Skip to content

Data Bricks develops the agent “Jenny” for autonomous management tasks: breaking CEOs’ isolation through a data ontology

Share
Data Bricks develops the agent “Jenny” for autonomous management tasks: breaking CEOs’ isolation through a data ontology

Listen to this article

Read by Anchor

The new updates Data Bricks has introduced to its intelligent agent “Jenny” reveal a technical effort to address a chronic communication and managerial gap that afflicts organisations as they scale, allowing the agent to carry out multi-step tasks that span long work periods without direct human supervision, by relying on an integrated software ontology that provides it with full context from the data stored within the company’s platform.

The system stems from a conviction articulated by Data Bricks’ chief executive, Ali Goodsi, who cites the “Dunbar number”, the metric defined by British anthropologist Robin Dunbar for the ceiling of stable relationships a person can maintain, about 150 individuals. In startups, the leader communicates directly with everyone and receives explicit negative feedback, but as an organisation grows to 12 000 employees like Data Bricks, sales and HR managers hesitate to relay failures or acknowledge stalled initiatives to senior management.

The intelligent agent thus becomes a channel for corporate candour, extracting from data layers what hierarchical filters conceal.

Practical use inside the company has shown its chief relying on the agent for tasks beyond traditional presentations: Jenny analysed employee satisfaction surveys, identified weak points that earned low scores, and helped discard remedial solutions proposed by some managers after they proved ineffective. It also summarised the details of a unsuccessful major sales meeting with a prospective client, highlighting the technical features the client required before purchase, and prepared interview agendas, prior media briefings, anticipated questions and real-time conversation suggestions.

This shift rests on moving the software briefing tools and model-guidance systems that served software engineers to executive management, turning tasks from moment-to-moment monitoring into autonomous, sustained operation. In contrast to security concerns tied to untrusted personal tools such as “Open Clo”, “Instinct” or “Muse”, the company offers its corporate agent as a permanent leadership tool that relies on data assets protected within its cloud platform.

Curbing rising inference costs is a survival condition for multi-step agents operating in production environments.

Because running independent agents can double cloud consumption bills, tools like the “Unity Gateway” portal allow predefined usage caps, with the option to replace expensive models with open-weight, lower-cost alternatives, giving organisations flexibility to tune operating budgets while granting agents execution independence.

This trajectory directly impacts business environments in the Gulf, Egypt and the wider region, where large entities such as banks, telecom groups and state-holding bodies manage divisions and units far beyond the Dunbar threshold and struggle to convey real-time operational pulse from branches to decision-making offices. Providing agents that rely on a central ontology while permitting open-weight models through control portals lets regional tech teams lower licensing costs and use financially-controlled local inference instead of exclusively relying on closed-source models with accumulating charges.

The true benchmark for executive agents’ success becomes the ability to sift hard operational facts from the thicket of corporate data, away from administrative polishing, while strictly adhering to spending limits that make automation economically viable over the long term.

Don't miss the next story

Subscribe for updates