How Unisense turned any employee into a sales expert with +15 years of experience.

We wanted to go from scattered inboxes to shared intelligence.

Like many B2B organisations, Unisense had spent over 15 years building deep expertise across sales, operations, pricing, and technical. The problem? Almost all of it was buried in individual Outlook inboxes, scattered documents, and datasheets that no one could find easily.

New employees had limited access to historical decisions, and experienced team members spent valuable time digging through old email threads to recall past decisions. Traditional search wasn't enough: it required knowing the right keywords and then manually reading through every result. Unisense needed a smarter way to surface that institutional knowledge and make it available to everyone, right where they already work.

Read more about Unisense.

We built a source-grounded AI assistant directly into outlook.

To solve this, Unisense partnered with Obsidian to build a Retrieval-Augmented Generation (RAG) assistant designed from the ground up for accuracy, transparency, and scalability.

Using a custom Python workflow, we processed and indexed over 375K sources across emails, documents, datasheets, and company information. To reflect the distinct nature of the two brands, the knowledge base was split into separate indexes for Unisense and Unisense Environment. The assistant was then integrated directly into Outlook, sitting behind the same login wall as Unisense's other internal systems.

In practice: When an employee reads an email, that email is automatically used as context: the system retrieves the most relevant sources from the RAG database and delivers a direct, cited answer - no document list, no manual reading.

Outdated sources can be removed dynamically, and system instructions can be edited on an ongoing basis. This scalable knowledge base also powers an internal open knowledge tool for company-wide questions, and an external website chat capable of answering simple customer inquiries without citing internal sources.

A scalable company intelligence, embedded in the daily workflow.

The implemented solution successfully transformed unstructured email data and company information into a persistent knowledge layer that grows over time. By prioritizing knowledge delivery over automated email writing, the solution enabled faster onboarding, stronger internal alignment, and more efficient decision-making.

New employees can now get up to speed faster, specialists can validate past decisions instantly, and valuable insights remain accessible and verifiable as teams evolve. The architecture is fully prepared to scale across thousands of additional emails and documents for both brands.