts."

Williamson emphasizes that competitive advantage is shifting back toward how companies build, rather than simply what they purchase.

"Open, interoperable platforms allow organizations to apply AI to their own enterprise data, rather than relying on embedded AI features that deliver short-term productivity but not long-term differentiation."

In highly regulated markets, this combination of openness and control is "what separates pilots from competitive advantage."

See also: Anthropic selected to build government AI assistant pilot

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.