Databricks Unveils Agent Bricks: Revolutionizing Enterprise AI Agent Development and Deployment

Why most enterprise AI agents never reach production and how Databricks plans to fix it

Join the event trusted by enterprise leaders for nearly two decades. VB Transform brings together the people building real enterprise AI strategy. Learn more Many enterprise AI agent development efforts never make it to production and it's not because the technology isn't ready. The problem, according to Databricks, is that companies are still relying on manual evaluations with a process that's slow, inconsistent and difficult to scale. Today at the Data + AI Summit, Databricks launched Mosaic Agent Bricks as a solution to that challenge. The technology builds on and extends the Mosaic AI Agent Framework the company announced in 2024. Simply put, it's no longer good enough to just be able to build AI agents in order to have real-world impact. The Mosaic Agent Bricks platform automates agent optimization using a series of research-backed innovations. Among the key innovations is the integration of TAO (Test-time Adaptive Optimization), which provides a novel approach to AI tuning without the need for labeled data. Mosaic Agent Bricks also generates domain-specific synthetic data, creates task-aware benchmarks and optimizes quality-to-cost balance without manual intervention. Fundamentally the goal of the new platform is to solve an issue that Databricks users had with existing AI agent development efforts. "They were flying blind, they had no way to evaluate these agents," Hanlin Tang, Databricks' Chief Technology Officer of Neural Networks, told VentureBeat. "Most of them were relying on a kind of manual, manual vibe tracking to see if the agent sounds good enough, but this doesn't give them the confidence to go into production." From research innovation to enterprise AI production scale Tang was previously the co-founder and CTO of Mosaic, which was acquired by Databricks in 2023 for $1.3 billion. At Mosaic, much of the research innovation didn't necessarily have an immediate enterprise impact. That all changed after the acquisition. "The big light bulb moment for me was when we first launched our product on Databricks, and instantly, overnight, we had, like thousands of enterprise customers using it," Tang said. In contrast, prior to the acquisition Mosaic would spend months trying to get just a handful of enterprises to try out products. The integration of Mosaic into Databricks has given Mosaic's research team direct access to enterprise problems at scale and revealed new areas to explore. This enterprise contact revealed new research opportunities. "It's only when you have contact with enterprise customers, you work with them deeply, that you actually uncover kind of interesting research problems to go after," Tang explained. "Agent Bricks....is, in some ways, kind of an evolution of everything that we've been working on at Mosaic now that we're all fully, fully bricksters." Solving the agentic AI evaluation crisis Enterprise teams face a costly trial-and-error optimization process. Without task-aware benchmarks or domain-specific test data, every agent adjustment becomes an expensive guessing game. Quality drift, cost overruns and missed deadlines follow. Agent Bricks automates the entire optimization pipeline. The platform takes a high-level task description and enterprise data. It handles the rest automatically. First, it generates task-specific evaluations and LLM judges. Next, it creates synthetic data that mirrors customer data. Finally, it searches across optimization techniques to find the best configuration. "The customer describes the problem at a high level and they don't go into the low level details, because we take care of those," Tang said. "The system generates synthetic data and builds custom LLM judges specific to each task." The platform offers four agent configurations: Agents are great, but don't forget about data Building and evaluating agents is a core part of making AI enterprise ready, but it's not the only part that's needed. Databricks positions Mosaic Agent Bricks as the AI consumption layer sitting atop its unified data stack. At the Data + AI