Voice Assist is an AI-powered call-handling assistant for high-volume, service-based businesses (plumbing, legal services, senior care). It screens callers, answers FAQs, and captures information so staff can focus on urgent or qualified leads.
CATEGORY
AI / FM
ROLE
Senior Product Designer
User Researcher
UI Designer
YEAR
2025
TIMELINE
January — Early interviews with target customers
April — Beta program launched
May — Conduct user testing with beta group
July — Full GTM launch
Why it matters
Many customers were overwhelmed by missed calls, unqualified leads, and staff burnout. Rising call center costs and poor experiences with offshore answering services also made them skeptical of AI. This created an opportunity to design a trustworthy, customizable assistant that could handle routine calls reliably while reflecting the tone and priorities of a real business.
Success metrics
“We reduced our missed calls by 44% using Voice Assist. That’s a huge win for our team and our clients.”
Voice Assist Customer – Property Management
Voice Assist launched on July 17, 2025.
291 trials in 2 weeks
44% drop in missed calls (Property Mgmt)
118% increase in answered calls (Home Services)
Beta launch
Before launching our beta, our team conducted discovery interviews. Real issues surfaced across interviews with customers in home services, senior care, and legal services:
Missed calls during overflow and after-hours
High volume of unqualified leads
Staff burnout from repetitive intake and FAQs
Poor experiences with third-party answering services
These insights shaped our early use cases for Voice Assist:
Lead capture — collect caller info and preferred appointment times during off-hours
Answer FAQs — handle general business questions when staff are unavailable
Lead qualification — screen calls with custom questions to filter out bad leads
Beta goals
Validate product market fit and gather real world feedback from early adopters. We focused on high call volume industries to evaluate how the assistant performed in real scenarios: where trust broke down, what users needed to feel in control, and which moments caused friction.
What I designed and why
I focused on creating a trustworthy, flexible assistant experience aligned with early customer needs and industry workflows:
Feature | Why it mattered |
|---|---|
Business Profile | Gave AI the context it needed to sound relevant and helpful (e.g., services, hours, FAQs). |
Caller Intake | Let users decide what info to collect so conversations felt purposeful, not generic. |
Voice and tone | Helped businesses sound on-brand and more human, which increased trust and adoption. |
Text follow-up | Re-engaged leads after the call and nudged them toward booking, boosting conversion. |
Beta process
Phase | What we did |
|---|---|
Beta interview | Conducted interviews with 7 high-volume businesses across senior care, legal, and home services |
Beta testing | Usability sessions and pilot testing with agency partners and SMBs |
Synthesis & prioritization | Grouped insights into pre-MVP must-haves and post-MVP enhancements |
Beta insights
Top 5 Insights from Beta
Insight | Why It Mattered |
|---|---|
Business Profile felt more useful over time but had unclear field hierarchy | Users weren’t sure which fields were required or relevant to AI |
Voice and tone options were appreciated, with interest in having more control | Users wanted more influence over pacing and retry behavior |
Users wanted a way to verify the assistant’s behavior before routing real calls | Users wanted confirmation that the AI understood their business info and was set up correctly before going live |
Users didn’t know when Voice Assist was “live” or fully configured | Unclear milestones left users unsure if setup was complete, which could delay launch or leave Voice Assist inactive |
Users wanted reassurance that callers could reach a human when needed | Users feared losing valuable leads if someone resisted AI or had an urgent issue |
MVP launch
The MVP's goal was to build on beta learnings to reduce onboarding friction, increase adoption, and confidently scale Voice Assist.
What we refined from the Beta launch
5-Step Onboarding Guide – Clarified when setup was complete and VA was ready to go live
Business Profile → Smart Profile – Separated AI-specific fields to improve clarity and scalability
Voice & Tone Settings – Expanded options like pacing, retry logic, and multilingual support while balancing customization with AI constraints to keep the system reliable
What we added post Beta launch
Feature | What it solved |
|---|---|
Call Flow Templates | Prebuilt starting points in call flows for common VA use cases, making activation faster |
Test Call Preview UI | Gave users a way to safely preview how the assistant would behave before routing real calls, addressing a major trust gap identified during beta. |
Micro-Surveys in App | Captured feedback without waiting for support tickets |
Human Transfer Fallback | Ensured AI never blocked a lead by offering human escalation |
Lessons
Build trust in AI by ensuring users feel confident in the help it provides and understand how their information is being used.
Pair smart defaults with targeted customization by leveraging existing customer context, making interactions feel intelligent and personalized.
Smooth Scroll
This will hide itself!
This will hide itself!
