AI and Large Language Models: A Reliable Solution Eludes Them

The Promise and the Problem

Artificial Intelligence (AI) and large language models (LLMs) have numerous practical applications, but their reliability is a significant concern. Despite their promise, these models are not very reliable, as highlighted in various studies and articles.

The Need for Enterprise Solutions

No one knows when this problem will be solved, so it makes sense that startups are finding opportunities in helping enterprises ensure that LLM-powered apps work as intended. This is particularly true for companies that are paying for these apps and want to ensure they are reliable and accurate.

Composo: A Startup with a Head Start

London-based startup Composo feels it has a head start in solving this problem due to its custom models that can help enterprises evaluate the accuracy and quality of apps powered by LLMs. The company’s approach is similar to that of other startups, such as Agenta, Freeplay, Humanloop, and LangSmith, which claim to offer a more solid, LLM-based alternative to human testing, checklists, and existing observability tools.

What Sets Composo Apart

Composo claims to be different because it offers both a no-code option and an API, which widens the scope of its potential market. This means that domain experts and executives can evaluate AI apps for inconsistencies, quality, and accuracy themselves, without needing to be developers.

How Composo Works

In practice, Composo combines a reward model trained on the output a person would prefer to see from an AI app with a defined set of criteria specific to that app. For example, a medical triage chatbot can have its client set custom guidelines to check for red flag symptoms, and Composo can score how consistently the app does it.

Launch of Public API

The company recently launched a public API for Composo Align, a model for evaluating LLM applications on any criteria. This move is expected to further expand the company’s reach and capabilities.

Customer Base and Funding

Composo has already secured a customer base with prominent companies like Accenture, Palantir, and McKinsey. The company recently raised $2 million in pre-seed funding, which is a notable achievement given the abundance of funding available in the AI landscape.

Low Capital Intensity

According to Composo’s co-founder and CEO, Sebastian Fox, the relatively low amount of funding is due to the startup’s approach being not particularly capital-intensive. Fox, a former McKinsey consultant, believes that the company’s focus on scaling its technology across existing clients will be more important than raising hundreds of millions of dollars.

Expansion Plans

With the fresh cash, Composo plans to expand its engineering team, acquire more clients, and bolster its R&D efforts. The focus from this year is on scaling the technology that the company already has across those companies.

Investment and Support

British AI pre-seed fund Twin Path Ventures led the seed round, which also saw participation from JVH Ventures and EWOR. Twin Path Ventures’ spokesperson stated that Composo is addressing a critical bottleneck in the adoption of enterprise AI.

Addressing the Bottleneck

Composo’s CEO, Sebastian Fox, believes that the bottleneck is a significant problem for the overall AI movement, particularly in the enterprise segment. Fox emphasizes that people are over the hype of excitement and are now thinking, "Well, actually, does this really change anything about my business in its current form?"

Industry Agnosticism and Resonance

Composo has chosen to be industry agnostic but still has resonance in the compliance, legal, healthcare, and security spaces. This approach is expected to make the company more valuable to companies that want to implement AI but could incur reputational risk from doing so.

Competitive Moat

Composo feels that the R&D required to get here is not trivial. The company’s approach is based on a large dataset of expert evaluations, which provides a competitive moat. Composo also believes that its first mover advantage and the data it accrues over time will help it stay ahead of competitors.

Agentic AI and Flexibility

Composo sees itself as better suited to the rise of agentic AI than competitors that use a more constrained approach. The company’s flexible set of criteria allows it to assess apps in a way that is not possible with more rigid approaches.

Conclusion

Composo is a startup that is addressing a critical bottleneck in the adoption of enterprise AI. With its custom models, no-code option, and API, the company is well-positioned to help enterprises ensure that LLM-powered apps work as intended. Composo’s approach is expected to make it a valuable player in the AI landscape, particularly in the enterprise segment.