Summit, Databricks also announced the general availability of its Lakeflow data engineering platform, which was first previewed in 2024. Lakeflow solves the data preparation challenge. It unifies three critical data engineering journeys that previously required separate tools. Ingestion handles getting both structured and unstructured data into Databricks. Transformation provides efficient data cleaning, reshaping and preparation. Orchestration manages production workflows and scheduling. The workflow connection is direct: Lakeflow prepares enterprise data through unified ingestion and transformation, then Agent Bricks builds optimized AI agents on that prepared data. "We help get the data into the platform, and then you can do ML, BI and AI analytics," Bilal Aslam, Senior Director of Product Management at Databricks told VentureBeat. Going beyond data ingestion, Mosaic Agent Bricks also benefits from Databricks' Unity Catalog's governance features. That includes access controls and data lineage tracking. This integration ensures that agent behavior respects enterprise data governance without additional configuration. Agent Learning from Human Feedback eliminates prompt stuffing One of the common approaches to guiding AI agents today is to use a system prompt. Tang referred to the practice of 'prompt stuffing' where users shove all kinds of guidance into a prompt in the hope that the agent will follow it. Agent Bricks introduces a new concept called - Agent Learning from Human Feedback. This feature automatically adjusts system components based on natural language guidance. It solves what Tang calls the prompt stuffing problem. According to Tang, the prompt stuffing approach often fails because agent systems have multiple components that need adjustment. Agent Learning from Human Feedback is a system that automatically interprets natural language guidance and adjusts the appropriate system components. The approach mirrors reinforcement learning from human feedback (RLHF) but operates at the agent system level rather than individual model weights. The system handles two core challenges. First, natural language guidance can be vague. For example, what does 'respect your brand's voice' actually mean? Second, agent systems contain numerous configuration points. Teams struggle to identify which components need adjustment. The system eliminates the guesswork about which agent components need adjustment for specific behavioral changes. "This we believe will help agents become more steerable," Tang said. Technical advantages over existing frameworks There is no shortage of agentic AI development frameworks and tools in the market today. Among the growing list of vendor options are tools from Langchain, Microsoft and Google. Tang argued that what makes Mosaic Agent Bricks different is the optimization. Rather than requiring manual configuration and tuning, Agent Bricks incorporates multiple research techniques automatically: TAO, in-context learning, prompt optimization and fine-tuning. When it comes to agent to agent communications, there are a few options in the market today, including Google's Agent2Agent protocol. According to Tang, Databricks is currently exploring various agent protocols and hasn't committed to a single standard. Currently, Agent Bricks handles agent-to-agent communication through two primary methods: Strategic implications for enterprise decision-makers For enterprises looking to lead the way in AI, it's critical to have the right technologies in place to evaluate quality and effectiveness. Deploying agents without evaluation isn't going to lead to an optimal outcome and neither will having agents without a solid data foundation. When considering agent development technologies, it's critical to have proper mechanisms to evaluate the best options. The Agent Learning from Human Feedback approach is also noteworthy for enterprise decision makers as it helps to guide agentic AI to the best outcome. For enterprises looking to lead in AI agent deployment, this development means evaluation infrastructure is no longer a blocking factor. Organizations can focus resources on use case identification and data preparation rather than building optimization frameworks.

