Research

The landscape of AI video agents was new… to build with confidence, we grounded every decision in deep research

We combined quantitative market analyses with qualitative founder insights and academic research to understand how AI could empower startups without compromising transparency or trust. This multi-layered approach helped us design a system that's both data-informed and human-aware.

Competitive Analysis

Understanding the AI sales landscape to find our white space

We analyzed 14 competitors spanning AI sales agents, chatbots, and avatar-based assistants from platforms like 11x.ai, Artisan, and Warmly, to conversational tools like Drift, Ada, and DeepBrain AI.

01

Key Finding

Most competitors automate tasks, few address the emotional dimension

02

Market Gap

Most AI sales rep tools drop SMBs into complex setups

03

Feature Opportunity

Adding video avatars could strengthen "all-in-one" platform positioning

White Paper Research

Users support honesty about AI nature

To ensure MyWorker.AI's conversational experience aligned with both user psychology and ethical AI principles, we conducted an in-depth review of eight academic and industry white papers. These papers spanned topics such as human-AI collaboration, embodied conversational agents, anthropomorphism, ethical transparency, and AI-driven customer engagement.

Key Insights

01

AI works best as a collaborator, not a replacement.

02

Users trust AI more when it clearly communicates its role and boundaries.

03

Human-like grounded traits, used thoughtfully in AI avatars, enhance trust.

PESTLE Analysis

Understanding the external forces shaping AI adoption

To design an AI solution grounded in real-world viability, we conducted a comprehensive PESTLE analysis. This framework helped us understand not just user needs, but also the broader market, ethical, and regulatory context shaping AI-led sales tools.

01

Political & Legal

AI regulation and privacy laws are tightening, making transparency and compliance non-negotiable.

02

Economic

AI adoption is accelerating, but SMBs now demand clear ROI and measurable value, not just automation.

03

Social

Users want control and authenticity, trust is built when AI feels human-aware, not manipulative.

04

Technological

Multimodal AI is rising, creating new UX possibilities for blending video, voice, and language models.

User Interviews

Uncovering what founders value most when AI becomes a voice for their business

To understand the pain points, motivations, and mental models of our target users, we conducted 7 in-depth interviews with founders and business owners across e-commerce, real estate, creative, and service industries. Each conversation uncovered the tension between automation and authenticity, revealing what founders value when AI becomes a face of their business.

We wanted to understand

01

How do start-ups handle new leads for their company?

02

What degree of avatar realism matters? Why?

03

What one job will deliver the fastest, clearest business impact?

Insights to Ideas

Challenges that founders were facing

Our ideation process began with a simple goal, staying close to the founders' reality. Across conversations, we looked closely at their pain points at a low-level and a consistent pattern emerged: time was their scarcest resource.

Design Thinking

Translating user pain points into tangible design possibilities

This stage was where our ideas truly came alive. We began with a Crazy 8's session, rapidly sketching diverse interpretations of how an AI avatar could ease founders' daily challenges. We explored designs from how the interface would look and what features it could have. They ranged from having focus on captions for accessibility, chat capabilities and others.

We also did storyboarding to conceptualize what features would solve the main challenges that users faced. These included digital concierge, guided product discovery to show them value, and qualification and booking and features like AI-powered follow-ups.

Storyboarding

Crazy 8's

Designing the User Flow

​Visualizing how the avatar interacts across distinct touchpoints

With our use cases prioritized, the next step was to translate ideas into motion. We approached this collaboratively but with intention starting first with individual exploration.

Each designer took ownership of one use case and independently mapped the user journey end-to-end. This phase of separation was deliberate; it allowed for divergent thinking before alignment.

Once every perspective was on the table, we converged to identify overlaps, refine transitions, and define a cohesive end-to-end flow for MyWorker.AI.

Moodboard

Understanding how conversational AI looks, feels, and earns trust

I made the moodboard to study from chat-based assistants, avatar-led interfaces, and enterprise AI touchpoints each exploring a different tone of voice and visual rhythm.

Wireframing

Giving life to early concepts

We explored the avatar's initial presence on the page? Should it be a subtle bubble or an overlay? We also explored the interaction model we considered a full-screen, video-centric experience for product demo. We also integrated chat interface to leverage familiar chatbot patterns to make this new type of interaction feel more intuitive for first-time users.

We moved quickly to detailed mid-fidelity designs. We kept it detailed because the feel of the AI interaction was just as important as the flow. We needed to test how users would react to a human-like presence.

Lo-Fi Wireframes

Mid-Fi Wireframes

Testing

Initial usability testing - gathering early feedback

With our mid-fidelity prototype ready, we moved into our first round of moderated usability testing with 2 founders.

The goal was to get early feedback on the most critical part of the user journey: the Product Demo workflow. Specifically, we wanted to see how users would react when the avatar proactively offered to give them a video demo of the platform's core features, like lead generation and personalized email crafting, meeting booking. The insights we gathered here were crucial and led to some significant design pivots.

2 Usability Tests with Founders

Focus: The Product Demo Workflow

Testing

Key findings - here our assumptions met reality

The feedback from these tests was incredibly direct and insightful, and it revealed two major flaws in our initial approach. First, the overall experience still felt like a chatbot to users and it lacked the immersive, conversational feel we were aiming for. Second, our users skipped the product demo. Every time. The founders we tested told us plainly that they don't have time for a tour. They wanted to understand the value of the product and how it can help them in an effective way.

Chat Experience

The experience lacked sense of immersion

"This feels like a slightly better chatbot."

Skipped Product Demo

Users skipped the demo

"Don't show me what it does, tell me how it helps me. I don't have time for a tour."

Iteration

The strategic pivot - from demo to do

This critical feedback forced us to completely rethink our approach. The core insight was that users didn't want us to show them the value; they wanted to experience it for themselves. So, we made a major strategic pivot. We killed the passive demo video.

Instead, the avatar now says, 'Let's do it together.' It guides the user through the actual steps of generating leads and crafting emails, live, right there in the interface. This gives them a real, tangible taste of the platform's power, not just a description of it.

To fix the immersion problem, we designed this entire flow to happen inside the avatar's UI, creating a seamless and magical experience. This pivot was the single most important decision we made in the entire project.

Before

Greet User

Passive Demo Video

Book Meetings

After

Greet User

Generate Leads

Craft Emails

Book Meetings

Validation

Testing the new workflow

We ran a second round of moderated usability tests with four founders, asking them to think out loud: Our goal was to answer three critical questions: First, is the new interaction clear? Second, do they perceive the generated leads as valuable? And third, and most importantly, do they feel confident and comfortable sharing information like email with the AI in this new, more intimate format?

01

Clarity

“ Is the initial interaction with the AI clear and intuitive?”

02

Value

“What is their reaction to the quality of the live-generated leads?”

03

Confidence

“Do they feel confused or uncomfortable at any point in the flow?”

Refinements

Iterating based on usability insights

User feedback helped us fine-tune the experience. Based on that here are few design refinements that we made:

Key Insight

Design Decision