The company's vision extends beyond the chat service, aiming to build an AI platform for analytics and solve various data problems. Numbers Station has already attracted several Fortune 500 customers, including global real estate services firm Jones Lang LaSalle. The platform learns continuously, enabling data teams to discover and verify hypotheses for impactful business outcomes.
Key takeaways:
- Numbers Station, a startup using large language models (LLMs) for data analytics, is launching its first cloud-based product, Numbers Station Cloud, which allows users to analyze their internal data using a chat interface.
- The company has developed a semantic catalog, an automatically curated source of a company’s metrics and definitions, which is specific to each company and ensures the model's understanding aligns with the company's usage of terms.
- Co-founder and chief scientist Ines Chami admits the team had underestimated the challenge of building the semantic catalog, which is crucial for turning vague questions into concrete queries.
- While the company is launching this chat service, the overall vision is to build an AI platform for analytics, with potential applications including enriching data with third-party sources and implementing classic algorithms like fuzzy matching.