Elqano Connects Enterprise Employees to Internal Expertise

01 /The Concept

The bigger a company gets, the less it knows about what its own people know. Elqano addresses that problem with a deceptively simple mechanism: a Microsoft Teams bot that connects employees who have questions with colleagues who have relevant experience.

An employee types a question into the Elqano bot – for example, describing a machine learning problem and asking whether anyone has tackled something similar. Elqano's AI engine searches the company's document repositories (SharePoint, in the typical deployment), identifies employees who have authored documents related to the query topic, ranks them by relevance – volume of relevant documents, recency, and terminological similarity to the question – and routes the question to the most likely experts.

Those experts respond directly in the chat, attaching relevant files or proposing a follow-up meeting. All Q&A activity also surfaces in a shared social feed, where any employee can like, comment, or tag additional colleagues who might contribute. A single question can propagate across departments through this layer, surfacing expertise the algorithm missed.

The bot continues learning over time. An initial indexing phase builds a map of the company's knowledge domain. A tuning phase of up to three months, guided by Elqano's team and employee feedback, sharpens the AI's ability to match queries to relevant documents. After that, the system updates automatically from ongoing usage signals.

Elqano targets mid-to-large enterprises, with three product tiers: Connect for 300–1,000 employees, Transfo for 1,000–150,000, and Transfo+ for companies above 150,000. The platform is built on an assembly of commodity components – PostgreSQL, Elastic Search, Azure hosting, Microsoft Teams for the interface – with the matching algorithm as the proprietary layer.

Founded in 2018, with a first version shipped in 2020, Elqano raised its initial funding of €900K at the end of 2022.

02 /Why It Matters

The theoretical foundation for Elqano dates to Ronald Coase's transaction cost theory of the firm – the Nobel-winning argument that companies exist because coordinating between independent actors is expensive, and internal organization reduces that cost. The original transaction costs were operational: contracts, supervision, handoffs. Computing and cloud services have largely automated those.

The remaining bottleneck is informational. As Bill Gates observed, the winners will be companies that restructure the flow of information internally. Knowledge about what employees know and have done is among the most valuable – and most poorly managed – of those information flows. Duplicate work, missed expertise, decisions made without awareness of prior solutions: these are the real costs of poor internal knowledge circulation, and they grow nonlinearly with headcount.

Elqano is not alone in recognizing the problem. A [previous review](/review/kak-najti-togo-kto-nuzhen) covered Beatrust, a Japanese startup with a substantially similar platform that raised $9.1M. The simultaneous emergence of similar solutions in different regions is a reliable signal that the underlying pain is genuine and not market-specific.

The platform's build-rather-than-assemble philosophy also stands out. Elqano assembled its stack from existing infrastructure – Teams, SharePoint, Postgres, Elastic, Azure – rather than rebuilding commodity layers from scratch. The proprietary value is the matching algorithm and the knowledge graph it builds. That approach reduces time-to-market and capital requirements significantly, which matters more at the €900K stage than it would at €9M.

03 /Opportunities

Large enterprise clients are the target here, and while the sales cycle is longer and more complex than SMB, the contract values justify the friction. The more salient consideration is whether the platform's design can keep pace with AI development.

ChatGPT-style interfaces that answer questions directly from indexed document corpora are an obvious evolution of the Elqano model. The next iteration of this category doesn't just route questions to human experts – it answers a portion of them automatically, using the same indexed document set, and escalates only what the AI can't resolve. That hybrid architecture reduces the burden on employees who currently have to respond manually to every routed query, and it creates a much faster response time for common or well-documented questions.

Any new platform in this space should be designed with that AI dialogue layer in mind from the start, rather than retrofitting it later. The companies that build the knowledge graph and the AI interface as integrated components – rather than adding a chatbot on top of a document search tool – will have a structural advantage as enterprise buyers start evaluating both capabilities together.