Knowledge Base in aiKO

A knowledge system that thinks like humans

Company

Aitomatic

Position

Head of Product & Design

Year

2024

Market

APAC & US, B2B

Problem

In manufacturing, expert engineers hold invaluable knowledge. They understand how machines operate, how to troubleshoot failures, and know which processes produce the best results.

But their critical insights often remain fragmented across emails, documentation, and casual conversations. When these senior experts leave, their hard-earned knowledge disappears, resulting in slower troubleshooting, prolonged downtime, and substantial operational costs.

At Aitomatic, we recognized the need for a better approach and believed we could redefine how human expertise was captured and leveraged, turning it into an accessible, living resource that supported engineers in critical moments.

Design principles

I grounded our design approach in one belief: an effective knowledge base should reflect how humans process, connect, and act on information.

My guiding principles were cognitive alignment combined with real-time actionability. The system had to organize knowledge and hierarchically, connect related ideas and live operational data seamlessly, turning static insights into dynamic decision-making help.

Design process

I led a team of 6: two FEs, two BEs, and a junior designer through research, engaging with engineers at companies like PepsiCo, RHI Magnesita, and Tokyo Electron. Through workflow mapping, we identified how engineers documented and utilized their knowledge.

I facilitated design sprints, with the other designer iterating the platform’s interactions. Towards the end, we send designs to engineers for unmoderated testings. My collaboration with engineers and data teams helped us integrate real-time sensor data, transforming the knowledge base into a dynamic, actionable decision engine.

Knowledge Hierarchy

Knowledge Hierarchy

Knowledge Hierarchy

Entities

Entities

Entities

What we build

Knowledge Base within aiKO was a new approach to capturing and using expert knowledge. Instead of forcing engineers into predefined boxes, the system mirrored their mental models, organizing knowledge hierarchically and linking related entities clearly.

Each insight, whether about a specific machine, troubleshooting steps, or best practices, was intuitively connected, making retrieval fast and frictionless.

With live data integration, troubleshooting guides pulled real-time sensor readings, displaying exactly what was happening on the factory floor at that precise moment so engineers could make quick decisions, and solve problems before they escalated.

Impact

The Knowledge Base within aiKO had been projected to reduce troubleshooting times by 25%, cutting both downtime and operational risks. It’s also expected to accelerate onboarding of new engineers by up to 40%, enabling operational effectiveness.

Customers such as Pepsi, RHI Magnesita, and Tokyo Electron recognized the knowledge base’s potential, had been integrating it into their processes.

Knowledge Base

Knowledge Base

Knowledge Base

Linked Agents

Linked Agents

Linked Agents

Add Knowledge

Add Knowledge

Add Knowledge

Add Plan

Add Plan

Add Plan

Add Files

Add Files

Add Files

Citations in Chat

Citations in Chat

Citations in Chat

Save Memory

Save Memory

Save Memory

Manage Memory

Manage Memory

Manage Memory

Connect to Web

Connect to Web

Connect to Web

Connect to API

Connect to API

Connect to API

Links in Chat

Links in Chat

Links in Chat