Databricks Blog

Building an agent to automate simple business tasks can be easy. But creating one...

In earlier posts, we introduced contextual policies in Omnigent, showed them blocking...

Your data estate holds far more than structured tables, metrics, and transaction...

AI coding tools deliver immense value: at Databricks, agentic coding has measurably...

AI assistants use language models, data retrieval, and reasoning to understand requests...

As organizations move beyond single-prompt AI interactions, agentic workflows are...

Tool calling is the ability of an AI model to interact with external tools, APIs,...

Today, we are releasing OfficeQA Pro V2, a new benchmark designed to evaluate whether...

A year ago, the best open-weight models trailed their proprietary counterparts by...

Migration as a strategic evolutionBigQuery is often the standard for starting fast, but for many enterprises...

The last six months have seen a rapid rise in AI-powered productivity and a proliferation...

Your dbt project runs 80 models every night. The warehouse bill doubled last quarter....

Databricks is a sponsor at Black Hat USA 2026 this week. Find us at Booths #5106...

A VP of Sales at a major CPG has three decks open before her first coffee. One from...

For years, ingesting semi-structured data like JSON, XML, or CSV meant a difficult...

Today, we are thrilled to announce that Databricks has officially completed the acquisition...

The article argues that healthcare organizations need to become 'AI-forward' by building a foundation of unified data, adaptive governance, and a scalable operating model, rather than just buying more AI tools. It identifies three common blockers—fragmented data, misaligned governance, and lack of a repeatable operating model—and explains why now is the right time to start.

This article presents a reference implementation for agentic media buying on Databricks, where autonomous buyer and seller agents transact using open standards (IAB Tech Lab's AAMP) and the Databricks platform (Model Serving, Lakebase, Unity Catalog, etc.), solving coordination bottlenecks and freeing teams to focus on strategy.

Databricks announces agentic code converter in Genie Code (Beta) that converts proprietary SQL dialects to open ANSI SQL using parallel subagents for iterative conversion, syntax and semantic validation. It simplifies data warehouse migration with project management, complexity scoring, lineage, and custom skills.

Databricks launches the agentic code converter (Beta) using Genie Code AI to automatically convert proprietary SQL dialects (T-SQL, Snowflake, Redshift, Oracle, BigQuery, Teradata) to open ANSI SQL, with features for migration projects, complexity scoring, lineage analysis, and custom skills, dramatically simplifying legacy data warehouse migration to the Lakehouse.

NBCUniversal achieved a 30% cost reduction and enhanced scalability and analytics by migrating to the Databricks Data Intelligence Platform. Partnering with EXL, they used a phased migration strategy with custom accelerators to automate code conversion, data migration, and validation, while unifying ML and analytics workloads.

Healthcare finance teams face high-risk decisions due to fragmented systems and outdated data. Databricks Genie acts as a governed, data-smart AI coworker grounded in an ontology that captures the meaning behind numbers, helping finance teams understand costs, revenue, and cash flow in business context.

Manufacturing ties up cash in inventory, receivables, and equipment. Finance's job is to free that capital to protect margins. AI agents like Databricks Genie help finance teams find where capital is trapped and take action.

Fraud and improper payments cost federal benefits programs hundreds of billions annually. Databricks applies its real-time fraud detection technology proven in banking and insurance, combining AI, real-time analytics, and cross-agency data sharing to score and block suspicious transactions without delaying legitimate aid. Over 80% of federal executive departments already use Databricks, and tools like OpenSharing and Clean Rooms enable secure fraud signal sharing.

ProdLine CoPilot leverages the Databricks Data Intelligence Platform to read live production line state, routing queries to specialized agents (downtime analysis, quality, supply chain, schedule optimization, etc.) that invoke real solvers and deliver actionable recommendations within minutes. Human-in-the-loop approval ensures traceability and control, reducing OEE losses by enabling in-shift decisions instead of post-shift analysis.

