Shireen

Home

Work

About

Resume

Get in touch

Microsoft Fabrics x Agent Lucid AI

PRODUCT DESIGN

WEB APP

AI

Role

UX research, Workflow design,
UI Design, User testing

Impact

75% less setup time,

80% increase in user confidence,

and 60% increased data insights

Team

1 Designer, 1 Project Manager,
4 Developers

Overview

What is Lucid Data Hub & AI Agents?

Lucid Data Hub unifies fragmented enterprise data and makes it accessible through AI-powered automation. Its built-in AI Agents analyze data, answer questions, and trigger actions using natural language reducing manual work and enabling faster, smarter decisions.

Why Integrate with Microsoft Fabric?

By embedding Lucid Data Hub and AI Agents into Microsoft Fabric, businesses unlock the power of real-time, secure, and scalable analytics within the tools they already use. Integration with Fabric’s data lakes, warehouses, and Power BI eliminates data silos and removes the need for extra ETL processes. This unified environment streamlines workflows, accelerates insights, and enables teams to move from data to decisions with unprecedented speed and efficiency.

Problem Context

Traditional sales and data operations face several challenges that hinder efficiency and data insights.

• Siloed Data: Enterprise info is fragmented across tools limiting insight.

• Manual Sales Processes: Data prep and outreach require time-consuming human effort.

• Usability Gaps: Most data/AI tools are hard for non-technical users.

Integrated Solution Vision

Embedding AgentLucid into Microsoft Fabric eliminates data silos, streamlines setup, and gives every authorized user instant access to actionable insights and AI-powered automation.

Outcome

Before

After

User

Satisfaction

Setup Time

User

Confidence

Time to

Insights

• User Statisfaction increased significantly after integrating AI agents and guided workflows.


• Setup Time reduced by 75%, enabling faster onboarding and agent deployment.


• User Confidence boosted by 80%, with non-technical users finding the interface intuitive and transparent.


• Time to Insights improved by 60%, allowing quicker decision-making compared to manual workflows.

Key UX Challenges

Problem:

Lucid’s original experience was unified and tightly integrated. But Microsoft Fabric’s architecture required it to be broken.


Solution:

Lucid's experience is broken into three standalone apps AI Agent, Context Hub, and Chatbot each appearing as a separate item in the Fabric ecosystem.

The flow now always starts with the Agent linking to an existing Context Hub automatically or guiding users to create one inline keeping the Agent central and the Hub a seamless prerequisite.This also supported with :

  • Shared UI

  • Deep linking between tools

  • Modular deployment compatibility with Fabric

3 Core Features

AI Agents - Lets you create smart helpers to watch data and do work automatically

Context Hub - Scan to Understand and explain your company's data.

Lucid Chatbot - Data Assistant that answers questions in plain language

AI Agent Builder : Creating your helpers

A no-code guided chat interface that enables users to create intelligent agents that monitor data, evaluate conditions, and trigger automated workflows like alerts, emails, and report generation.

Why it's Important

It empowers non-technical users like operations managers, analysts, or admins to independently automate workflows that would otherwise require developer support. The visual structure helps users reason about the logic, while tight integration with the Context Hub ensures the agents are always referencing well-defined, trustworthy data.

How it works

  1. The guided builder breaks the AI Agent creation into structured steps, making a complex process approachable:

    1. Context Hub Selection – Users begin by naming and selecting a Context Hub (industry, region, business context) or create a new one. This ensures the agent has the right foundation for interpreting data.

    2. Data Pool Configuration – Users choose datasets the agent can access and define SQL queries when needed. This creates clarity on what data powers the agent.

    3. Business Logic Setup – Once data is linked, users move to logic configuration, connecting datasets to rules and conditions.

    4. Actions & Suggested Next Steps – Finally, users define actions (e.g., alerts, reports, dashboards).

Iterations

• The first exploration used a freeform canvas where users drag and connect logic blocks. While powerful, it risked being overwhelming new users might not know where to start or how to structure an agent correctly.

• To address this, a second iteration introduced a guided interface. Instead of an open canvas, the flow now provides clear steps selecting a Context Hub, adding data pools, configuring business logic, and defining actions.

Why this direction felt stronger

• Establishes a clear starting point and structured flow.

• Reduces confusion around prerequisites like the Context Hub.

• Lowers the learning curve while still allowing modular flexibility later.

Context Hub : Setting the stage for Data Use

The Context Hub serves as the foundational intelligence layer of the platform. It connects to enterprise data sources and builds a data dictionary that describes the meaning of each table and field in human-understandable terms.

Why it's Important

This ensures that every AI action, automation, or insight is grounded in shared, accurate understanding of the data. It avoids misinterpretation by downstream tools like agents and chatbots, and removes ambiguity for both technical and non-technical users interacting with business-critical data.

How it works

  1. Users give Context Hub a name, industry, and region to tailor how data is interpreted.

  1. Users begin by connecting a secure data workspace (like a lakehouse).

  1. The system performs an AI-driven scan of the data structure identifying columns, field types, and naming conventions.

  1. Once connected to a data source, the system scans and extracts fields; AI then suggests definitions through a guided review interface to build the data dictionary.

Chatbot : Your Data Assistant

An AI-powered chatbot that allows users to ask natural language questions or issue commands and get live answers from their data without navigating dashboards or writing queries.

Why it's Important

It democratizes access to insights by removing the technical barrier of BI tools or SQL. Users can self-serve information or activate automations via a simple chat interface, improving productivity and reducing dependency on technical teams.

How it works

  1. Users begin by selecting an existing or create a new Context Hub.

  1. Chat Interface to ask natural language questions or issue commands.

Case Example

Smart Inventory Management at a Retail Store

A retail store that stocks hundreds of products. Traditionally, the store manager spends hours checking inventory levels, sales reports, and manually contacting suppliers to reorder stock. This can cause delays or stock shortages.How AgentLucid Helps :


1. Context Hub:

The store’s data is stored in their Microsoft Fabric workspace and lakehouse. The manager sets this up in Context Hub by explaining that this data covers “Product Inventory” and “Sales Records.” The system scans all the tables and builds a detailed dictionary explaining what each column means, such as “Product ID,” “Quantity in Stock,” and “Reorder Threshold.”


2. AI Agent Builder:

Using the drag-and-drop tool, the manager builds a smart helper: “Check inventory levels daily, and if any product is below the reorder threshold, send a notification to the purchasing department to reorder.” This agent works automatically without constant supervision.


3. Chatbot:

At any time, the store manager can ask the chatbot inside Microsoft Fabric: “Which products are running low today?” The chatbot replies with a clear list, including notes on which orders are already placed, thanks to the AI agent’s work.

Outcome

The store avoids running out of popular products, reduces manual workload, and speeds up its restocking process all with simple tools anyone can use.

Testing & Validation


  • Prototype Walkthroughs Simulated user flows were tested with stakeholders to estimate task completion effort and confidence levels.

  • Comparative Benchmark - The freeform canvas approach was compared with the guided setup to project differences in setup time and error likelihood.

  • Heuristic Evaluation - Applied usability heuristics to identify friction points and score improvements in satisfaction and learnability.

  • From these exercises, we estimated potential improvements in setup time (~75% faster), user confidence (~80% higher), and time-to-insight (~60% quicker).

Create a free website with Framer, the website builder loved by startups, designers and agencies.