Learnings for transforming experimentation into real business impact
Introduction
Today, Artificial Intelligence holds a central position on the organizational agenda. However, as interest grows, a gap is becoming increasingly evident between initial expectations and the actual impact many initiatives manage to generate for the business.
At BCI Consulting, we have been working alongside organizations across various industries to tackle this challenge. Recently, in collaboration with IAE Business School, we hosted an exchange session with IT, Finance, and Operations leaders to discuss how to advance with AI in a concrete, sustainable way that is aligned with the business.
This article synthesizes the key insights currently shaping organizations and analyzes them in light of recent data and studies (Q4 2025).
Why Do Many AI Initiatives Fail to Scale?
One of the most repeated patterns is starting AI projects from a technology-first rather than a process-first perspective. The result is usually the same: pilots that work in controlled environments but fail to scale or integrate sustainably into the business.
Several recent studies agree that the majority of AI initiatives do not make it past the testing stage. The causes are not related to the capability of the models, but rather to structural factors such as poorly prepared data, difficulties integrating with existing processes, and unrealistic expectations regarding timelines and results.
The Role of Data in Artificial Intelligence Adoption
AI does not create value on its own. Its impact depends directly on the quality, traceability, and governance of the data it operates on.
In organizations with consolidated core systems—such as SAP S/4HANA—transactional data becomes the primary enabler for applying AI in an integrated and scalable way. Without reliable data, there is no sustainable automation or real improvement in decision-making.
This also applies to the use of large document bases (document grounding), where the relevance and context of the data are more important than the volume.
Real Cases of AI Applied to Business
When AI is applied to specific processes, results begin to emerge. Across different sectors, real impacts are already being observed in:
- Automation of repetitive administrative tasks.
- Analysis and validation of documentation.
- Support for clinical decision-making.
- Fraud and anomaly detection.
- Optimization of planning and supply chain.
- Assistance for developers and technical teams.
In healthcare organizations, for example, the application of AI in diagnostic image analysis is reducing processing times and improving the quality of information that the medical professional subsequently validates.
In financial areas, the intelligent automation of invoicing and document validation is beginning to offload a significant administrative burden, allowing teams to focus on higher-value tasks.
People, Change Management, and Expectations
AI adoption is, above all, a process of organizational change. One of the main challenges is expectation management: the belief that technology, on its own, will solve complex problems or immediately reduce organizational structures.
Recent analyses distinguish between direct financial Return on Investment (ROI) and Return on Employee (ROE): improvements in productivity, decision quality, and the reduction of operational friction. In many cases, the initial value of AI appears first on this second level.
Embedded AI in Business Processes: The SAP Approach
One of the most significant evolutions of AI in corporate environments is its direct integration within core systems. SAP has advanced in this direction with Joule, the AI layer embedded in SAP S/4HANA, and the development of intelligent agents that operate on real business processes.
These agents allow for the automation and assistance of specific tasks such as:
- Contextual search for customer and order information.
- Procurement summaries by vendor and period.
- Bulk copying and modification of sales orders.
- Validation and approval of operational changes.
- Monitoring and embedded analytics on critical processes.
This approach reduces operational friction and facilitates progressive adoption while keeping people “in the loop” of the decision-making process.
Conclusion
Artificial Intelligence is not an immediate destination, but a journey—a journey that requires a focus on processes, data quality, change management, and clear rules.
The competitive advantage does not lie in adopting AI first, but in adopting it with sound judgment, aligned with the business, and prepared to scale.
About BCI Consulting
BCI Consulting is a technology consultancy specialized in business transformation, focusing on critical processes, data, and the adoption of AI integrated into SAP environments. We support organizations in defining realistic and sustainable roadmaps.
📄Download the Paper
AI in Organizations: From Experimentation to Real Impact