Turning missed calls into conversations: a 0→1 AI assistant
CONTEXT
Voice Assist is an AI-powered call-handling assistant for high-volume, service-based businesses (plumbing, legal services, senior care). It screens callers, answers FAQs, and captures information so staff can focus on urgent or qualified leads.
CATEGORY
AI / FM
ROLE
Senior Product Designer
User Researcher
UI Designer
YEAR
2025
TIMELINE
January — Early interviews with target customers
April — Beta program launched
May — Conduct user testing with beta group
July — Full GTM launch
Why it matters
Many customers were overwhelmed by missed calls, unqualified leads, and staff burnout. Rising call center costs and poor experiences with offshore answering services also made them skeptical of AI. This created an opportunity to design a trustworthy, customizable assistant that could handle routine calls reliably while reflecting the tone and priorities of a real business.
Success metrics
“We reduced our missed calls by 44% using Voice Assist. That’s a huge win for our team and our clients.”
Voice Assist Customer – Property Management
Voice Assist launched on July 17, 2025.
291 trials in 2 weeks
44% drop in missed calls (Property Mgmt)
118% increase in answered calls (Home Services)
Beta launch
Before launching our beta, our team conducted discovery interviews. Real issues surfaced across interviews with customers in home services, senior care, and legal services:
Missed calls during overflow and after-hours
High volume of unqualified leads
Staff burnout from repetitive intake and FAQs
Poor experiences with third-party answering services
These insights shaped our early use cases for Voice Assist:
Lead capture — collect caller info and preferred appointment times during off-hours
Answer FAQs — handle general business questions when staff are unavailable
Lead qualification — screen calls with custom questions to filter out bad leads
Beta goals
Validate product market fit and gather real world feedback from early adopters. We focused on high call volume industries to evaluate how the assistant performed in real scenarios: where trust broke down, what users needed to feel in control, and which moments caused friction.
What I designed and why
I focused on creating a trustworthy, flexible assistant experience aligned with early customer needs and industry workflows:
Feature | Why it mattered |
|---|---|
Business Profile | Gave AI the context it needed to sound relevant and helpful (e.g., services, hours, FAQs). |
Caller Intake | Let users decide what info to collect so conversations felt purposeful, not generic. |
Voice and tone | Helped businesses sound on-brand and more human, which increased trust and adoption. |
Text follow-up | Re-engaged leads after the call and nudged them toward booking, boosting conversion. |
Beta process
Phase | What we did |
|---|---|
Beta interview | Conducted interviews with 7 high-volume businesses across senior care, legal, and home services |
Beta testing | Usability sessions and pilot testing with agency partners and SMBs |
Synthesis & prioritization | Grouped insights into pre-MVP must-haves and post-MVP enhancements |
Beta insights
Top 5 Insights from Beta
Insight | Why It Mattered |
|---|---|
Business Profile felt more useful over time but had unclear field hierarchy | Users weren’t sure which fields were required or relevant to AI |
Voice and tone options were appreciated, with interest in having more control | Users wanted more influence over pacing and retry behavior |
Users wanted a way to verify the assistant’s behavior before routing real calls | Users wanted confirmation that the AI understood their business info and was set up correctly before going live |
Users didn’t know when Voice Assist was “live” or fully configured | Unclear milestones left users unsure if setup was complete, which could delay launch or leave Voice Assist inactive |
Users wanted reassurance that callers could reach a human when needed | Users feared losing valuable leads if someone resisted AI or had an urgent issue |
MVP launch
The MVP's goal was to build on beta learnings to reduce onboarding friction, increase adoption, and confidently scale Voice Assist.
What we refined from the Beta launch
5-Step Onboarding Guide – Clarified when setup was complete and VA was ready to go live
Business Profile → Smart Profile – Separated AI-specific fields to improve clarity and scalability
Voice & Tone Settings – Expanded options like pacing, retry logic, and multilingual support while balancing customization with AI constraints to keep the system reliable
What we added post Beta launch
Feature | What it solved |
|---|---|
Call Flow Templates | Prebuilt starting points in call flows for common VA use cases, making activation faster |
Test Call Preview UI | Gave users a way to safely preview how the assistant would behave before routing real calls, addressing a major trust gap identified during beta. |
Micro-Surveys in App | Captured feedback without waiting for support tickets |
Human Transfer Fallback | Ensured AI never blocked a lead by offering human escalation |
Lessons
Build trust in AI by ensuring users feel confident in the help it provides and understand how their information is being used.
Pair smart defaults with targeted customization by leveraging existing customer context, making interactions feel intelligent and personalized.
Smooth Scroll
This will hide itself!
This will hide itself!
