“We’re building the world’s most trusted digital teammate - the AI avatar who sells, qualifies, and delights as only the best founder could, available to every business, 24/7.”
Find my leads...
Talk
Find my 5 free leads
Fix details of ICP
Here's the Ideal Customer Profile (ICP) generated from your site. Shall I find 5 qualified leads now, or would you like to refine the ICP before I proceed?
Powered by AI
Ideal Customer Profile
Industry
Real Estate – Residential Property Management
Company Size
11–100 employees
Role
Property Manager, Regional Manager
Primary Goal
Increase Net Operating Income (NOI) by reducing vacancy costs
Target Profile Defined
CasaYa
Overview
About the project
Startup
MyWorkerAi
Role
UX Designer + Researcher
Target Users
Founders, Startup Leaders, Sales/Solutions Teams
Duration
10 Weeks
Team
6 Members
Tools
Figma, Figjam, Google Suite, DoveTails
Background
MyWorker.AI is a platform helping startups and small businesses to automate outreach processes, from researching prospects to booking meetings. Generates predictable pipelines with 24/7 autonomous prospecting.
PROBLEM - SOLUTION OVERVIEW
The challenge lies in creating a digital human that communicates warmth, intelligence, charisma, and trust—without crossing into the uncanny valley.
Example text here
Solution
Why MyWorker.AI is better than existing chat bots?
1
Design and avatar that feels like a founder, not a bot
Create depth of immersion by making the AI feel like a natural part of the user’s experience
AI should feel like it’s setting the stage for the founder to shine, not pretending to be the founder
Avatars and animations risk uncanny valley when too “human-like” therefore we focused on the avatar’s awareness, transparency, and timing over mimicking human traits.
2
Transparency
Visually communicating when the AI is working/thinking
Users told us they don’t want overly human-like AI - they want to clearly understand what it’s doing, so instead of a static loading spinner, we created an immersive loading state.
AI walks user through each step: analyzing their site, identifying the ICP, and preparing leads.
3
Progressive Disclosure
Research showed full-screen avatars felt intrusive and lead to user drop-off
To build trust from the very first interaction, we placed a minimized avatar near the header - always present, but never overwhelming, and expandable when the user is ready
Founder-level conversations that feel genuine
Maximize Founder Efficiency by freeing up founders’ time from repetitive sales.
Opportunity to scale sales operation without expensive sales staff
Design Approach
How did we transform user needs into an effective business solution?
This holistic approach: discovery -> define -> ideate -> design -> iterate ensured that our solution was not only innovative, but also deeply rooted in genuine user needs.
UNDERSTANDING AI
The landscape of AI video avatars was very new for us to navigate. To design with confidence, we knew we had to ground every decision in deep research.
We initiated our research process individually and collectively by diving into competitors, research papers, and interviewing startup founders to broaden our knowledge on human-AI interaction, sentiment and Psychological & Ethical Implications of Human-like AI.
Insights on AI
Full disclosure of how the AI works/operates helps with user trust
Users want AI to feel human-aware, not human-like.
Users want human empathy at the start, efficiency later.
Trust in AI increases with transparency, not sophistication
Interviewed 7 early-stage startup founders
We collaboratively structured interview questions to conduct thorough interview with 7 participants from founders to people in the sales industry to better understand founders’ perspectives and how they handle attracting leads to help us build our AI product.
What users said during the interviews...
Founders want AI to handle mechanics to amplify their authenticity, not replace their personal touch.
“Warmth up front, speed later”
“There’s a strong feeling that these tools fail to build a human connection”
“Don’t pretend it’s a person - just make it socially intelligent”
Based on our findings
Sales motivation shifts from emotional to utilitarian throughout the journey, and users want AI to feel human-aware, not human-like.
LITERATURE REVIEW
58% of people advocate for full transparency about AI use
We dove into 8 research papers focused on human-computer interaction within customer service and the anthropomorphization of AI. This was done to grasp an understanding of human sentiment towards the current use of AI within a sales field.
Insights
Full disclosure of how the AI works/operates helps with user trust
58% of people in the study advocate for full transparency about AI use
Users want human empathy at the start, efficiency later
Trust in AI increases with transparency, not sophistication
AI is most effective as a collaborative tool
Anthropomorphizeation (attribution of human characteristics, emotions, or behaviors to non-human entities) can enhance trust
COMPETITIVE ANALYSIS + THE GAP
Competitors fell short in combining complex contextual conversations with lifelike avatars for empathetic sales.
In this phase, we wanted to explore the existing implementations and limitations of current AI Large Language Models (LLM). From our previous research, we identified....
Insights
Leader
11X.AI is an AI-powered outbound SDR named Alice that autonomously prospect, messages, follow-ups, and books meetings via email, Linkedin, and calls. Framed as a “hire-able digital rep” that runs 24/7.
Contenders
ArtisanAI and Tavus.io primarily emphasize the use of AI to improve a customer’s experience from a chat-bot interaction to an AI one, notably through teaching machines to be more human.
Competitive Analysis
THEMATIC ANALYSIS
Grouped and connected the insights.
After collecting data from our user interviews, secondary research and competitive. analysis, we constructed an affinity diagram. This allowed our team to collaboratively discuss and reflect on the findings we had each uncovered, fostering a unified comprehension of the prevalant issues and user requirements.
We grouped repeated data and insights into 7 key themes that were essential to emphasize:
New Leads
Preference
Appearance
Dislikes
Willingness
Challenges
Use of AI
Affinity Map
Pain Points
Follow ups are exhausting founders despite being critical for sales success. Increased outreach creates exponential follow up pressure that founders struggle to manage.
Users want a natural conversation. Replicating the feeling of talking to another human, still knowing its AI.
Emphasis: Founders want AI to help amplify their authenticity, not replace their personal touch.
