Microsoft announced the acquisition of Osmos, bringing agentic AI capabilities directly into Microsoft Fabric to automate time-consuming data engineering workflows. The Seattle-based startup specializes in using autonomous AI agents to transform raw data into analytics and AI-ready assets within OneLake, the unified data lake at the core of Microsoft Fabric. The acquisition aligns with Microsoft’s capital expenditure strategy and is projected to exceed $80 billion for the full fiscal year as the company reshapes its cost structure to prioritize AI infrastructure.
How Autonomous Data Engineering Changes Daily Operations
For technology executives managing ERP and enterprise data ecosystems, this acquisition signals a fundamental shift in how data preparation work will be performed. Organizations currently face a persistent challenge: teams spend most of their time preparing data rather than analyzing it, creating bottlenecks that slow business intelligence and analytics initiatives. Osmos addresses this by deploying agentic AI that operates autonomously alongside data professionals, reducing operational overhead for connecting, preparing, analyzing, and sharing data across the organization.
The integration into Microsoft Fabric creates direct implications for enterprises running SAP, Oracle or other ERP systems that generate massive volumes of transactional data requiring transformation before analytics consumption. Microsoft Fabric, launched in 2023 as a unified data and analytics platform, aims to consolidate all organizational data and analytics into a single secure environment. By embedding Osmos’s autonomous data engineering capabilities, Microsoft is positioning Fabric to compete more aggressively with Databricks, which also provides automated ETL tools on Azure.
This move reflects broader market dynamics where companies are investing heavily in AI and automation to eliminate data silos and accelerate time-to-market for innovative products.
Three Benefits for ERP Professionals
For ERP professionals evaluating data engineering platforms, several criteria become critical.
- Autonomous AI agents must demonstrate the ability to handle complex data transformation tasks without constant human intervention, directly reducing the labor intensity of data preparation.
- Integration with existing ERP systems and data lakes should be native rather than requiring extensive custom development.
- The platform must support real-time or near-real-time data processing to enable operational analytics that inform immediate business decisions rather than historical reporting.
Best practices for integrating autonomous data engineering into SAP or ERP environments emphasize starting with high-volume, repetitive data preparation tasks that currently consume significant manual effort. Organizations should establish clear governance frameworks that define when AI agents can operate autonomously versus when human review is required, particularly for master data management and financial consolidation workflows. The shift toward unified data platforms like Fabric requires enterprises to reassess their current point-to-point integration architectures and consider whether consolidating data engineering operations into a single platform reduces complexity and cost.
Common challenges in adopting autonomous data engineering include ensuring data quality when AI agents make transformation decisions, maintaining visibility into what changes agents are making to data pipelines, and managing the organizational change as data professionals transition from hands-on data preparation to oversight of AI-driven processes.
What This Means for ERP Insiders
Agentic AI shifts data engineering from human-led to AI-supervised operations. Microsoft’s Osmos acquisition fundamentally challenges the traditional data engineering operating model where human developers build and maintain ETL pipelines manually. This development pressures ERP vendors to embed autonomous capabilities within their own data management tools or risk becoming data sources rather than analytics platforms.
Unified data platforms threaten specialized integration and middleware markets. By consolidating autonomous data engineering directly into Fabric, Microsoft is building an end-to-end data and analytics ecosystem that reduces dependency on third-party ETL tools, data integration platforms and specialized middleware. Transformation leaders should assess whether Microsoft’s unified approach offers sufficient flexibility for complex enterprise requirements or whether specialized tools remain necessary.
Capital expenditure priorities reveal cloud platform competitive dynamics intensifying. Microsoft’s projected $80 billion annual AI infrastructure spend demonstrates the scale of investment required to compete in enterprise AI, creating barriers to entry that favor hyperscale cloud providers over regional or specialized competitors. For ERP vendors considering cloud strategy, this signals that partnership with major cloud platforms may be more viable than building independent AI infrastructure.