Databricks Introduces Agent Bricks to Build Production-Ready AI Agents on Enterprise Data | AIM

Databricks also launched MLflow 3.0, a redesigned version of its AI lifecycle management platform. At the Data + AI Summit, Databricks announced the launch of Agent Bricks, a new offering that allows businesses to build and deploy AI agents using their own data, without the need for manual tuning or complex tooling. Available in beta starting today, Agent Bricks is positioned as an automated system that transforms a high-level task description and enterprise data into a production-grade agent. Ali Ghodsi, CEO and co-founder of Databricks, described it as "a whole new way of building and deploying AI agents that can reason on your data." He added, "For the first time, businesses can go from idea to production-grade AI on their own data with speed and confidence, with control over quality and cost tradeoffs." Agent Bricks automates the entire process of AI agent development. It uses research developed by Mosaic AI to generate synthetic data tailored to a customer's domain and builds task-specific benchmarks to evaluate agent performance. The system then runs a series of optimisations, allowing users to choose the version that best balances accuracy and cost. The result is a deployable agent that operates with consistency and domain awareness. The platform supports a range of use cases across industries. An Information Extraction Agent can convert unstructured content like PDFs and emails into structured fields such as names and prices. A Knowledge Assistant Agent provides accurate, data-grounded answers to user queries, reducing the kind of errors often seen in traditional chatbots. The Multi-Agent Supervisor allows coordination between multiple agents and tools like MCP to manage workflows, including compliance checks and document retrieval. Meanwhile, a Custom LLM Agent handles specific text transformation tasks, such as generating marketing content that aligns with an organisation's brand voice. Databricks said the product addresses a key issue in the AI agent space, which is that most experiments fail to reach production due to a lack of evaluation standards, inconsistent performance, and high costs. According to the company, Agent Bricks resolves these challenges by offering domain-specific, repeatable, and objective evaluations, all within a workflow that requires no stitching together of multiple tools. Early adopters are seeing results across sectors. AstraZeneca used Agent Bricks to extract structured data from over 400,000 clinical trial documents without writing any code. Joseph Roemer, head of data & AI at the company, said they had "a working agent in just under 60 minutes." At Flo Health, the tool helped improve the medical accuracy of AI systems while meeting internal standards for safety and privacy. "By leveraging Flo's specialised health expertise and data, Agent Bricks uses synthetic data generation and custom evaluation techniques to deliver higher-quality results at a significantly lower cost," said Roman Bugaev, the company's CTO. The announcement was accompanied by the release of two additional tools. Databricks now offers support for serverless GPUs, giving teams access to high-performance compute without the operational burden of managing GPU infrastructure. This enables users to fine-tune models and run AI workloads on demand. Databricks also launched MLflow 3.0, a redesigned version of its AI lifecycle management platform. Tailored for generative AI, MLflow 3.0 includes prompt management, human feedback loops, LLM-based evaluation and integration with existing data lakehouses. The new version allows teams to monitor and debug AI agents across any platform and is downloaded over 30 million times a month. According to Databricks, the combination of Agent Bricks, serverless GPU support, and MLflow 3.0 makes its platform the most complete environment for building, tuning and deploying enterprise-grade generative AI systems.

Building for a new era: Databricks takes on pain points of complexity, lock-in and cost for enterprise AI - SiliconANGLE

Building for a new era: Databricks takes on pain points of complexity, lock-in and cost for enterprise AI Ali Ghodsi, co-founder and chief executive of Databricks Inc., believes that data and artificial intelligence estates are fragmented, costly and much too complex. Today, he explained how his company's flurry of announcements will address these issues and more. "It's actually really hard still to succeed with data and AI," Ghodsi (pictured) said during his keynote remarks today at the company's Data+AI Summit in San Francisco. "It's a complexity nightmare of high costs and proprietary lock-in. It's slowing down the organizations. We think this is the biggest problem." Part of the solution will involve AI agents and an ability to simplify their development for enterprise use. Today Databricks unveiled Agent Bricks, a unified workspace that automates agent building and optimization using customers' enterprise data and synthetic equivalents. A key part of the release