Webmotors accelerates data queries for 353 users with Genie Agents on Databricks

Webmotors, Brazil's largest automotive solutions platform, connects buyers and sellers across millions of vehicle transactions. Previously, business units depended on the central data team for every analytical request. The ticket queue became a primary bottleneck, forcing business users to wait days for reports and limiting data autonomy to a small group of SQL-literate employees. By implementing Genie Agents and AI/BI Dashboards on the Databricks Data + AI Platform on AWS, Webmotors replaced ticket queues with direct self-service. Today, 353 active users extract instant insights from governed data without writing code.

Overcoming data access bottlenecks with Genie

Webmotors needed to speed up data delivery across commercial, product, technology and support teams. Under the legacy model, business teams consumed data through static reports in Power BI and Metabase. Metabase served as the primary self-service portal for roughly 150 users, but in practice, autonomy depended on the ability to write SQL. As a result, every new question carried an operational cost and required joining a support queue managed by the data team.

The model created three distinct operational challenges:

• The ticket queue forced the data team to prioritize ad hoc support over building new analytics infrastructure.
• Executive decision-making suffered because business units built custom views directly from legacy tables without applied business rules, preventing a single source of truth.
• Business leaders could not explore hypotheses in real time during critical operational events, such as daily incident reviews or live Upstream strategy sessions.

"Before Genie, the data team was a mandatory bottleneck," said Henrique Sales Leite, Senior Data Analyst at Webmotors. "Every question carried a cost to ask, so users only asked what justified joining the queue. The autonomy existed on paper, not in practice."

Building a governed Lakehouse with Unity Catalog and Genie

To eliminate the ticket bottleneck, Webmotors deployed the Databricks Data + AI Platform on AWS. The team built a Lakehouse foundation using Unity Catalog for centralized governance. They also implemented a medallion architecture, cleaning and curating data into gold-layer tables with standardized business rules and taxonomies. This governed foundation became the prerequisite for reliable natural-language querying at scale.

Deploying Genie Agents for self-service analytics

On top of this governed layer, Webmotors deployed Genie One, Genie Agents and Genie Code. The Insights team curated 26 official Genie Agents dedicated to specific business domains, such as B2C revenue, customer churn, reimbursement tracking and dealer metrics. Non-technical business users across seven departments ask natural language questions in Portuguese, refine cuts interactively and interrogate data without writing SQL.

"Genie only works well because the data is properly cataloged in Unity Catalog with well-written table descriptions and standardized taxonomy," said Vitor Pereira, Data Analyst - Insights at Webmotors. "Genie arrived at exactly the right moment to serve as the self-service edge of an organized foundation."

Integrating AI/BI Dashboards for real-time insights

Webmotors also adopted AI/BI Dashboards, growing from 88 to 154 Genie-enabled dashboards between May and August 2026, with Genie Agents embedded directly into visualization layouts. This enables business stakeholders to view prepared KPIs and ask follow-up questions within the same screen. In a sample of dashboards, 92% were built with AI assistance, with almost one distinct creator per dashboard, showing that business teams are building their own views.

Unlocking efficiency and speed with Genie Agents

Implementing Genie Agents on Databricks transformed how Webmotors interacts with enterprise data. Active self-service platform users grew from 150 on legacy tools to 353 autonomous users on Databricks, and 86.7% of employees with platform access now use Genie. Over 16 months, negative feedback on data queries dropped from 41.9% to 10.6%, reflecting improved satisfaction with data answers. In August 2026, 81% of questions were handled by official Genie Agents curated on certified data.

The most direct business impact shows up in incident response. Webmotors conducts a daily incident review, and when an issue arises, product managers need to identify the cause before the damage grows. Instead of joining the data team's queue, a small group of product specialists now uses Genie Agents to run root cause analysis on their own, in minutes rather than the hours an analyst would need.

"In incident response, the goal is to limit damage, so the analysis has to happen as fast as possible," said Henrique. "Every hour of investigation is accumulated damage. With Genie Agents, our PMs run exploratory analysis that would take an analyst hours to develop."

Integrating natural language querying directly into daily routines unlocked tangible speed and efficiency across specific operational use cases:

• Customer base and churn tracking: Teams identify customers who appear one month and drop off the next, a churn analysis that previously required an analyst writing SQL. Teams also pull actionable lists, such as customers with balances expiring in the next 30 days, so that they can act before those balances lapse.
• Intraday customer satisfaction monitoring: An NPS dashboard tracks satisfaction hour by hour across collection channels, so teams respond the same day instead of waiting for month-end close.
• Optimized financial and reimbursement processing: Cross-referencing unit costs against revenue eliminated the need for manual table downloads, while reimbursement verification reports that once took days now complete in minutes.
• Agile pipeline development: Data engineers pair Genie Code with Claude Code to validate transformation code, cutting notebook development time from 1–2 days down to half a day.
• Automated executive reporting: Generating and updating monthly executive committee reports shifted from a full day of manual data aggregation to immediate retrieval.
• Self-service historical and dealer analytics: Historical queries that previously required Jira tickets and days of wait time now take three clicks, and teams build custom dealer performance dashboards independently.
• Commercial CRM autonomy: Commercial teams built consolidated CRM data views to support committee presentations, structure sales training and supply accurate metrics without developer help.
• Strategic portfolio and project insights: Cross-referencing B2C revenue, vehicle inventory and lead generation delivers an integrated portfolio view that surfaces growth levers. Teams on strategic initiatives such as Revenda Mais also gather research data rapidly.
• Report consolidation: Users eliminated, consolidated or optimized roughly 139 manual reports and dashboards, reducing repetitive work and the risk of manual error.
• Consolidated operational savings: In an internal survey, 85% of users reported saving analyst hours, averaging 3.5 hours per person weekly. Audit log data show roughly 504 hours returned to the business each month, equivalent to about 7 working days per year for each person who asks Genie questions.

"Genie brought speed to decision-making, especially in meetings where querying data in real time is critical," said Vitor Pereira. "Data stopped being queried and started being interrogated."

Scaling conversational AI across the enterprise platform

Looking ahead, Webmotors aims to expand Genie usage to 100% of its internal workforce while managing query efficiency and costs. The company is also rolling out official Genie Agents and Databricks Apps to its broader network of affiliated group companies representing over 1,000 employees. By unifying a governed Lakehouse foundation with conversational Genie Agents and custom experience layers, Webmotors continues to scale analytical capabilities across its entire automotive platform without linearly increasing data team overhead.