Vale accelerates ERP migration with Accenture’s AI, gaining time for planning

3–4 minutes

Mining company Vale is using Accenture’s Cognitive Digital Brain AI platform to modernise its SAP enterprise resource planning (ERP) system. The project, known internally as Neo Vale, involves migrating to SAP’s latest software version, S/4HANA. The tool is used in workshops where business teams outline their needs; it records discussions and converts requests into technical requirements, specifications, code, and even test plans.

During the initial phase, the AI processed over 270 requirements from the commercial, maintenance, and financial departments. According to the companies, this brought forward the definition stage by up to three weeks—a task that previously required several rounds of interaction between consultants and users. The extra time was used for more detailed planning rather than shortening the overall schedule, explains Theo Penedo, Vale’s technology director, who represents the business in the project.

The change is significant because Vale’s current ERP connects to over 350 systems and serves more than 30,000 users. The project was divided into two parts: a brownfield component, where processes are transferred without changes, and a greenfield component, where processes are revised from scratch. Only areas with the greatest potential for improvement underwent a complete redesign.

From workshop to code

Cognitive Digital Brain records meetings, extracts requests, and suggests adjustments in SAP. When the standard system does not meet a requirement, the platform automatically generates the code for the necessary extension and, if integrated into the development environment, activates it immediately. It then creates a test plan and specifies the data for execution. Everything is reviewed by Accenture consultants before being approved by Vale.

Daniel Gonzalez of Accenture says the tool also creates a traceability matrix of functional and technical requirements and specifications. Additionally, it can connect to the legacy ERP, map all existing configurations, relate them to business processes, and compare them with industry best practices. This preliminary diagnosis allows teams to arrive at workshops already prepared.

Impact on work and the team

The AI takes on tasks previously performed by junior professionals, such as writing documents and drafting program outlines. Senior consultants are left with review and decision-making tasks, which require experience and judgment. According to Gonzalez, the goal is not to cut jobs but to free up specialists to spend more time discussing with clients what has been produced.

Accenture, which laid off 11,000 people in September 2025, says only those who learn to work with AI will remain. At the same time, the consultancy is expanding teams in the Northeast, using AI as justification for new hires.

Challenges beyond technology

Accenture uses Cognitive Digital Brain for around 40 clients in Latin America, most of them in Brazil. The biggest impact, according to the consultancy, is in the natural resources sector, which is why Vale’s case was chosen to illustrate the tool. The first challenge was technical: to go beyond what existing AI solutions already offer. The solution was to incorporate Accenture’s industry knowledge, so that specifications and code are already adapted for mining, for example.

The second challenge is organisational. Making AI work in parallel with consultants and client teams requires a redistribution of tasks and changes in routines on both sides. The consultancy says the biggest obstacle now is reorganising people’s work around the new technology.

Why the migration is urgent

SAP will end standard support for the ECC ERP at the end of 2027, with an optional paid extension available until 2030. Large companies still using ECC are accelerating migration to avoid additional costs and operational risks.

For Vale, in addition to meeting the two-year deadline and maintaining the budget, the project aims to improve indicators such as DSO (days sales outstanding), the average time for receiving sales payments.

Conclusion

Vale’s experience demonstrates how consultancies are packaging decades of industry knowledge into AI agents in a market where generic models are no longer sufficient. The tool is available in Brazil, but Accenture does not disclose prices. What is clear is that artificial intelligence is redefining not only what is possible to do but also how people collaborate to deliver large projects.