involves the use of large language model automated "judges" that generate questions and expected answers to assess model performance. This is presumably to resolve situations such as the one described by Ghodsi where one automaker expressed concern to him about an agent that was recommending a competitor's cars. "This maps to the business problems you have," Ghodsi said. "You get an agent that can basically evaluate. We will optimize an AI system that does really well on the LLM judge benchmark." The other major announcement from Databricks today involved the database world. The company's release of Lakebase, a managed Postgres database built for AI, added an operational layer to the firm's Data Intelligence Platform. Databricks' participation in the operational database or OLTP world is designed to address what the company's believes to be an outmoded, inflexible model. "If you look at these databases, they were really built for a different era," Ghodsi said. "There's just this massive lock-in of data in traditional databases. It's such a sticky thing." Lakebase builds on the company's acquisition of Neon Inc., which was announced in May. Neon's serverless PostgreSQL platform allows developers to add support for data structures where AI models keep information. "We think this is going to be the future of all databases," Ghodsi said. "Postgres has won. You really should properly separate compute and storage in that database using Postgres. It should be built for the AI era." Today's announcements were designed to address what Databricks has been hearing about AI deployment from its extensive enterprise customer base. Cost remains in issue, and data management is a significant part of the equation. This was echoed during an appearance by the chief executive of financial giant JPMorgan Chase and Co., Jamie Dimon, who told attendees that his company now spends about $2 billion per year on AI. "The hardest part is the data, it isn't the AI," Dimon said. "Getting the data in the form that it can be usable is the hard part." Making data usable will require automation, based on the sheer size and complexity of the information involved. Among the many announcements from Databricks was an offering called Lakeflow Designer, a no-code capability that allows users to author production data pipelines by using a drag-and-drop interface and a natural language generative AI assistant. It's the latest entry in the field of "vibe coding," a recently coined description where users give into the "vibes" of AI and express their intentions without needing to know how to write a single line of code. "It's kind of like vibe coding if you are not coding at all, we call it vibe designing," Ghodsi said during a briefing for the media following his keynote address. "It's a new market for us, a new market opportunity." Tools such as Lakeflow Designer foreshadow a major sea change in how software will be designed and deployed for enterprise applications. During a brief appearance at the Databricks keynote session, Neon's Chief Executive Nikita Shamgunov described how AI agents were creating four times more databases than humans on his company's platform. "I think we are at the dawn of the AI software revolution," Shamgunov said. An example of this trend can be seen in Goose, an interoperable open-source AI agent framework developed at the digital finance firm Block Inc. Goose not only automates code generation, it can enable non-programmers to prototype new apps for features. It was developed by Block using Anthropic Inc.'s Claude model, employed by developers for coding and tool use. In his remarks at the Databricks conference, Anthropic co-founder and CEO Dario Amodei acknowledged that virtual collaborators will soon become commonplace. "Coding is moving the fastest," Amodei said. "It's a foreshadowing of what's going to happen across all applications." A byproduct of this could be that the notion of devices as disposable, replaceable resources, often characterized as "cattle, not pets," could soon apply to enterprise software as well. Software that once ran enterprise tasks for years could suddenly be replaced by a model where key new pieces are added and then removed within a matter of hours. "There are going to be new paradigms for how software is produced and delivered," Matt Dugan, vice president of data platforms for AT&T Inc., said in an interview with SiliconANGLE. "These are disposable. I wonder if we're heading into a spot where we start to think of software assets the same way. It's a change." Through tools such as Agent Bricks and Lakebase, Databricks is building the infrastructure to support this change in how software is created and deployed. It's a game plan that has proven to work well for the 12-year-old company. Ghodsi noted that despite raising $10 billion late last year on a valuation of $62 billion, the privately held Databricks is not dependent on external capital, and it has plans to hire 3,000 employees this year. "What's our secret sauce?" Ghodsi said. "We have the customers, and we have the data. That's the differentiator."

