Grupo Intelsis attended SAP House Chile 2025, where leaders, specialists and companies discussed process integration, operational efficiency and artificial intelligence.
The team spoke with clients and partners and attended presentations on business uses of technology. One covered the preparation and management of data used in AI projects.
A space to listen to clients and partners
Alongside the presentations, SAP House Chile offered a space to exchange experiences and discuss the challenges facing participating organizations.
These conversations help identify integration difficulties, differences in how indicators are calculated and obstacles to AI adoption. The team can use that context when assessing projects.
SAP BTP as the data foundation for AI
In his talk, Matías Müller explained how SAP BTP enables data management and what that means for artificial intelligence projects. His core point was that the platform integrates, governs and transforms data from many sources into a single, reliable layer, so AI models work with information that is high quality, current and grounded in business context.
In practice, that means pulling data out of the silos where it usually sits: the ERP, satellite systems, spreadsheets and partner applications. SAP BTP services such as SAP Integration Suite, which connects SAP and non-SAP systems, and SAP Datasphere, which organizes data and makes it available, help build that layer without customizing the ERP core, keeping it clean. For a business-level overview of the platform, read what SAP BTP is.
Why data quality sets the limits of AI
Müller also stressed that data quality directly determines three attributes of any AI model: accuracy, reliability and scalability. In day-to-day business terms, the effect looks like this:
- Accuracy: duplicate master records and incomplete history can compromise a model’s forecasts and recommendations.
- Reliability: sales, finance and operations need to understand and validate the source data to assess recommendations.
- Scalability: manual spreadsheet cleaning can make it harder to repeat a pilot in other business units or countries.
According to Müller, by reducing silos, improving traceability and enabling advanced analytics, SAP BTP supports smarter, data-driven decisions. He also made the case that investing in sound data management is a competitive advantage, not just a technology choice.
Checklist: is your data ready for AI?
To turn that message into action, we put together five checks to assess whether your company's data can support an artificial intelligence use case. Run them before you choose a tool or a model.
- Start with the use case. Pick one concrete decision, such as forecasting demand for a product line or prioritizing maintenance on critical assets, and map only the data it needs. Trying to organize all of your company's data first tends to stall the project.
- Give every critical data domain an owner. Customer, material, supplier and pricing data need a business owner who defines what each field means and approves changes. Without owners, quality becomes IT's problem alone and nobody fixes the source.
- Write quality rules and measure them. Define what a valid record looks like in terms of completeness, uniqueness, format and freshness, track those metrics and fix issues in the source system, not in the report.
- Document lineage. Record each data source, its transformations, and its last update. This helps teams investigate an AI recommendation and identify where to correct the information behind it.
- Govern access. Set role-based profiles, protect personal and sensitive data and log who accesses what. An AI model should never become a shortcut to information people aren't supposed to see.
Let's talk about your data
The Grupo Intelsis team can assess the data required for an AI use case. Our data and analytics and artificial intelligence teams support source assessment, governance design and architecture on SAP BTP.
