What is an intelligent knowledge hub?
Organizations working on international development handle a great amount of data. They stand to benefit from knowledge management tools that consolidate available information. Laterite is building AI-powered intelligent knowledge hubs to centralize diverse types of data – from entire datasets to research papers to policy documents – into a cohesive and user-friendly format.
Laterite’s intelligent knowledge hubs are accessed through a chat format, like ChatGPT, and provide nuanced answers on the topic they’re trained on. To do this we first work with experts on a topic to curate a database of critical documents and get a holistic understanding of the concept. Then, we structure this data as a knowledge graph and connect the graph to large language model (LLM) ‘agents’ and knowledge graphs. The result is a sidekick specialist on a topic.
What makes these intelligent knowledge hubs so useful?
As researchers working in international development for 14 years, Laterite is aware of a common problem in our field: underused data. Information is often fragmented across various domains, many of which are reviewed once and forgotten. When a new program starts, it all goes back to square one and efforts are duplicated.
There is a better way!
Information is only valuable if it’s accessible. Imagine a diverse set of file types on a particular topic. What if you could organize all this data by their content, instead of file format? The intelligent knowledge hubs have an aerial view of all the knowledge related to your topic of choice and can connect the dots across file types. You can then access their insights through a user-friendly and secure chatbot interface.
Intelligent knowledge hubs also do more than just store a repository of data. They can generate entirely new insights and outputs. You can interact with knowledge hubs as a virtual assistant, can ask it to plot graphs, summarize a paper, interpret statistics, or draft a policy brief.
How is this different from ChatGPT or other chatbots?
Laterite’s knowledge hubs are specialists, while publicly available generative AI products like ChatGPT are generalists. Because they are trained with masses of data, ChatGPT’s responses are generic and lack depth, especially on low-resource contexts like sub-Saharan Africa. By working with experts on a specific portfolio of work, we can apply the power of LLMs to a narrow, curated set of information. This gives it access to important context and the data, research, and insights already available.
Laterite’s intelligent knowledge hubs are also capable of their own calculations and inferences, while many chatbots built today are not. The key distinction is in the way information is stored and accessed in the backend.
We will be sharing a technical brief on these distinctions in the coming weeks. Stay tuned!
Try it for yourself
This is EduStats – Laterite’s prototype for the Ministry of Education in Rwanda.
EduStats is a proof of concept, built on a selection of documents regarding education statistics in Rwanda. It can answer a wide range of questions on education in Rwanda, providing helpful tables, graphs and nuanced answers.
Here is the link to the ChatGPT integrated version, with additional features like two-way chat and graphing capabilities. (Note: signing up or linking to a Google account is required for both.)
Contact us for a demo
Get in touch if you’d like a live demo of EduStats or for any queries. We’d be happy to explore building a knowledge hub together for your specific social impact work.
Reach out to Senior Research Associate John DiGiacomo through jdigiacomo@laterite.com.