Startup tackles knowledge graphs to improve AI accuracy

The lack of reliable context has become a critical barrier to the adoption of AI agents, but startup Lovelace says it has solved the problem.

The answer, said co-founder Andrew Moore, the former head of Google Cloud AI, is his company’s new platform, Elemental, an AI-powered system for building knowledge graphs that he said is cheaper, faster, and more accurate than its competitors. It can help ground large language models (LLMs) in accurate context, he explained, while also providing full auditability so that enterprises know the exact information on which decisions were made.

“You cannot do safety-critical reasoning for agents purely based on trying to do the same sort of thing you normally do for chatbots,” Moore told CIO. “You need something else to help make sure that the reasoning that’s going on is properly coordinated.”

Given hallucination rates across 26 top models range from 22% to 94%, according to Stanford’s 2026 AI Index, released in mid-April, it’s an important problem to solve.

According to Moore, Elemental builds knowledge graphs from a customer’s own data, figuring out entities, relationships, and time and location attributes, as well as reporting the original sources of information.

For example, he said, “There are going to be 500 ships going through [the Strait of Hormuz, contested in the war with Iran] over the next week, and we know that some of them are going to have Iranian weapons in them. How do we figure out which ones are worth boarding to do a check on?”

With Elemental, the system can look at the history of the ships and their captains, their cargo manifests, even market and weather conditions, to figure out whether the ships are supposed to be where they are, or whether they’re acting in a suspicious way.

Elemental is not a replacement for an LLM, he explained, but something that helps companies improve the performance of the LLMs they’re already using. While it uses an LLM to help build the knowledge graph, once the graph is built, the process of using it involves traditional coding.

Knowledge graphs provide context

Carm Taglienti, public sector CTO at Insight, a Chandler, Arizona-based solution integrator, explained the value of knowledge graphs to AI. “Once knowledge is stored in the knowledge graph, and that knowledge can be used to set context or answer questions or produce results, then in that case, the probabilistic nature of the LLM can be made to operate in a more consistent fashion,” he said.

Ray Wang, principal analyst and founder at Constellation Research added that knowledge graphs can help integrate more information sources and provide better context for LLMs. “You can reduce errors, address hallucinations, and add more data sets,” he said.

As a side benefit, they also significantly reduce the amount of information a company needs to send to the LLM. According to Moore, complex queries can use tens of millions of tokens; Elemental can reduce this to around 10,000.

Lovelace also operates YottaGraph, a separate context engine that uses public and licensed data to enrich customers’ knowledge graphs. YottaGraph currently holds close to a trillion facts and is growing by about a billion a week. It is available as an experimental preview, free online for queries related to financial information.

This is all part of a move towards context engineering, in which companies focus on getting the LLMs better data, and knowledge graphs are a key part of the effort. According to DataM Intelligence, the global knowledge graph market was $1.34 billion in 2025, and is expected to increase to over $19 billion by 2033, growing at nearly 31% per year.

Lovelace ‘makes this super scalable’

Lovelace is targeting the intelligence and financial services sectors for its initial launch, which positions it directly against established players such as Palantir and Neo4j.

“We are compared to the early days of Palantir,” Moore said. “But the ability to reason on hundreds of thousands of new facts arriving every second, at the moment, the Palantir infrastructure can’t handle.” That makes Palantir, as well as AI model vendors themselves, potential partners for his company, Moore said.

Lovelace isn’t alone in this market, R.V. Guha, technical fellow at Microsoft, and former technical advisor at OpenAI, noted, but the work involved in creating schemas for the knowledge graphs has been a huge bottleneck. What makes Lovelace different, he said, is that Moore and his team “have done a bunch of clever things that make this super scalable. I haven’t seen anything quite like this yet from others, but I suspect that in 12 months this will become the norm.”