Gravity launches AI agent for sustainability teams

Gravity has launched an AI agent for sustainability and energy teams and published a buyer's guide on how to assess AI agents sold to enterprise customers.

The new product, the Gravity Agent, is an all-purpose tool that can carry out any action a user can take within the Gravity platform. It includes more than 85 pre-built skills developed by the company's climate strategists and energy specialists.

The launch comes as software suppliers across sectors increasingly apply the term "AI agent" to products that often perform only narrow tasks. Gravity is using the release to argue for a stricter definition of the label, particularly for buyers in areas such as carbon accounting, regulatory disclosure and energy management, where data quality and auditability matter.

The accompanying buyer's guide sets out nine questions for procurement teams and software buyers. They focus on whether an AI agent can work beyond a scripted demonstration, handle incomplete or inconsistent real-world data, and perform a broad range of actions inside a software platform rather than a small set of isolated functions.

Buyer scrutiny

The guide reflects a wider concern in enterprise technology markets that AI terminology has become difficult for customers to compare. Gravity is seeking to distinguish between agents that can execute multi-step work and tools that assist with one task at a time, such as spreadsheet cleaning or question answering.

Ted Kornish outlined that distinction in comments accompanying the launch.

"Everyone says they have an agent, but almost no one means the same thing," said Ted Kornish, Co-Founder and CTO, Gravity. "A true agent handles a wide variety of tasks and works toward them: deciding the steps, taking action, and circling back only when it needs more guidance or has a draft ready for review. Gravity is the only API-first sustainability platform, meaning the agent can use the software in all the ways users can."

Early pilot users reduced the time needed for data migration, disclosure drafting and identifying energy cost savings by 70% or more, according to Gravity. It did not disclose the number of pilot customers or the sectors they operate in.

The platform's initial skill library includes functions for moving historical data from spreadsheets and other systems, drafting regulatory disclosures and identifying potential energy-saving projects. Users can also create additional skills tailored to their own processes.

Governance focus

Gravity is placing governance at the centre of the product launch, reflecting the sensitivity of sustainability data and the compliance demands on corporate reporting teams. Any change proposed by the agent requires human review and approval, and every action is logged so records remain audit-ready, according to the company.

Customer data is not used to train the agent, Gravity said. For emissions calculations, results continue to run through its deterministic calculation engine so sustainability teams and auditors can reproduce the figures.

Those safeguards address one of the central tensions in enterprise AI adoption. Companies want automation that cuts manual workload, but they also need clear oversight, especially where software is used for financial, regulatory or operational decisions.

Gravity's target users are sustainability and energy teams, which often spend large amounts of time on data collection, validation and reporting cycles. The company argues that AI tools should reduce administrative work rather than replace specialist judgment in emissions accounting or energy planning.

The launch also highlights the growing overlap between sustainability software and mainstream enterprise AI. Rather than limiting automation to dashboards or single-task assistants, vendors are starting to build tools that can act within core systems, complete workflows and return work for approval.

The agent can work on several tasks at once, operate in the background and retain memory to speed up repeated work, according to Gravity. That design points to a model in which software increasingly functions as a semi-autonomous assistant inside specialist business platforms.

AI emissions

Gravity also linked the launch to its work on measuring emissions associated with AI use. It said it developed an emissions accounting methodology for AI last year and now uses it to assess the footprint of the Gravity Agent, as well as help customers measure emissions from their own AI usage.

That is a notable point for buyers in the sustainability market, where companies are under pressure not only to adopt digital tools but also to account for their resource demands. As AI usage grows, energy consumption and associated emissions from model training and inference are becoming a more visible issue for procurement and sustainability teams.

The company says its aim is to free staff from repetitive work so they can spend more time on operational decisions and emissions reductions. "The Gravity Agent does not replace these teams, but rather gives them superpowers - freeing them to spend less time wrangling data and more time taking action to decarbonize operations, reimagine supply chains, and cut energy spend."