New Databricks tools aim to cut costs and complexity in AI agent deployment - SiliconANGLE

New Databricks tools aim to cut costs and complexity in AI agent deployment Databricks Inc. today unveiled a new suite of tools aimed at simplifying the development of artificial intelligence agents for enterprise use. The centerpiece of the announcement is Mosaic Agent Bricks, a unified workspace that automates agent building and optimization using customers' enterprise data and synthetic equivalents. Agent Bricks enters public beta at the company's annual Data + AI Summit today in San Francisco, alongside new offerings that include serverless graphic processing unit support and MLflow 3.0, Databricks' latest platform for managing machine learning and generative AI applications. Databricks also announced updates to its Unity Catalog that expand support for Apache Iceberg -- the open-source table format designed for large analytical datasets in data lakes -- and introduce new features designed to bridge the gap between data platforms and business users. Databricks positions Agent Bricks as a response to a growing challenge in enterprise AI: the complexity and cost of bringing prototype agents into production. The company said many organizations rely on trial-and-error methods, spot-checking AI outputs or relying on subjective judgments about quality in an inconsistent, time-consuming and financially inefficient process. "One of the biggest things that keeps these models from getting into production is that there's no good way to evaluate whether or not agents are going to do what you expect them to do," said Joel Minnick (pictured), Databricks' vice president of marketing. Agent Bricks attempts to eliminate that guesswork through a series of automated steps. Users start by describing the task they want an agent to perform and connecting their enterprise data. The platform then generates domain-specific synthetic data that matches data already collected and furnished by the organization. It also creates evaluation benchmarks using its optimization engine called Test-time Adaptive Optimization. Rob Strechay, managing director at theCUBE Research, a SiliconANGLE sister company, said Databricks' focus on cost and performance addresses "a huge fear for organizations moving or looking to move from proof-of-concept to production." Agent Bricks' ability to generate synthetic data targets "the lack of training data for specific use cases, giving agents a larger pool of information to learn from." Large language model "judges" generate questions and expected answers to assess model performance. This enables the system to iterate through various configurations of models, retrieval setups, and tuning parameters and suggests options that balance performance and cost. Minnick said the approach allows organizations to experiment with the trade-offs between accuracy and cost. "Maybe we chose Llama 7B and were able to achieve 98% quality at this cost, while we achieved 85% quality using Anthropic at a much, much lower cost," he said. "You have a lot of control over exactly how the LLM judges perform." Strechay called the AI judges "one of the most important parts" of the announcement. "It creates evaluation criteria, generates vetting data and provides detailed insights on agent performance," he said. "It also creates a way for users to customize and create their own judging criteria for each agent." Built-in governance and integration with existing enterprise controls allow organizations to move AI projects from experimentation to deployment without requiring additional tooling or infrastructure. Agent Bricks comes with agents that address several common use cases. The Information Extraction Agent pulls structured data from unstructured documents like PDFs and emails. This allows retail companies, for example, to extract pricing and product details from supplier catalogs with varied formatting. A Knowledge Assistant Agent improves chatbot accuracy by grounding responses in verified internal documents. The agent aims to enable technicians to retrieve answers from standard operating procedures or manuals without searching. A Custom LLM Agent tackles tasks like summarization or classification, with the ability to tailor output to industry-specific languages. Healthcare providers, for example, can deploy models that reformat patient notes into clinician-friendly summaries, Databricks said. A Multi-Agent Supervisor allows organizations to orchestrate several agents working in tandem. A financial services example combines agents specializing in intent detection, document search and compliance checks to deliver more personalized responses for advisors and clients. Minnick said customers are seeing dramatic results in early trials. Biopharmaceutical company AstraZeneca plc "built an agent in 60 minutes that has parsed over 400,000 clinical trial documents, pulled out relevant information and compiled it for their researchers," he said. Another healthcare use case is summarizing patient notes and lab results to support clinicians. The updated Unity Catalog now supports Apache Iceberg managed tables through Representational State Transfer Catalog application programming interfaces. Databricks said this makes Unity the only catalog that allows external engines such as the open-source Trino, Snowflake Inc., and Amazon Web Services Inc.'s Elastic MapReduce to read and write to performance-optimized Iceberg tables with governance controls intact. The company said this eliminates table format lock-in and enables interoperability across data environments. Unity Catalog now offers three key Iceberg-related features: the ability to create Iceberg-managed tables, transparent governance of Iceberg tables in external catalogs, and integration with the Delta Sharing ecosystem. That means organizations can manage and share data regardless of table format or compute engine. Databricks Assistant, a natural language interface embedded in Unity Catalog, now helps users explore data, assess its quality, and understand context. Databricks is also targeting business users with Unity Catalog Metrics and an internal marketplace that aims to make data more accessible by surfacing curated data assets organized by business domain. Unity Catalog Metrics allows key performance indicators and business metrics to be defined and governed as first-class data assets. They can be queried using SQL and are decoupled from specific business intelligence tools for consistent reporting. Serverless GPU support enables users to run machine learning and generative AI workloads without directly managing GPU infrastructure. This option, now in beta test, lowers the barrier for smaller teams or pilot projects that previously struggled with the complexity and cost of GPU provisioning. In addition, Databricks released MLflow 3.0, the latest version of its open-source platform for managing machine learning workflows. The update includes a new architecture called LoggedModel, which directly ties model weights and code to training runs. Enhanced visualization and debugging tools allow teams to compare performance across environments, while tighter integration with the Databricks Lakehouse aims to simplify governance, traceability and production deployment. The company said the three offerings constitute an effort to make AI development more predictable and cost-efficient for enterprises that may lack deep in-house machine learning expertise. It's betting that many organizations will adopt advanced AI tools if the process can be simplified and performance made more transparent. Agent Bricks and serverless GPU compute are available in beta test starting today. MLflow 3.0 is generally available.