PROBLEM STATEMENT
“How might we design a transparent AI solution that earns radical trust quickly by blending empathy and human-aware intelligence to nurture, engage, and convert curious visitors into qualified leads?”
BRAINSTORMING
What creative directions we explored to solve the problem?
3
brainstorming
sessions
20+
design
iterations
4
concepts for
solution
Examining the viability of the proposed ideas via
Concept Validation & Founder feedback
Refinement based on:
Feedback, time, and
technological feasibility
SOLUTION
Immersive & Transparent AI Avatar
We focused on honing three key themes: Immersion, Transparency & Progressive Disclosure.
JOURNEY MAPPING + STORYBOARDING
Visualizing the experience of the user
For our Product Requirements Brief (PRD) we began with six top use cases. As the project continued, we iterated and scaled them down to four based on feasibility and time-constraint as we focused on building around our Founder/CEO persona.
Understanding the functionality of our top use cases
Top Use Cases (PRD)
Digital Concierge: Instantly greets, qualifies, and routes visitors with a founder-level touch
Guided Product Discovery: Interactive, personalized video walkthroughs tailored to prospect interest
Qualification & Booking: Smart branching flows to understand need, handle objections, and schedule meetings/demos
FAQ & Objection Handling: Natural-Language Q&A
Re-engagement: Proactively reaches out to scale signups,
Onboarding Guide: Walks new users through their first steps and solves issues before they become churn
Outlining the Minimum Viable Product (MVP)
Digital Concierge
Guided Product Discovery
Qualification & Booking
AI-Powered Follow-Up
Tackling the use cases with storyboarding
WIREFRAMING
Iterated user flow based on founder feedback
Before we could dive into solutions, we took the time to review our understanding of the current user flow MyWorker.AI has, and found the process to be a bit too tedious, and discovered the turnaround time to receive updates via email is around 2 weeks.
Current MyWorker.AI flow for finding leads
Pain Point 1
New campaign flow is redundant, un-engaging and long process
Pain Point 2
Unlock 5 leads for free is easy to miss as it’s inside the third “product” dropdown menu. User would have to click on product section to know feature exists
Pain Point 3
Process to generate 5 leads takes time, and users have no alternative action suggested while waiting. Also does not tell users avg wait time
Understanding the Flow of the User
Initial user flow designs were separated into 3 simple flows/actions a user would take to get us started.
Separate Flows
Example text here
Refined User Flow for Client-Side
As we dove deeper into the ideation phase, we pivoted into two separate MVP states (one for the MyWorker.AI landing page for founders to see the Avatar in action and one for ‘future states’ where we display the AI Avatar in action within a customer’s landing page).
User Flow for MyWorker.AI Path
Separate Flows
Separating the project into two user flows allowed us to not only showcase what the Avatar could do for client’s/founders to meet the business goals of the project, but aso incorporate all the use cases we introduced and also display how the Avatar interacts with a regular user on a client’s landing page.
Low-Mid Fidelity Wireframes
USABILITY EVALUATION
In what ways can it be improved?
We conducted think aloud sessions with 6 founders and sales field agents, each tasked with generating new leads.
Insights
1
“I would never read all the text; just get to the point.”
Users felt there was too much copy/text
Long text would take the AI 20-30 seconds to finish, “what would happen if I click in the middle?”
Some users preferred speaking to the AI entirely, so they are able to multi-task rather than have the chat experience
Design decisions:
Condense the amount of words on output messages while still keeping the experience/journey clear and seamless
Keeping it to 2-3 lines maximum to make it more immersive/interactive
2
“I thought it was just another customer service bot - I didn’t realize it could generate eads.”
Initial sight of Avatar is often glanced over or missed, but users still understood its purpose on the website
Majority of users did not feel the need to click on avatar, nor was it the first thing they noticed/action they took
Design decisions:
Slightly increased the size of the Avatar making it more prominent making sure it’s one of the first things that users see
Added in message banner to clearly state the purpose of the AI Avatar and its capabilities
3
“I loved seeing the email previews - it makes the process feel real and actionable.”
Users did not want to give their email until they saw that the tool could do
Seeing leads and email drafts early in the process builds trust and demonstrates the AI’s value
Participants saw value in connecting their email accounts for email/draft refinement after seeing the preview drafts
ITERATING
What did we change?
After we analyzed all the feedback received from the usability testing, we iterated and came up with a few more solutions.
Included a built in demo/meeting booking system to streamline appointments with the founder for any questions left unanswered or for a more personal-human conversation.
Instead of a static loading spinner, we created an immersive loading state. The AI walks the user through each step—analyzing their site, identifying the ICP, and preparing leads. This visible process reduces uncertainty, builds comfort, and increases completion and conversion rates.
We used Google Veo for generating how the AI will look and feel. With detailed prompts, we designed exactly what we want— for example here an avatar introducing herself that feels present but not distracting, listens without staring, and communicates progress clearly, all while feeling natural. This attention to micro-states is what makes the avatar trustworthy and comfortable for users.
FEASIBILITY
A feasible MVP after meeting with the co-founder
Once the project was polished, we met with the co-founder of MyWorker.AI who is also an engineer. The biggest takeaway was the realization that the technology may not be there yet, or not cost-effective. Therefore, the team pivoted once more and we came up with a design without the AI Avatar for MVP if the company decided to take a simpler route.
Landing Page & 50% version of chat
Enter a company website...
Powered by AI
I can generate leads by analyzing a website. Please enter a company website
Unlock 5 free leads
My Learning & Takeaway
Reflection
How has this experience contributed to my growth
TBD
TBD
TBD
TBD
TBD
TBD
LET’S COOKUP SOME
MAGIC TOGETHER
get in touch
work
about
resume