AI and Graph Data Technology: Why 1 + 1 = Much More

AI is revolutionizing business models and technology infrastructures.

Graph database technology, critical to enabling social media, e-commerce, mobile apps, and more, is revolutionizing data management.

But what about the intersection of the two?

• Are they fully independent concepts with little crossover value?

• Can graph databases make AI applications better?  And if so, how?

• Or is it the other way around? Can embedding AI capabilities within graph database technology make graph databases faster to set up, easier to use, and more effective to operate?

In other words, when it comes to AI and graph databases, does 1 + 1 = 3, 4, 5 or more?

Meet the Math Majors

To help us with this math, we recently lined up an outstanding group of graph database technology and AI thought leaders in AI and Graph Databases: Where 1+1=3 (Or More) | The Jam Session.

• William McKnight, President of McKnight Consulting Group: I’ve known William for at least 15 years. He is a top-ranked influencer in big data, cloud, and analytics, the author of three books, and, most importantly, a man who has led large-scale data management projects at some of the largest organizations in the world.

• Danfeng Xu, Co-founder of PuppyGraph: After a distinguished engineering career at LinkedIn, Facebook, and Microsoft, Danfeng started PuppyGraph with the mission of radically simplifying querying a data warehouse and lake as a graph regardless of structure, avoiding traditional graph database setup.

• Imran Chaudhri, Chief Architect, AI, Healthcare & Life Sciences at Progress: Imran is a software Industry veteran serving in critical roles such as CTO, Chief Architect, and founder and advisory board member of several software firms.

AI + Graph Use Cases

We started by discussing use cases where graph database technology enabled AI applications. Examples included a global-scale legal knowledge base, real-time fraud detection, and risk management. Common to these was the need to utilize seemingly unrelated data relationships across diverse sources.

Removing Barriers to Graph + AI Adoption

Next, we discussed the barriers constraining the greater use of graph database technology. According to the panelists, the biggest issue is the learning curve required for data management professionals steeped in relational technology to master graph-style queries and network-based data models.

PuppyGraph’s AI-driven text-to-query capabilities were purposefully designed to shorten this learning curve from weeks and months to minutes and hours.

Another barrier has been the need for graph-specific databases such as Neo4J. But new graph query tools now allow graph-style queries from various database types, including multi-modal databases such as Progress’s MarkLogic Data Platform.

The AI inside Graph Opportunity

We asked the experts whether embedding AI capabilities within graph database technology could make it faster to set up, easier to use, and more effective to operate. The answer was a strong yes. AI-enabled graph capabilities included intelligent data and schema discovery, automated schema and query development, adaptive query execution, and more.

Best Practices and Other Advice

The experts also shared advice on AI + graph starting points, best practices, and lessons learned.  These included

• Focus on the Right Use Case: Select early use cases where AI + graph technology are both essential. This prevents reverting to prior methods whenever an issue arises.

• Deliver Value Quickly: Keeping your initial project scope small is okay. You will earn more resources to expand your efforts if you show results.

• Look for Shortcuts: Consider graph-style query enablers that query existing data in place. These tools avoid the need to lift and shift data into graph-specific databases as a project prerequisite.

Learn More

To watch the entire expert session, check it out on the Insight Jam Channel on YouTube.