Agent Bricks teaches users to follow best engineering practices - SiliconANGLE

The best judge of artificial intelligence could be AI -- at least that's the idea behind Databricks Inc.'s new tool, Agent Bricks. Built on Databricks' Mosaic AI platform, Agent Bricks allows users to request task-specific agents and then generates a series of large language model "judges" to determine that agent's reliability. "Agent Bricks is really the generalization of the best practices, the verticals that we saw, the styles that people use, the techniques that we saw work the best, all in one product," said Jonathan Frankle (pictured), chief AI scientist of Databricks Inc. "It reflects philosophically how we think people should build agents. It reflects what worked and what didn't work. Now it's ready for prime time." Frankle spoke with theCUBE's John Furrier at the Databricks' Data + AI Summit, during an exclusive broadcast on theCUBE, SiliconANGLE Media's livestreaming studio. They discussed how Agent Bricks evolved from internal best practices into a full-fledged product designed to evaluate AI with AI. (* Disclosure below.) The seed for Agent Bricks came from customers' need to evaluate their agents, according to Frankle. Ensuring that an agent is reliable starts with defining a criteria and a set of practices for comparing agent performance against it. "AI is a little bit unpredictable, non-deterministic, fuzzy," Frankle explained. "That's where LLM judges come in. You have an LLM that evaluates when the LLM is working well. To do that, you have to make sure the LLM judge knows what you're trying to do, knows how to measure it. It's really about, 'Does the LLM judge agree with a human judge?'" Getting all of the humans to agree on what the model should look like can be half the battle, Frankle suggested. That's why humans are in the loop throughout the agent development process. Databricks has essentially created scaled reinforcement learning, wherein the judges can train an agent to behave how developers want it to. "You don't need to give a bunch of labeled data," Frankle said. "Getting labeled data is really hard for humans. But getting a judge is not that hard. And we took a lot of time to figure out what was easy and hard for our customers to get, how we could do the science to make it possible to customize an LLM using that data." Despite the rise of vibe coding -- which Databricks' recent updates enable -- Frankle hopes that tools such as Agent Bricks will push all its users to think more like software engineers. Agent Bricks forces customers to test and evaluate over and over again until the model is extremely reliable. "An AI demo, you can slap together, you can show to your CEO, it'll have some cool behaviors and everybody will be excited," Frankle said. "That's not how you get into production. AI engineering is building a system that is carefully calibrated to solve a particular problem. You can measure how well it's solving that particular problem. When it doesn't work the way you want it to, you add more measurement to make sure you never see that problem again." Here's the complete video interview, part of SiliconANGLE's and theCUBE's coverage of the Databricks' Data + AI Summit:

If You Build Them: Databricks To Launch New Data Workflow, AI Agent Development Tools

Lakeflow Designer and Agent Bricks technology unveilings for building data pipeline workflows and AI agents, respectively, are on tap at Wednesday's Databricks Data + Summit. With new technologies for constructing data pipelines and building AI agents, attendees at this week's Databricks Data + AI Summit can be forgiven if they feel like they should be wearing hard hats. Data and AI platform provider Databricks Wednesday launched Agent Bricks, a unified workspace for building production-scale AI agents, and Lakeflow Designer, a no-code tool that data analysts can use to build reliable data pipelines. Agent Bricks and Lakeflow Designer top the announcements coming out of the Data + AI Summit in San Francisco as Databricks looks to maintain its momentum as one of the IT industry's fastest-growing companies and one of the most influential in the critical juncture of big data and AI. [Related: Meeting The Data Needs Of The AI World: The 2025 CRN Big Data 100] "There's a lot of pressure for organizations to scale their AI efforts. Getting high-quality data to the right