A VALUED CUSTOMER OF THIRDEYE DATA

Inter-American Development Bank: finding the right expert by asking a question

The Inter-American Development Bank funds social and development projects across Latin America, working at a scale where programmes run across entire countries. An organisation like that runs on expertise, and its most valuable asset is often knowing which of its people has already solved a version of the problem in front of it. That knowledge tends to live in individual memory rather than anywhere searchable.

LOCATION

Washington, DC, United States

COMPANY SIZE

Large-scale Enterprise

INDUSTRY

Banking, Finance and Insurance › Explore solutions for this industry

AI-based knowledge management and expert discovery

Finding the right internal expert is a question that directories answer badly. A directory holds job titles, which describe what someone is employed to do rather than what they have actually worked on. The real record of expertise sits scattered through articles, blogs, publications and project documentation, in prose rather than in fields.

ThirdEye Data built a system that reads that material and turns it into something answerable. It processes the unstructured content and extracts what matters from it: the topics a piece of work covered, the domain it sat in, the people and organisations involved, the location, and when it happened. From that it builds a detailed profile for each person, linking them to the topics, domains, places and periods their work actually touched.

The interface is a conversation. Someone asks, in ordinary English, who knows about a particular subject, and gets back a considered answer with a ranked list of the people best placed to help. No query syntax, no guessing which category somebody filed their work under.

The difficulty in a system like this is not retrieval. It is ranking. Almost everyone in a large organisation has touched almost every topic at some point, so a system that returns everyone who matches is no better than a search box. The value is in judging depth.

Expertise across the organization made findable by asking a question in ordinary language.

Related solution: knowledge management and expert discovery

Most organisations have more knowledge than they can find. It accumulates in documents, reports and correspondence written for a purpose at the time and never indexed for any purpose since. The usual response is to ask people to tag things, which works for about a month.

ThirdEye Data builds systems that read the material as it is, rather than requiring it to be organized first. That means extracting the entities that matter, building profiles from evidence instead of self-declaration, and putting a conversational interface in front of it so nobody has to learn a query language. We have built this for a development bank finding internal expertise, for a publisher searching millions of research documents, for a court making case law genuinely searchable, and for a government agency extracting lessons from operational reports. The subject matter differs. The problem that the knowledge exists and cannot be reached does not.

Explore our knowledge management and enterprise search solutions ›