The Project
In 2017, as AI was beginning to make its way into large enterprises, the need to explore and adopt intelligent technologies was becoming increasingly critical. Organizations were seeking ways to embed AI into daily operations—improving efficiency, enabling omnichannel services, and reshaping how communication itself was evolving with the participation of AI in day-to-day interactions.
It was in this context that we designed AiDA (Artificial Intelligence Digital Assistant), Avanade’s first AI-powered assistant, built on Microsoft’s LUIS interface. More than a chatbot, AiDA represented the company’s entry into conversational AI, bringing together advanced natural language processing with user experience techniques that made the transformation accessible, intuitive, and scalable across the entire enterprise.
Team & Collaboration
Client: Avanade
Timeline: 8 months
Tools: Sketch, Lucidchat, Microsoft PowerPoint and Excel
Role: UX Designer Lead; I was responsible for coordinating design and development efforts across the teams.
Creative Team:
Design Lead (myself) & 2 UX/UI Designers
Development Team:
Based in Seattle, USA, Poland, and India
Business:
1 Product Owner based in San Diego
The Need
Avanade recognized that employees were losing time and energy navigating more than a dozen disconnected systems. What should have been simple tasks often meant switching tools, duplicating steps, or searching across fragmented platforms. At the same time, the organization wanted to demonstrate its ability to harness AI for meaningful business impact.
As UX Design Lead, I guided the effort to ensure this wasn’t treated as just a technical experiment. By elevating the role of the experience design team, we reframed the challenge as a human-centered transformation. The vision was not simply to build a chatbot, but to design AiDA as an assistant that could streamline workflows, simplify access to information, and ultimately improve both efficiency and employee satisfaction—while positioning the company as a leader in intelligent solutions.
Process
1. Research & Discovery
In the early stages, I worked closely with the Product Owner to clarify business goals and define the project scope. I led the effort to consolidate fragmented service information and uncover where the assistant could provide the most value.
-
Mapped the 12+ existing services and internal systems
-
Identified knowledge gaps and prioritized what required validation
-
Conducted 30+ employee interviews across departments to capture daily friction points
-
Combined analytics and survey data into a task matrix, ranking opportunities by urgency, service availability, and integration feasibility.
2. Conversational Design
As the team expanded, I brought in two UX/UI designers under my leadership and guided them in applying user-centered design methods to conversational AI. I facilitated a Design Thinking workshop to co-create AiDA’s persona, tone, and behavioral traits, aligning the assistant’s character with employee expectations and brand values.
-
Designed conversation flows in Lucidchart, mapping user intents to backend services for developer clarity
-
Authored modular conversational scripts tailored by scenario, ensuring consistent tone and clarity
-
Divided script work by complexity, using the process as a mentorship opportunity for junior designers
-
Focused on creating interactions that felt intuitive, supportive, and emotionally intelligent
3. Emotional Intelligence & AI Integration
We used Microsoft LUIS (Language Understanding Intelligent Service) to infuse the assistant with emotional awareness. We outlined use cases where user sentiment detection would guide tonal shifts and escalation logic—helping the assistant respond more empathetically and efficiently.
4. Cross-Platform Visual Design
Because AiDA needed to operate across web chat, Microsoft Teams, a Windows desktop widget, and embedded sales tools, I directed the creation of a modular UI framework. Working with the visual designer, we tested and refined the system to:
-
Reflect company branding
-
Align with platform-specific design conventions
-
Maintain consistent hierarchy, usability, and clarity across all channels
5. Reflexion
Although the project budget ended before we could evolve AiDA into a more advanced conversational assistant like the ones we see today, it was a pivotal first step. The work proved how AI could reshape communication and access inside large organizations, laying the groundwork for future enterprise adoption and showing the transformative potential of user-centered AI.
And on a personal note—I had a lot of fun figuring out how to bring a bot to life.
Project Success
The outcome was a smart, scalable assistant that continues to support more than 400,000 users across Avanade and Accenture. Beyond its immediate benefits, AiDA became a reusable blueprint for enterprise AI solutions, demonstrating how conversational interfaces can deliver both measurable business value and a more intuitive, human experience.
Articles & Presentations
After completing the project, I was eager to share my knowledge, especially in the emerging field of AI-driven conversational interfaces. I wrote a series of articles and gave public presentations to contribute to the tech community.
Articles
-
Can Your Bot Provide True Business Value to Your Company?
-
8 Critical Elements for Your Brand's Next Chatbot
Conferences & Talks
-
Conversational Interfaces with Microsoft Azure – Mississauga, Canada
-
UX in Conversational Interfaces – Avanade Internal Conference
Leadership in Action
This was more than a design project—it was a team evolution journey:
-
Our team was new to conversational AI, so I created a space for exploration and shared learning. We embraced iterative prototyping, with daily check-ins to align on progress, share insights, and maintain momentum.
-
Despite being spread across three continents, we worked as one team with a shared direction. Everyone contributed their expertise, and I ensured open communication to keep us aligned and supported throughout the process.
-
Conflicting regional priorities and unclear internal documentation were challenges we faced. To address this, I led workshops and decision-making sessions, guiding the team toward consensus on scalable, global use cases that worked for all regions.