How agentic AI can help telecom finance teams protect the margin when every moment matters | Databricks Blog Skip to main content How preventing revenue leakage became finance's front line In telecom, revenue is earned in billions of small moments: a call connected, a gigabyte used, a subscription renewed, another carrier's traffic routed across the network. Each one has to be rated, billed, and collected correctly, and at that scale even a small percentage that slips becomes a very large sum. Revenue leaks in three places: services that go unbilled or uncollected, charges lost to fraud and partner-settlement errors, and customers who churn and take their spending with them. In a market where prices are flat and the cost of switching is low, keeping the revenue you have already earned matters as much as winning new revenue. This is the environment in which telcos operate, and their finance departments are the constant through all of it, helping the business understand and act on rising complexity. But before finance can stop a leak, it has to see one, and that is where the real bottleneck lives. Those billions of moments are recorded across ordering systems, billing platforms, ERPs, and spreadsheets, and no two of them describe the business quite the same way. At month-end close, the picture only comes together after reports are pulled from each system and reconciled by hand, which means finance teams often spend more time locating, validating, and reconciling data than actually using it to make decisions. That wasted time is not just an efficiency problem. A billing discrepancy that surfaces weeks after the fact is usually a write-off; the same discrepancy surfaced today is recoverable revenue. Accurate, real-time reporting and billing is not adjacent to revenue assurance. It is the precondition for it. It’s all compounded by agents shaping how usage is rated, how a charge is flagged, and how a retention offer is made. Studies show that global operators experience about $40B in revenue leakage every year (TM Forum). The pace will vary by operator, but as waves of agents reshape billing and finance systems, AI spend, and fraud, the race is on for finance departments to harness them to help accurately and reliably identify and prevent leakage. Billions of small moments earn the revenue. Finance protects revenue retention. Why a word like ontology now matters to telecom finance Finance has always been good at finding the number, even when it is buried in complexity. But they are also the first to let the business know the numbers do not tell the whole story. What matters is the meaning behind them: which service, which plan, which contract, and which partner settlement, and how each of those is changing as the business moves. In a telecom, that meaning is scattered by design. The same service can carry a different name in each of a dozen billing and ordering systems, and the fields that describe it were often built for the system, not the business. An answer can be perfectly accurate and still not be correct, because it rests on a partial or dated picture of how the business actually works. Put plainly, is the number seen in the full context of the business? That is what an ontology does: it captures meaning and keeps it current as the business changes. As Ali Ghodsi puts it, most enterprise AI is guessing with false confidence, a context problem, not an intelligence problem. But, as with every technology, it is how the capability is delivered that makes all the difference. Which brings us to a new kind of ontology, built for the demands telecom finance places on it. Accurate is the right figure. Correct is the same figure, rooted in the service, the plan, and the settlement. Where Genie becomes the answer In telecom, catching a leak in time is the difference between recovering the revenue and writing it off, and the picture goes stale quickly as plans, usage, and fraud patterns change. So the ontology itself has to keep moving. It has to learn from the systems the business runs, sharpen with every question, and adapt as rates and patterns shift, so the context stays fresh rather than captured once and left behind. This is where Genie becomes the answer. Databricks built Genie as a data-smart AI coworker: a coworker to whom a finance leader can ask a direct question and get a trustworthy, sourced answer in return, grounded in Genie's ontology and governed at every step. That means the complex reporting, invoice, and billing questions that used to wait on a reconciliation cycle - which invoices are past payment terms? What is our spend with our top vendors year to date? How did this month's variance form? - get answered in the moment, from one governed view of the data. And because Genie learns the business rather than just querying it, it proactively surfaces the anomalies and discrepancies a person working across a dozen systems would miss. It is built to help finance have more accurate answers and, more importantly, deliver trusted actions, beyond just providing readouts of what has happened. Finance teams from leading telecoms are already seeing the impact of Genie. At Lumen Technologies, for example, thousands of employees within the Office of the CFO are using Genie as a reasoning agent to accelerate time to insights across billing, finance, accounts payable, and more. Answers on complex spend and invoice data that once took days of manual Excel reconciliation can now be retrieved in minutes with Genie agents. Watch the webinar→ Consider the three