Fern Halper, Ph.D.

The DGI Data Governance Framework addresses funding two ways. But what type of funding is needed? Data Governance programs need to establish funding for four types of Data Governance and Stewardship efforts. Read more here: http://ow.ly/ppvC30rIP2Y #PracticalGuidance #DataGovernance

3

Insights from AIM Research's 2025 PeMa Quadrant on Top Mid-Market and Specialized Data Science Service Providers 𝗚𝗧𝗠 𝗦𝗵𝗶𝗳𝘁𝘀 𝗦𝗵𝗮𝗽𝗶𝗻𝗴 𝘁𝗵𝗲 𝗠𝗮𝗿𝗸𝗲𝘁 - The data science services market is undergoing a decisive shift in how vendors go to market. Firms are moving away from one-off projects toward productized AI systems, agentic workflows, and repeatable accelerators that reduce buyer risk and compress time-to-value. GenAI co-pilot workshops, no-code agentic platforms, and domain-specific accelerator suites are becoming standard entry points, giving buyers faster clarity on ROI before committing to scaled engagements. - Distribution is increasingly hyperscaler-led. AWS, Azure, GCP, Snowflake, and Databricks have evolved from infrastructure providers into deal accelerators, with co-sell motions and joint GTM programs now central to how vendors win enterprise accounts. Partner statuses and dedicated cloud CoEs are emerging as meaningful differentiators. - Verticalization is now the default motion. Buyers in financial services, healthcare, retail, CPG, and manufacturing want domain-tuned solutions, not generic AI promises. Vendors are packaging vertical depth into repeatable accelerators across trade promotion, supply chain, compliance, and patient analytics, shortening time-to-value and strengthening win rates. - Once in the door, the best vendors grow through expansion rather than new logos. The market consistently shows 70-80% of revenue coming from deepening existing accounts via embedded pods and co-innovation programs. AI budgets scale only after proof of operational value, making land-and-expand the defining GTM motion. - Thought leadership has become a genuine demand-generation tool. Roundtables, executive forums, and partner events now qualify deals early in a crowded market where buyers are skeptical of unproven claims. Vendors investing here are earning advisor status before the RFP begins. - M&A is being used to plug capability gaps, expand geography, and accelerate domain depth, particularly in agentic AI, computer vision, and security. - Messaging has pivoted from model sophistication to measurable business outcomes. Buyers are now evaluating vendors on revenue uplift, cost savings, and operational efficiency. Engagement models are evolving alongside this, efficiency gains are real, and experimentation with how value is structured and priced is actively underway. Check the complete report here: https://lnkd.in/gwZg4RHZ Featured Vendors in the study (in alphabetical order): Aays, Acuity Analytics, Agilisium Consulting, Aventra Group, Blend, C5i, Chetu, Inc., Coforge, Dentsu Global Services, Evalueserve, Exponentia.ai, Fractal Analytics, Happiest Minds Technologies, Indus Net Technologies (INT.), InfoCepts, JK Tech, LatentView Analytics, MathCo, Polestar Analytics, Prodapt, Qubika, Simform, Straive, Tiger Analytics, Tredence Inc. #DataScience #AI #GenAI #AgenticAI #MLOps #PeMaQuadrant #AIMResearch

18

Krishna Govindan of Discover just published an excellent deep-dive on implementing Graph RAG at scale. 📈 We're excited to see PyAirbyte playing a key role in their data pipeline architecture. Graph RAG systems need intelligent Change Data Capture (CDC) for unstructured data sources. Modified files must be identified, processed incrementally, and integrated into knowledge graphs without reprocessing entire document sets. Krishna's solution leverages PyAirbyte with Airbyte source connectors to: ‣ Detect and process updated files from GitHub, SharePoint, and Google Drive ‣ Define primary keys and cursor fields for incremental updates ‣ Maintain citation anchors during the query phase This is exactly the kind of real-world AI data pipeline we love seeing - where structured and unstructured data movement becomes the foundation for more accurate, context-aware AI applications. https://bit.ly/3IuWih3

8