Summary

• Quintas Energy has selected Box to create a governed content layer across its renewable-asset documentation.
• Client and asset onboarding provide the initial operational focus, while more advanced agentic and extraction features remain future capabilities.
• The programme illustrates why enterprise automation inherits weaknesses in permissions, metadata and records before it can improve workflows.

Quintas Energy is building the information layer beneath its AI programme before attempting wider automation, using Box to bring technical, contractual, financial and regulatory documents into a governed environment connected with its renewable-asset workflows.

The Seville-headquartered asset manager has selected Box Enterprise Advanced as part of its 2026 digital programme, with client and asset onboarding among the first processes targeted. Those workflows can involve substantial volumes of engineering material, contracts and compliance documentation that arrive from several counterparties and have to be reconciled before an asset can be managed consistently.

Box will provide a content layer intended to capture, classify and govern that material while integrating with Quintas Energy’s proprietary platform. More advanced functions, including tailored AI agents, automated metadata extraction and additional workflow tooling, are described as future capabilities rather than systems already operating across the portfolio.

Aida Durnes, chief operating officer at Quintas Energy, said AI’s “value depends on the quality, structure, and governance of the information behind it”. The phrase lands on a persistent enterprise problem: automation can accelerate the use of organisational information, but it can also accelerate the consequences when that information is outdated, misclassified or available to the wrong people.

Automation inherits the information estate

Renewable-asset management provides a particularly clear environment in which to see those dependencies. Contracts govern commercial obligations, engineering records describe physical assets and maintenance history, while regulatory material changes across jurisdictions and financial records have to remain connected to the correct project throughout its operating life.

If those sources are fragmented across shared drives, email and separate systems, an AI layer does not remove the fragmentation by itself. Retrieval can return the wrong version of a document, access controls can expose material beyond its intended group and an automated process can act on an error before a person notices the context is missing.

Traditional content management therefore remains part of the AI architecture. Permissions determine what models can retrieve, metadata associates records with the correct assets, retention rules determine what remains available and audit trails create evidence of how automated systems have interacted with business information.

Those functions are familiar, although their operational weight grows once software can generate actions rather than simply help somebody find a file. A weak filing structure is inconvenient when a person is searching manually; the same weakness can propagate through onboarding or reporting when software is allowed to process documents automatically.

Agents come after the records

Quintas Energy’s initial focus on onboarding gives the programme a process that can be measured without relying on broad claims about AI productivity. The business can compare how long onboarding takes, how much manual coordination is required, where documents are re-entered and how often staff have to resolve inconsistencies before and after the new environment is introduced.

That evidence should provide a better basis for deciding where more autonomous workflows are justified. Box has outlined future use of agents, extraction and workflow automation, but treating those capabilities as a roadmap rather than a completed deployment also makes it easier to separate demonstrable operational improvements from vendor functionality that has yet to prove itself in this organisation.

For Box, the deployment supports a broader attempt to make governed enterprise content the substrate for AI rather than leaving customers to move information into separate automation platforms. That proposition has to compete with cloud, productivity and enterprise software vendors that are also trying to place their own AI layers closer to corporate data.

Quintas Energy will have a more practical measure of success. If its staff can move renewable assets into operation faster while maintaining access control, compliance and document quality, the AI programme will have improved a real workflow before the company asks agents to do anything more ambitious.