questions on the minds of every telecom finance team, each tied to one of three outcomes that compound, one feeding the next. For each, Genie does more than retrieve the data and answer. Its ontology learns the business, sharpens with every question, and shows its work: Are we capturing every dollar we earn, every service billed, recognized, and collected to its terms? Start with the revenue you have earned. Every call, gigabyte, and subscription has to be rated, billed, and collected correctly, and across billions of events a small share always goes unbilled or uncollected. Which charges and partner settlements are at risk of fraud or error, draining revenue before it is booked? Then the leakage to fraud and error. Fraud and partner-settlement mistakes drain revenue before it is ever booked, so the value is in catching a charge as the pattern forms, not writing it off later. Where is churn driving up the cost of replacing revenue we already had? Then the revenue you already won. Keeping a customer costs a fraction of replacing one, so the value is in seeing which high-value customers are about to churn while there is still time to hold them. Three questions, three outcomes, one mechanism. Genie learns the business, sharpens with every question, and shows its work. That is the difference between reporting what already happened and continuously learning about your business, getting smarter with every interaction. And because every figure traces to its source, every permission holds, and the cost of the AI itself stays governed under one model, it is an answer finance can trust to act on. Genie readies the move, to recover an unbilled charge, to block a fraudulent one, to make a retention offer, and a person in the loop makes the call. Finally, Genie's learning across all three comes together. Billing and collecting every service, closing the fraud and settlement gaps where revenue leaks, and keeping the customers you have already won means every dollar you earn becomes a dollar you keep. That turns three separate fights into one reinforcing mechanism: each move sets up the next, and the momentum compounds. Each move sets up the next. Genie's learning across all three is what makes the momentum compound. A data-smart AI coworker built for the way telecom finance works This is the force multiplier built for what finance departments require. The critical people driving rigor and discipline across the business can now lean on a data-smart AI coworker that is always getting smarter, always current, always governed, truly understanding the business. Networks will keep carrying more of the world's traffic, while a tool like Genie will help finance keep more of the revenue it earns. See what a data-smart AI coworker looks like for telecom finance. Databricks Genie is available today. Frequently asked questions What is changing for finance in telecom? More of the decisions that move revenue, billing, fraud, and retention, are made by agents. Finance's mission to capture and keep the revenue it earns is unchanged; what has grown is the speed and complexity of change, which finance tools must understand and govern. Does Genie make billing, fraud, or pricing decisions? No. Those calls belong to revenue assurance, the fraud and interconnect teams, and marketing and care. Genie gives finance an accurate, governed view to see a forming risk early and guide or direct the owners who act on it. Why do ontology and governance matter to a telecom CFO? Ontology captures what the numbers mean for your business and keeps it current, so an answer is correct and not just accurate. Governance keeps every figure traced, permissioned, and cost-controlled. Together they make an answer safe to act on. How is Genie different from an AI dashboard or BI tool? A dashboard shows you what the data says. Genie is a data-smart AI coworker that helps you act on it, grounded in your ontology and governed end to end, with a person deciding. Get the latest posts in your inbox Subscribe to our blog and get the latest posts delivered to your inbox. Sign up View all blogs

Media finance teams must accurately quantify audience value, optimize subscription and advertising pricing, and ensure content investment ROI. With rising complexity from automation and agents, Databricks Genie offers a data-smart AI coworker that leverages ontology and governance to provide trustworthy answers, helping finance teams protect margins.

Databricks AI Search now supports high-QPS scaling, enabling standard endpoints to reach thousands of queries per second with a single parameter, without managing replicas or load balancers. Ideal for search bars, recommendation systems, and real-time entity resolution, with built-in production observability.

Databricks governs AI spend at scale by routing every coding agent through Unity AI Gateway, enforcing budgets, visibility, and policies across all models and tools. They balance innovation with cost control using separate daily and monthly budgets: daily limits for runaway protection with self-service raises, and monthly limits for extraordinary spend with manager approval and time-limited overrides.

Genie One is an AI coworker that integrates with existing business tools to automate recurring tasks. This article presents four practical use cases: automated business reviews and reporting, meeting preparation and follow-up, knowledge work and document automation, and operational monitoring and alerts. Each use case includes step-by-step instructions to set up workflows in minutes.