AI, Data, Teams and Innovation

I have spent my life building high-performing teams ranging from my startups to roles at the largest tech companies in the world. The one truth that has never failed me is that the investment in hiring and retaining the best possible people pays for itself over and over repeatedly. Different approaches and strategies exist to do this, but it is always the first principle I enter any enterprise with: find, recruit, empower, retain. 

The key, in my opinion, starts with empowerment. The people who provide the most impact in an organization are not only experts in their craft, but also seek to improve and increase impact as core principles in how they work. I think of them as “drivers” of positive change. Drivers know their rare value, so the investment in creating a culture that brings these people into the fold is the first principle of building and competing in technology (and other areas). When a new technology emerges, top talent gravitates to other top talent around the opportunity. Departures of less motivated individuals are backfilled with more motivated ones seeking to be associated with a tier-1 empowering team. When a problem exists across organizations, the driver seeks out the stakeholders and facilitates an improvement considering stakeholders’ and the primary business concerns. 

One of the great things about moving from building new products directly to the venture side of supporting innovative product-building technology companies is encountering new teams, new approaches, and new ways of solving these same challenges. As a Partner and CTO at Inovia, I have encountered many such teams. In this time of significant technology capability shift brought by transformer architectures and generative AI possibilities, the teams created and nurtured as empowered drivers are the first to apply, exploit and extend their capabilities. I think of them as AI-tailwind teams. When coupled with experts in customer pain points, we are seeing magic happen in a step change with product and value growth in all metrics and the best of all – delighted customers.

One such company is Neo4J. Neo4j has been entrusted for nearly two decades by innovative organizations from huge to small with the most valuable of all assets: their data. Along the way, they have developed a technical capacity on par with any product and technical organization in the world. Specifically, they are fundamentally informed by customer needs and by technical and product drivers with a fierce commitment to customer value, and appropriate application of new technologies and capabilities that will help their customers, in turn, drive value to their own customers. A better example of the AI-accelerated category is hard to find.

If you are the company entrusted with assisting companies themselves to drive value to customers through better organized and effective data, you are at the core of their business. As we have outlined in our recent whitepaper

“The coming iterations [of LLMs] will likely not just be bigger LLMs but will be augmented with additional tech like Retrieval Augmented Generation (RAG), etc, to improve explainability and patch some of the LLM weaknesses. As these iterations advance, tools that enable LLMs to be grounded in fact and internal data (i.e. knowledge graphs and vector databases) will be embedded as a core part of the Generative AI tech stack, with knowledge graphs growing in importance as users look to LLM-enabled solutions to answer more complex queries. “

RAG is a solution to what is commonly referred to as hallucinations – where LLMs present false information as fact. As enterprises are deploying LLMs into products and workflows, RAG has been widely adopted to ground LLMs, in fact, by using their own internal data. 

Developers have seen success using vector embeddings for most RAG solutions today, but this technique relies on similarity search and struggles to support complex queries and advanced reasoning. Enter Neo4J knowledge graphs for LLMs. The graph structure of Neo4j makes finding factual answers to complex queries possible by capturing the deep contextual relationships in enterprise data. 

It has been a pleasure to watch the drivers at Neo4J capitalize on their position as custodians of enterprise data to embed themselves as a core part of the New Gen AI Tech Stack and earn customer trust by delivering new and expanded value to their users. 

If you want to learn more about how we think about application companies in the generative AI space, read our blog about our latest investment here.