Test and Evaluate Any Model Now

Ridgeline Founder Stories: Ben Harvey of AI Squared aims to transform the way organizations analyze and leverage data




The story behind a technology intended for many uses, from making more sales and saving more lives.

As businesses and organizations move into the digital age, they are increasingly using AI solutions to streamline their operations and sales processes, understand their customers, predict behaviors, and stay ahead of the curve. AI Squared, a Ridgeline-backed company founded by Benjamin Harvey, is at the forefront of this revolution. Through cutting-edge AI solutions, AI Squared helps companies across industries integrate AI into their existing technologies and achieve higher levels of productivity, efficiency, and innovation. In June 2022, AI Squared received $6M seed funding to expand the product and meet growing demand. We asked Ben about his journey so far and predictions for the future of AI and AI Squared.

Note: This interview has been edited for length. 

How did you become interested in AI?

BH: I was inspired to pursue computer engineering after seeing my older brother Joseph attend college for it. I grew up in a disadvantaged household in the inner city, where success was often equated with rappers, drug dealers, and sports athletes, who were products of their environment. However, after seeing my brother physically leave the inner city and find success, I knew what I wanted to do. I pursued computer science as an undergraduate and PhD, but was introduced to genetics and genomics during my postgraduate research at Harvard and MIT. There, I applied computer science techniques to analyze large-scale cancer genomic datasets. This was during the time of the Human Genome Project, and my job was to analyze genomic information after it was sequenced. I was introduced to AI and machine learning during this time, and we used these algorithms to predict disease susceptibility for different patients based on their genetic makeup.

What inspired you to co-found a company?

BH: After leaving Harvard and MIT, I spent a decade at the National Security Agency (NSA) working in data science, including as Chief of Operations Data Science and Head of Data Science for the Edward Snowden leaks. As Chief of Operations Data Science, I created an operating model that deployed data scientists into mission spaces to work on data science problems. When I was Head of Data Science at the NSA, I had two brothers that were deployed in hostile territory overseas, one in Qatar and another in Afghanistan. I knew that the insights that our organization created from AI and machine learning could be the determining factor that can help them ultimately make a decision that could save their lives. This was a problem that I was trying to solve that directly affected members of my family. I knew that I wanted to try to figure out how to create a dual-purpose technology in an industry where I could come back to the federal government and be able to provide that as a solution to the problem that I had identified. But, the federal government has a lot of bureaucratic processes, like the Federal Acquisition Process (FAR), which makes it difficult and time-consuming to create innovative technology solutions for given problems and get them integrated into a mission production application. That’s where I wanted to try to figure out how I can work in industry, go to Silicon Valley, and create what could ultimately be a dual-purpose technology. And hence, that’s why we work with Ridgeline, because Ridgeline works well with companies that have dual-purpose technology. I wanted to infuse that back into the NSA. I understood that I had spent my entire career at the NSA, but I didn’t yet know how to build a company, how to incubate technology, or how to go to market with the technology. So, I left NSA and worked at Databricks for a few years before starting AI Squared.

Your company helps other companies integrate AI into their existing technology. What’s a typical use case of this, and have there been any non-typical or surprising use cases?

BH: In finance and retail, sales executives use CRM tools like Salesforce and Dynamics to convert leads into clients. However, they often have to switch between different machine learning tools to gather information and make decisions, which causes them to lose context and accuracy. AI Squared integrates machine learning capabilities directly into existing CRM tools to provide real-time results that are optimized for decision-making. This is also true for the intelligence community’s CRM tools, which provide enriched and contextualized information about terrorist targets to help analysts and warfighters make informed decisions and mitigate risks.

Tell us about your team: Who’s on it, and how did you meet?

BH: I met Jacob Renn, co-founder and Chief Technologist, at the National Security Agency in the Operation Data Science hub. I brought Jacob on because he was one of the smartest data scientists that I had ever met. Brian Landron, another co-founder, was one of the best engineering leaders I worked with at Maxar Technologies. Where things may seem as if they were impossible, Brian can always figure out a way from an engineering perspective to make things happen and bring solutions to some of the most challenging problems. I brought Ian Sotnek, Head of Product, on board for his background in cognitive neuroscience and product development. I met him in an AI policy class that I was providing a lecture for at George Washington University. Alvin McClerkin, co-founder and Chief Operations Officer, has been my friend since childhood and right-hand man, providing senior-level support. Lloyd Pierre, co-founder and Head of Field Engineering and Sales, has a strong federal background and has helped win Small Business Innovation Research grants. John Daly, co-founder and Director of Solutions Architecture, has a background in computer science from the University of Maryland and is knowledgeable in 3D and augmented reality.

Where do you see AI headed in the future?

BH: We’ve seen ChatGPT from a large language model (LLM)—and it is a type of conversational AI. Machine learning algorithms have been built over time and used in systems, but this is the first time where there has been a machine learning or AI capability with a human in the loop, where the human can ask questions and get responses that are based upon general knowledge that could be found across the internet or derived from interacting with other individuals. Many companies have complex machine learning infrastructures, but now they just want to be able to take the trained mode like a ChatGPT, and in minutes, integrate it into a currently existing application without all of the machine learning infrastructure that has historically been necessary to integrate these types of models.

What does success look like to AI Squared in the short term and long term?

BH: Currently, our main focus is achieving the milestones necessary for our Series A funding round. We aim to increase revenue to rank among the top 5% of seed companies and demonstrate the maturity of our product to attract customers faster. In the short term, we aim to increase speed, agility, and customer engagement, while achieving revenue, customer, and user growth milestones. In the long term, we plan to build a team that can help us become a unicorn. We will build our sales and marketing organizations and develop a product lead growth strategy around our open source technology. This will enable us to build a funnel of open source users that we can work with to gain more customers. We will also focus on building the community edition of our technology to grow a community around the AI Squared platform. This will allow us to offer a free version that integrates AI and machine learning into existing applications. Our goal is to expand how we work with large language models and integrate them seamlessly into existing applications.

Learn more about AI Squared at squared.ai. Stay up-to-date with the latest news on Ridgeline’s portfolio by subscribing to our monthly newsletter!

This article was originally published at ridgeline.vc.