places accelerates the path to building intelligent applications," said Ali Ghodsi (pictured), Databricks co-founder and CEO, in the's Lakeflow Designer announcement -- a comment that could also serve more broadly as the company's mission statement. Ghodsi is scheduled to deliver the Data + AI Summit keynote Wednesday morning, joined by Anthropic co-founder and CEO Dario Amodei, JPMorgan Chase CEO Jamie Dimon, and Nikita Shamgunov, CEO of Postgres database startup Neon that Databricks acquired in May. Microsoft Chairman and CEO Satya Nadella is also scheduled to deliver a recorded address. While Databricks is seen as a strong candidate for an IPO, it remains privately held and does not disclose details about its finances. But in December the company raised $10 billion in a funding round that put its value at $62 billion and at the time said it expected to surpass an annual revenue run rate of $3 billion shortly and achieve positive cash flow. The $10 billion funding round increased to $15 billion in January with the addition of a $5.25 billion credit facility. At the core of Databricks' momentum is the company's flagship Databricks Data Intelligence Platform, a unified data and AI platform built on a lakehouse architecture and powered by the company's Data Intelligence Engine. The system offers a range of capabilities including data analytics, data integration, data catalog, data governance and security, and more. The company continues to expand the platform's functionality: The Neon acquisition, for example, added serverless Postgres database capabilities to the Data Intelligence Platform for developers and for AI agents. And that's the driver behind Wednesday's Lakeflow Designer and Agent Bricks announcements, both being demonstrated at the Data + AI Summit event. Agent Bricks is a new unified workspace for building AI agents that Databricks says works with an organization's unique data to achieve "production-level accuracy and cost efficiency." Agent Bricks builds on technology Databricks acquired when it bought generative AI startup MosaicML in June 2023 for $1.3 billion. Last year Databricks unveiled Mosaic AI Agent Framework for building AI agents on the Databricks platform. "Databricks' strategy is data intelligence. How do we build AI systems that can reason over your enterprise data on your enterprise tasks? Agent Bricks is really a personification of that," said Hanlin Tang, Mosaic co-founder and now Databricks CTO for neural networks, in an interview with CRN. "I think for this Agent Bricks product, we really took a step back and rethought -- and really took a new approach to -- how we think the world really should be building these agents," Tang said. Tang said Agent Bricks is designed to overcome several common problems around AI agents and agent development, including organizations' lack of enough data to build agents and difficulty evaluating how well they are working once in production. And another major challenge: "The industry has definitely evolved from just using models to building entire agent systems that have tools, or vector databases, or all these sorts of different components," Tang said. "And then suddenly, there is an explosion of choices. What model do you use? How do you string these things together? What kind of agent workflow should you use?" Agent Bricks offers an automated way to create high-performing AI agents tailored to a business, according to a preview of the announcement provided to CRN. Developers provide a high-level description of the agent's task and connect it to enterprise data. Agent Bricks "automatically generates task-specific evaluations and LLM judges to assess quality" and creates synthetic data to substantially supplement the agent's learning. Agent Bricks then searches across "the full gamut of optimization techniques" to refine the agent, according to Databricks. Agents developed using Agent Bricks can be used for a number of common use cases, including information extraction, knowledge assistant, and custom LLM agents for such tasks as summarization, classification or rewriting within specific industries. Tang said Agent Bricks can be used by Databricks partners who provide AI development services for customers, including developing AI agents for clients. Agent Bricks is currently available in beta. Databricks is also announcing new AI features for Mosaic AI including support