StarZero All shows Try the podcast clipper →

Jan 8, 2026 · 35m · mixergy

#2292 AI Automation that makes cold calls

Evgeny Matze · 21m spoken Andrew Warner · 8m spoken Aiden Richards · 1m spoken
0:00 / 0:00

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

In this interview hosted by Andrew Warner, Rezora co-founders Evgeny Matze and Aiden Richards explain how they transformed a high-demand voice AI cold calling agency for real estate professionals into a scalable, fine-tuned SaaS platform.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Andrew holds 29.1% of the talking time here. How this is scored →

Andrew as informed peer 3.7 Guest teaching 4.5 Guest disagreement 0.8 Andrew pushing back 2.7
05100:0010:0020:0030:001:26–3:53 · Andrew as informed peer 2/10 The Cold Calling Grind in Real Estate Andrew makes optimistic assumptions about expired listings having high conversion rates and asks why a 1-2% rate was frustrating. Evgeny educates him on real estate reality, explaining that disgruntled homeowners often reject agents and that 200 dials are needed just to reach one homeowner.3:54–7:43 · Andrew as informed peer 3/10 Developing the First AI Voice Agent Prototype Andrew challenges Evgeny on why he did not simply scale a real estate brokerage empire with his tool instead of pivoting to software. Evgeny explains his CS background and personality mismatch with the rigid, decades-old real estate business model.7:43–13:00 · Andrew as informed peer 4/10 Client Acquisition and Early Low-Code Agency Architecture Andrew digs into the lean customer acquisition funnel using native Meta lead capture forms without a landing page. Evgeny describes qualifying leads via Google Meet and screening out unqualified agents despite high demand.13:00–18:59 · Andrew as informed peer 5/10 Voice Automation Workflows and Model Fine-Tuning Andrew maps out how an agency could build outbound voice agents for other trades like plumbing using Zapier and Vapi. Evgeny explains that true conversational salesmanship requires complex supervised fine-tuning data pipelines and LLM-as-a-judge scoring rather than basic prompt engineering.19:00–23:38 · Andrew as informed peer 3/10 Transition to Self-Serve SaaS and Onboarding Aiden Evgeny discusses transitioning away from client services to build software full-time and finding Aiden via YC Co-founder matching. Andrew brings Aiden in to define operational responsibilities and playfully notes Aiden's persistent outreach.23:39–35:48 · Andrew as informed peer 5/10 SaaS Customization, AI Coding Tools, and Company Naming Andrew presses on why Evgeny shut down paying agency clients instead of maintaining a $50k/mo agency model. Evgeny explains that custom prompt adjustments were unsustainable, and breaks down why AI coding tools still demand deep software architectural knowledge.1:26–3:53 · Guest teaching 6/10 The Cold Calling Grind in Real Estate Andrew makes optimistic assumptions about expired listings having high conversion rates and asks why a 1-2% rate was frustrating. Evgeny educates him on real estate reality, explaining that disgruntled homeowners often reject agents and that 200 dials are needed just to reach one homeowner.3:54–7:43 · Guest teaching 4/10 Developing the First AI Voice Agent Prototype Andrew challenges Evgeny on why he did not simply scale a real estate brokerage empire with his tool instead of pivoting to software. Evgeny explains his CS background and personality mismatch with the rigid, decades-old real estate business model.7:43–13:00 · Guest teaching 3/10 Client Acquisition and Early Low-Code Agency Architecture Andrew digs into the lean customer acquisition funnel using native Meta lead capture forms without a landing page. Evgeny describes qualifying leads via Google Meet and screening out unqualified agents despite high demand.13:00–18:59 · Guest teaching 7/10 Voice Automation Workflows and Model Fine-Tuning Andrew maps out how an agency could build outbound voice agents for other trades like plumbing using Zapier and Vapi. Evgeny explains that true conversational salesmanship requires complex supervised fine-tuning data pipelines and LLM-as-a-judge scoring rather than basic prompt engineering.19:00–23:38 · Guest teaching 2/10 Transition to Self-Serve SaaS and Onboarding Aiden Evgeny discusses transitioning away from client services to build software full-time and finding Aiden via YC Co-founder matching. Andrew brings Aiden in to define operational responsibilities and playfully notes Aiden's persistent outreach.23:39–35:48 · Guest teaching 5/10 SaaS Customization, AI Coding Tools, and Company Naming Andrew presses on why Evgeny shut down paying agency clients instead of maintaining a $50k/mo agency model. Evgeny explains that custom prompt adjustments were unsustainable, and breaks down why AI coding tools still demand deep software architectural knowledge.1:26–3:53 · Guest disagreement 1/10 The Cold Calling Grind in Real Estate Andrew makes optimistic assumptions about expired listings having high conversion rates and asks why a 1-2% rate was frustrating. Evgeny educates him on real estate reality, explaining that disgruntled homeowners often reject agents and that 200 dials are needed just to reach one homeowner.3:54–7:43 · Guest disagreement 1/10 Developing the First AI Voice Agent Prototype Andrew challenges Evgeny on why he did not simply scale a real estate brokerage empire with his tool instead of pivoting to software. Evgeny explains his CS background and personality mismatch with the rigid, decades-old real estate business model.7:43–13:00 · Guest disagreement 0/10 Client Acquisition and Early Low-Code Agency Architecture Andrew digs into the lean customer acquisition funnel using native Meta lead capture forms without a landing page. Evgeny describes qualifying leads via Google Meet and screening out unqualified agents despite high demand.13:00–18:59 · Guest disagreement 1/10 Voice Automation Workflows and Model Fine-Tuning Andrew maps out how an agency could build outbound voice agents for other trades like plumbing using Zapier and Vapi. Evgeny explains that true conversational salesmanship requires complex supervised fine-tuning data pipelines and LLM-as-a-judge scoring rather than basic prompt engineering.19:00–23:38 · Guest disagreement 0/10 Transition to Self-Serve SaaS and Onboarding Aiden Evgeny discusses transitioning away from client services to build software full-time and finding Aiden via YC Co-founder matching. Andrew brings Aiden in to define operational responsibilities and playfully notes Aiden's persistent outreach.23:39–35:48 · Guest disagreement 2/10 SaaS Customization, AI Coding Tools, and Company Naming Andrew presses on why Evgeny shut down paying agency clients instead of maintaining a $50k/mo agency model. Evgeny explains that custom prompt adjustments were unsustainable, and breaks down why AI coding tools still demand deep software architectural knowledge.1:26–3:53 · Andrew pushing back 3/10 The Cold Calling Grind in Real Estate Andrew makes optimistic assumptions about expired listings having high conversion rates and asks why a 1-2% rate was frustrating. Evgeny educates him on real estate reality, explaining that disgruntled homeowners often reject agents and that 200 dials are needed just to reach one homeowner.3:54–7:43 · Andrew pushing back 3/10 Developing the First AI Voice Agent Prototype Andrew challenges Evgeny on why he did not simply scale a real estate brokerage empire with his tool instead of pivoting to software. Evgeny explains his CS background and personality mismatch with the rigid, decades-old real estate business model.7:43–13:00 · Andrew pushing back 2/10 Client Acquisition and Early Low-Code Agency Architecture Andrew digs into the lean customer acquisition funnel using native Meta lead capture forms without a landing page. Evgeny describes qualifying leads via Google Meet and screening out unqualified agents despite high demand.13:00–18:59 · Andrew pushing back 3/10 Voice Automation Workflows and Model Fine-Tuning Andrew maps out how an agency could build outbound voice agents for other trades like plumbing using Zapier and Vapi. Evgeny explains that true conversational salesmanship requires complex supervised fine-tuning data pipelines and LLM-as-a-judge scoring rather than basic prompt engineering.19:00–23:38 · Andrew pushing back 1/10 Transition to Self-Serve SaaS and Onboarding Aiden Evgeny discusses transitioning away from client services to build software full-time and finding Aiden via YC Co-founder matching. Andrew brings Aiden in to define operational responsibilities and playfully notes Aiden's persistent outreach.23:39–35:48 · Andrew pushing back 4/10 SaaS Customization, AI Coding Tools, and Company Naming Andrew presses on why Evgeny shut down paying agency clients instead of maintaining a $50k/mo agency model. Evgeny explains that custom prompt adjustments were unsustainable, and breaks down why AI coding tools still demand deep software architectural knowledge.

