Prophecy launches DataGPT to run verifiable data workflows in AI tools

Solving the trust gap in AI

Many business analysts currently use AI chatbots to handle complex tasks, such as reconciling thousands of transactions from a spreadsheet. However, these chatbots often write their own internal scripts to reach an answer, which most non-technical users cannot read or verify. This creates a situation where employees must take the final numbers on faith without knowing the underlying logic.

Prophecy is launching DataGPT to address this specific problem by making the logic behind every AI answer visible. Instead of just providing a final number, the service generates a visual map of the steps used to reach that conclusion. This allows analysts to see exactly how the AI processed the information, ensuring the results align with company definitions and approved exceptions.

The company notes that closing a chat session currently wipes out the logic and rules explained during that specific interaction. DataGPT aims to prevent this loss by creating persistent workflows that can be reviewed and approved. As reported by SiliconANGLE, the goal is to put business data users on the same footing as software developers who already have tools to verify AI-generated code.

How the workflow functions

DataGPT works by connecting popular AI tools like ChatGPT, Claude, and Copilot to an organization's existing data infrastructure. An analyst can upload a messy file and ask the AI to cross-check it against records held in places like SharePoint or Snowflake. The system then lays out each stage of the process on the screen alongside the results.

If an analyst spots an error in one specific part of the process, they can fix that individual stage before approving the entire workflow. Once a workflow is finalized and signed off, it is backed by open code stored in Git. This ensures that any future reruns follow that exact code perfectly without the AI model changing the logic or incurring additional costs.

The service also builds a knowledge graph for each customer based on their specific business rules and data catalogs. This allows the system to understand unique company requirements over time. For example, an analyst could set up a rule for discounted product bundles and have the system run that specific logic every month automatically.

Security and IT governance

A major hurdle for many companies is 'shadow AI,' where employees use unauthorized tools to process sensitive company information. Prophecy pitches DataGPT as a way to bring these activities under the control of the IT department. When a user uploads a file, it does not sit inside the AI chat interface; instead, it goes through Prophecy to the organization's own data platform.

This architecture allows IT staff to see and govern every workflow just like any other software they manage. The service supports major data platforms including Snowflake, Databricks, and Google BigQuery, allowing companies to keep their data where it already lives. This setup ensures that data remains within the company's security perimeter while still providing the benefits of generative AI.

The move toward governed AI comes as organizations prepare for stricter data standards. Prophecy points to a Gartner forecast suggesting that by 2028, half of all organizations will adopt a zero-trust approach to managing data governance. By providing a verifiable path for data, DataGPT helps companies meet these requirements while still allowing employees to be productive.

Availability and company growth

DataGPT is available immediately through the connector marketplaces for ChatGPT, Claude, and Copilot. For companies that have not yet approved these specific AI tools, the service can also be run within Prophecy's own dedicated application. This provides a flexible entry point for different levels of corporate AI adoption.

The product can be rolled out centrally by an administrator or started by individuals using a free account. This tiered approach allows organizations to test the tool with small teams before committing to a full company-wide rollout. The service is already being used by major clients including SAP, HSBC, and Pfizer.

Prophecy has been active since 2017 and has raised more than $100 million in funding to date. Their most recent funding was a $47 million Series B extension led by Smith Point Capital in early 2025. This investment supports their mission to provide verifiable data workflows for the enterprise market.