Book a meeting

Building an Internal Knowledge System with AI: How Sesame Access Is Transforming Information into Intelligence

Turning transcripts into intelligence for better lift engineering

The Sesame Victoria Chair Lift

Key Takeaways

  • Learn how Sesame Access built a structured, AI-driven internal knowledge system to convert transcripts and meetings into usable intelligence.

  • Discover how tools like Gumloop, Asana, and advanced LLMs (ChatGPT, Claude, Gemini) form the backbone of the Sesame Intelligence System (SIS).

  • Understand how the project, first piloted in Q2 2024, has already processed more than 1,200 transcripts, cutting internal search time by over 40 percent.

Introduction

Before 2024, engineers at Sesame Access were spending an average of five hours each week hunting through emails and notes to find past project decisions. This information loss slowed response times and created risk in technical projects.

To solve this, Sesame Access began building an internal knowledge system with AI — a structured framework that transforms every recorded conversation and meeting transcript into searchable, traceable knowledge. The result is the Sesame Intelligence System (SIS), a continuously evolving digital library designed to make company expertise instantly accessible and reusable across teams

How the AI Knowledge System Works

The SIS uses automation to move data through a clearly defined multi-stage workflow inside Asana:

  1. Capture (Q1 2024): Teams began recording structured meetings through Microsoft Teams channels.

  2. Normalise: Each transcript is automatically stored, labelled, and quality-checked in a single SharePoint location.

  3. Extract: Gumloop runs automations that feed transcripts to LLMs, which identify decisions, design notes, and key risks.

  4. Consensus & QA: Engineers validate AI-extracted data to maintain accuracy.

  5. Launch (Q3 2025): Approved knowledge items are exported into both the internal SIS and the external Knowledge Hub.

This process ensures that vital knowledge is retained, searchable, and auditable — replacing scattered emails with an intelligent, central source of truth.

Why AI Matters for Knowledge Consistency

Without structured AI support, valuable insights can vanish after each meeting. With SIS, every transcript becomes a searchable dataset. AI identifies repeating topics and potential overlaps, helping engineers connect related projects and anticipate design challenges earlier.

As Steven Lyons explained during the rollout, “I don’t want to take a transcript and populate the website the way ChatGPT thinks it should be populated — I want it processed through the runway.” This reflects Sesame Access’s guiding principle: AI assists, but humans approve.

Integrating Gumloop, Asana & LLMs

Sesame Access integrated Gumloop with Asana to automate document movement and status tracking:

  • Asana acts as the “runway,” managing each knowledge item’s lifecycle.

  • Gumloop handles the background automations and API calls to AI models.

  • LLMs such as ChatGPT, Claude, and Gemini provide semantic understanding, summarisation, and cross-validation.

Together, these tools have enabled the team to automate over 70 percent of manual documentation tasks, freeing engineers to focus on innovation and client delivery.

From Knowledge to Application

This AI foundation isn’t abstract — it drives tangible outcomes. Engineers can now retrieve precise technical insights in seconds when working on complex lift projects such as the Mayfair Stairlift, Wellington Lift, or Victoria Chair Lift.

By referencing previous design solutions, risk logs, and accessibility compliance notes, Sesame Access ensures that lessons learned in one project improve the next. This accelerates delivery, strengthens safety assurance, and reinforces long-term engineering quality.

Building Toward 2026: A Smarter, Searchable Future

The next development milestone, scheduled for Q1 2026, is a fully interactive AI search assistant inside the Knowledge Hub. Users will be able to type natural-language queries such as “Hidden Lift Design Standards: BS 6440 & Annex VIII” and instantly receive curated results drawn from SIS data.

This evolution turns Sesame Access’s website into a living knowledge ecosystem — one that grows smarter with every transcript uploaded.

Frequently Asked Questions (FAQ)

What is the Sesame Intelligence System (SIS)?
It’s an AI-powered internal platform that captures, processes, and organises company knowledge through an automated multi-stage workflow.

When did the SIS project begin?
The first internal pilot launched in Q2 2025, with a company-wide rollout during 2025.

How much data has been processed so far?
Over 1,200 transcripts and meeting notes have been structured, leading to a 40 percent faster average retrieval time.

Which tools make this possible?
Sesame Access combines Microsoft Teams, Asana, Gumloop, and advanced LLMs for automated classification and verification.

Can external partners access the SIS?
No. Only approved summaries appear in the public Knowledge Hub; internal data remains securely stored for authorised staff.

How can I learn more?
You can book a Teams Meeting with one of our Project Managers to explore how Sesame’s AI systems can enhance your project workflow.