for serverless GPUs (available in beta) and the MLflow 3.0 unified platform for managing the AI life cycle (now generally available). The new Lakeflow Designer provides nontechnical data analysts with a no-code data ETL (extract, transform and load) approach to building reliable, production-grade data pipelines, according to Databricks. Data pipelines are traditionally built by data engineering teams. But with such projects often facing backlogs, data analysts sometimes take on the job themselves using low-code/no-code tools that Databricks says can sacrifice data governance, and system reliability and scalability. "There is this shadow data engineering that's happening outside of Databricks, and it's typically in the nontechnical section of [customers] -- roles like business analysts, operations analysts, just nontechnical people," said Bilal Aslam, Databricks senior director of product management, in an interview with CRN. "Their job isn't to write code. Their job isn't to be sitting in Databricks every day. What these people are doing, to get their jobs done, to get fresh, clean data, they're using tools that are typically in the no-code or low-code category, and they're doing their own data preparation. "And this sort of creates this shadow data engineering," Aslam said. "You end up solving this real [data pipeline development] problem, it's a multibillion-dollar problem, but with tools that are essentially limited, there's a dead end ... because these tools don't integrate back into the data intelligence platform." Lakeflow Designer, which will be in preview later this summer, uses a visual drag-and-drop user interface and natural language generative AI assistant to create data pipelines with the same scalability, governance and maintainability as pipelines built by code-first data engineers using more sophisticated tools, according to Databricks. While targeted toward nontechnical users, Lakeflow Designer is backed by Databricks Lakeflow, the company's data pipeline development platform for technical data engineering teams that was unveiled in June 2024. Databricks is unveiling the general availability of Lakeflow at Data + AI Summit. Aslam said Lakeflow Designer is "an accelerant" for solution provider and systems integrator partners looking to close deals because it opens up the range of use cases for the Databricks platform, and data preparation projects are no longer slowed by the lack of data. "We're getting very positive feedback from [systems integrator] partners on this capability because it simplifies this bottleneck in their individual [data preparation] processes," he said. The company is also unveiling several new Lakeflow capabilities, available in private preview, including a new IDE (integrated development environment) for data engineering and new data ingestion connectors for Google Analytics, ServiceNow, Microsoft SQL Server, SharePoint, PostgreSQL and Secure File Transfer Protocol. Lakeflow also can now direct-write to the Databricks Unity Catalog using Zerobus.

Databricks Launches Agent Bricks: A New Approach to Building AI Agents

Agent Bricks addresses several common customer use cases across key industries: Information Extraction Agent turns documents, like emails, PDFs and reports into structured fields like names, dates and product details. Retail organizations can easily pull product details, prices and descriptions from supplier PDFs, even if the documents are complex or formatted differently. Knowledge Assistant Agent solves the issue of getting vague or flat-out wrong answers from chatbots, with fast, accurate answers grounded in your enterprise data. Manufacturing organizations can empower technicians to get instant, cited answers from SOPs and maintenance manuals without needing to dig through binders. Multi-Agent Supervisor enables you to build multi-agent systems that seamlessly stitch together agents across Genie spaces, other LLM agents and tools such as MCP. organizations can orchestrate multiple agents to handle intent detection, document retrieval, and compliance checks, creating complete, personalized responses for advisors and clients. Custom LLM Agent transforms text for custom tasks such as content generation or custom chat, optimized for your industry. Marketing teams can build customized agents to generate marketing copy, blogs or press releases that respect their organization's brand.