Thought Leadership: Knowledge graphs is the missing layer between enterprise data and trustworthy AI
July 20, 2026
Boston, MA, July 20, 2026 - AI81 Insights – Thought Leadership. Insights from Dr. Satyajeet (Satya) Raje, AI81Works Technology Advisory Board member and global data science leader at Samba.
Most organisations are not short of data. They are short of connected data. Customer records sit in one system, behavioural signals in another, content and product data in a third. Each answers a narrow question, but none can explain how the pieces relate. As enterprises move from AI pilots to AI agents that act on their behalf, that missing context becomes the single biggest risk to trust, accuracy and value.
Dr. Satya Raje has spent his career on that problem, from his years as Chief Ontologist and CTO lead for data governance and data science at IBM to his current work at Samba. Samba is a global media intelligence company that recently introduced the Samba Knowledge Graph (SKG). His view is simple: knowledge graphs are no longer a niche data-engineering tool. They are becoming core infrastructure for enterprise AI.
From Identity Graphs to Knowledge Graphs
Traditional data platforms are good at matching records. They can tell you that two profiles belong to the same customer. What they cannot do is reason about behaviour, context and meaning. A knowledge graph models the relationships between people, content, products and events. Systems can then ask not only “who is this?” but also “what do they care about, and what are they likely to do next?”
At Samba, the SKG applies this approach to a fragmented media landscape. It forms a semantic layer over the company’s first-party data, linking television viewing, web behaviour, content, ad exposure and audience signals through a single identity spine. The result is audience intelligence built on observed behaviour across screens rather than broad demographic buckets, delivered in a privacy-conscious way and without relying on third-party cookies.
Why It Matters for Every Enterprise, Not Just Media
The lessons travel well beyond advertising. Whether in healthcare, manufacturing, financial services or retail, the same pattern applies:
Context beats volume. Billions of records have limited value until their relationships are modelled. Connected data turns description into prediction.
Grounded AI is trusted AI. Large language models hallucinate less when they reason over verified, structured knowledge instead of raw data. Knowledge graphs give retrieval-augmented generation a reliable foundation.
Agents need a map of the business. Agentic AI systems that take actions on their own need an explicit model of how customers, products, assets and rules relate. A knowledge graph provides that map.
Privacy by design is a competitive advantage. Modelling behaviour at the level of relationships and segments reduces dependence on fragile identifiers and supports compliance with tightening privacy rules.
Governance becomes visible. Ontologies make definitions, ownership and lineage explicit, which is exactly what data governance programmes struggle to achieve with documents alone.
Leadership Quotes
“Every organisation is racing to deploy AI agents, but an agent is only as good as its understanding of the world it acts in,” said Dr. Satya Raje. “Knowledge graphs give AI that understanding. They connect data into meaning, and meaning into decisions you can trust and explain. The companies investing in that foundation today will lead the next decade of AI.”
“Satya’s work shows what separates AI that impresses in a demo from AI that performs in production,” said Carl-Johan Nakamura, Co-Founder & CEO of AI81Works LLC. “Trust in AI starts with trust in data, and knowledge graphs are one of the most powerful ways to build it. We are proud to have Satya on our advisory board as we help clients design that foundation.”
What Leaders Should Ask Next
For executives planning their next phase of AI investment, four questions are worth putting to the leadership team:
Do we have a shared, machine-readable model of our core business entities and how they relate?
Are our AI assistants and agents grounded in verified enterprise knowledge, or only in documents and raw tables?
Can we explain where an AI recommendation came from and which data shaped it?
Who owns our ontology, and is it governed with the same rigour as our financial data?
Readers who want the technical depth can find the full Samba Knowledge Graph white paper, co-authored by Samba CEO Ashwin Navin and Dr. Satya Raje, at https://samba.com/insights/the-samba-knowledge-graph-powering-the-next-generation-of-media-intelligence
About the Author
Dr. Satyajeet (Satya) Raje is a member of the AI81Works Technology Advisory Board and leads global data science at Samba. He was formerly Chief AI Officer at AnthologyAI and Chief Ontologist and CTO lead for data governance and data science at IBM. He is an industry expert in data fabric and knowledge graphs.
Media Contact
For more information or to schedule an interview, please contact:
Liliana Trinkle
Email: ltrinkle@ai81works.com
Phone: +1-781-650-6800