Turning missed calls into conversations: a 0→1 AI assistant
CONTEXT
Voice Assist is an AI-powered call-handling assistant for high-volume, service-based businesses (plumbing, legal services, senior care). It screens callers, answers FAQs, and captures information so staff can focus on urgent or qualified leads.
CATEGORY
AI / FM
ROLE
Senior Product Designer
User Researcher
UI Designer
YEAR
2025
TIMELINE
January — Early interviews with target customers
April — Beta program launched
May — Conduct user testing with beta group
July — Full GTM launch
Why it matters
Many customers were overwhelmed by missed calls, unqualified leads, and staff burnout. Rising call center costs and poor experiences with offshore answering services also made them skeptical of AI. This created an opportunity to design a trustworthy, customizable assistant that could handle routine calls reliably while reflecting the tone and priorities of a real business.
Success metrics
“We reduced our missed calls by 44% using Voice Assist. That’s a huge win for our team and our clients.”
Voice Assist Customer – Property Management
Voice Assist launched on July 17, 2025.
291 trials in 2 weeks
44% drop in missed calls (Property Mgmt)
118% increase in answered calls (Home Services)
Beta launch
Before launching our beta, our team conducted discovery interviews. Real issues surfaced across interviews with customers in home services, senior care, and legal services:
Missed calls during overflow and after-hours
High volume of unqualified leads
Staff burnout from repetitive intake and FAQs
Poor experiences with third-party answering services
These insights shaped our early use cases for Voice Assist:
Lead capture — collect caller info and preferred appointment times during off-hours
Answer FAQs — handle general business questions when staff are unavailable
Lead qualification — screen calls with custom questions to filter out bad leads
Beta goals
Validate product market fit and gather real world feedback from early adopters. We focused on high call volume industries to evaluate how the assistant performed in real scenarios: where trust broke down, what users needed to feel in control, and which moments caused friction.
What I designed and why
I focused on creating a trustworthy, flexible assistant experience aligned with early customer needs and industry workflows:
Feature | Why it mattered |
|---|---|
Business Profile | Gave AI the context it needed to sound relevant and helpful (e.g., services, hours, FAQs). |
Caller Intake | Let users decide what info to collect so conversations felt purposeful, not generic. |
Voice and tone | Helped businesses sound on-brand and more human, which increased trust and adoption. |
Text follow-up | Re-engaged leads after the call and nudged them toward booking, boosting conversion. |
Beta process
Phase | What we did |
|---|---|
Beta interview | Conducted interviews with 7 high-volume businesses across senior care, legal, and home services |
Beta testing | Usability sessions and pilot testing with agency partners and SMBs |
Synthesis & prioritization | Grouped insights into pre-MVP must-haves and post-MVP enhancements |
Beta insights
Top 5 Insights from Beta
Insight | Why It Mattered |
|---|---|
Business Profile felt more useful over time but had unclear field hierarchy | Users weren’t sure which fields were required or relevant to AI |
Voice and tone options were appreciated, with interest in having more control | Users wanted more influence over pacing and retry behavior |
Users wanted a way to verify the assistant’s behavior before routing real calls | Users wanted confirmation that the AI understood their business info and was set up correctly before going live |
Users didn’t know when Voice Assist was “live” or fully configured | Unclear milestones left users unsure if setup was complete, which could delay launch or leave Voice Assist inactive |
Users wanted reassurance that callers could reach a human when needed | Users feared losing valuable leads if someone resisted AI or had an urgent issue |
MVP launch
The MVP's goal was to build on beta learnings to reduce onboarding friction, increase adoption, and confidently scale Voice Assist.
What we refined from the Beta launch
5-Step Onboarding Guide – Clarified when setup was complete and VA was ready to go live
Business Profile → Smart Profile – Separated AI-specific fields to improve clarity and scalability
Voice & Tone Settings – Expanded options like pacing, retry logic, and multilingual support while balancing customization with AI constraints to keep the system reliable
What we added post Beta launch
Feature | What it solved |
|---|---|
Call Flow Templates | Prebuilt starting points in call flows for common VA use cases, making activation faster |
Test Call Preview UI | Gave users a way to safely preview how the assistant would behave before routing real calls, addressing a major trust gap identified during beta. |
Micro-Surveys in App | Captured feedback without waiting for support tickets |
Human Transfer Fallback | Ensured AI never blocked a lead by offering human escalation |
Lessons
Build trust in AI by ensuring users feel confident in the help it provides and understand how their information is being used.
Pair smart defaults with targeted customization by leveraging existing customer context, making interactions feel intelligent and personalized.
Smooth Scroll
This will hide itself!
This will hide itself!