Expert Intelligence is a primary player in the infrastructure layer of the AI agent stack, specifically focusing on the Model Context Protocol (MCP). They are one of the few service providers explicitly offering a path for enterprises to migrate legacy data into MCP-compatible knowledge graphs. This is critical for developers building agents that need to perform complex reasoning over proprietary or domain-specific data without the risk of hallucination.
Their relevance lies in their ability to turn "dirty" or unstructured enterprise data into a machine-ready format. By advocating for semantic ontologies over pure RAG (Retrieval-Augmented Generation) based on vector search, they are pushing the ecosystem toward more deterministic and reliable agentic behavior. They are active in the transition from simple chat interfaces to sophisticated agents that can navigate complex organizational rules and relationships.
Expert Intelligence is a consultancy and implementation firm focused on the transition from statistical AI to semantic reasoning. Established in 2022, the company is a specialized arm of Point Topic, a UK-based data provider with over 30 years of history in the broadband and telecommunications sectors. Based in Lisbon and London, the team is led by Oliver Johnson and Arnold Kürsteiner, who bring decades of experience in ETL pipelines, audience modeling, and earth sciences to the problem of machine-readable data.
The core of the Expert Intelligence thesis is that modern large language models are only as useful as the structures they can access. They argue that organizations are currently drowning in data that is unusable for AI agents because the meanings and relationships within that data are not explicit. To address this, they provide a tiered service model that begins with a "Semantic Audit" to assess machine readiness and culminates in full semantic implementation.
Perhaps the most relevant part of their current work for the agent community is their focus on the Model Context Protocol (MCP). As AI agents increasingly require access to private, complex datasets, the industry has moved toward MCP as a standard for tool-use and data retrieval. Expert Intelligence has productized this through their "my first MCP" offering. This service provides organizations with a lightweight implementation of an ontology and a knowledge graph that exposes data through MCP servers.
This approach effectively turns a company's internal knowledge into a logical system that an agent can query with high confidence. By creating these explicit maps, they move away from the risk of the "black box" LLM making incorrect inferences about organizational data. Their work is a practical application of the semantic web, focusing on how names, meanings, and relationships can be mapped to create reliable outputs for automated systems.
While their methodology is applicable to any data-heavy industry, Expert Intelligence maintains a strong foothold in the broadband sector. Their parent company, Point Topic, provides the underlying domain expertise that allows them to build highly specific models. This history is visible in their focus on geographic and human-centric data—asking where people are, how they live, and what characteristics define their markets.
The company sits in a unique competitive space. They are not a general-purpose AI software vendor, nor are they a typical high-level strategy consultancy. Instead, they operate as technical architects for the data infrastructure that agents require. For organizations that have spent heavily on data collection but struggle with "machine readiness," Expert Intelligence provides the translation layer necessary to make that data functional in an agentic workflow. Their leadership team, including Knowledge Director Arnold Kürsteiner and Planning Director Peter Donaghey, emphasizes a human layer in analysis, ensuring that the initial classification of data relationships is grounded in actual domain expertise before being handed over to automated reasoning engines.
Assessment of an organization's current data readiness for machine reasoning.
Lightweight implementation of ontologies and knowledge graphs using Model Context Protocol.
Expert Intelligence is hiring.