speaking balance: gold is Andrew, purple is the guest (3 minute bins)

0:00 · Andrew 38.7% · guest 61.3%0:00 · Andrew 38.7% · guest 61.3%3:00 · Andrew 17.4% · guest 82.6%3:00 · Andrew 17.4% · guest 82.6%6:00 · Andrew 40.3% · guest 59.7%6:00 · Andrew 40.3% · guest 59.7%9:00 · Andrew 29.5% · guest 70.5%9:00 · Andrew 29.5% · guest 70.5%12:00 · Andrew 48.7% · guest 51.3%12:00 · Andrew 48.7% · guest 51.3%15:00 · Andrew 36.1% · guest 63.9%15:00 · Andrew 36.1% · guest 63.9%18:00 · Andrew 19.1% · guest 80.9%18:00 · Andrew 19.1% · guest 80.9%21:00 · Andrew 16.4% · guest 83.6%21:00 · Andrew 16.4% · guest 83.6%24:00 · Andrew 25.3% · guest 74.7%24:00 · Andrew 25.3% · guest 74.7%27:00 · Andrew 10.9% · guest 89.1%27:00 · Andrew 10.9% · guest 89.1%30:00 · Andrew 29.9% · guest 70.1%30:00 · Andrew 29.9% · guest 70.1%33:00 · Andrew 36.3% · guest 63.7%33:00 · Andrew 36.3% · guest 63.7%
Sharpest disagreement ▶ 25:43 Evgeny rejects 'little tweaks' framing

Evgeny firmly refutes Andrew's assertion that client requests were just minor adjustments, explaining the deep instruction logic and backtesting required for each niche.

 Hardest push from Andrew ▶ 26:25 Andrew defends the recurring agency business model

Andrew challenges Evgeny's decision to drop clients, arguing that generating $40k in 40 days could easily support a $50k monthly recurring agency.

 Biggest teaching moment ▶ 17:37 Deep dive into model fine-tuning mechanics

Evgeny educates Andrew on the technical complexities of audio transcription, LLM-as-a-judge scoring, and supervised fine-tuning epochs versus simple Zapier prompting.

 Andrew holds their own ▶ 13:33 Andrew diagrams the voice automation tech stack

Andrew demonstrates his automation expertise by mapping the exact integration chain from CRM triggers to Zapier, Vapi API voice calling, and calendar booking.
the scores for every segment, with the reasoning behind each
ChapterTopicAndrew as informed peerGuest teachingGuest disagreementAndrew pushing backWhy
The Cold Calling Grind in Real Estate 2613 Andrew makes optimistic assumptions about expired listings having high conversion rates and asks why a 1-2% rate was frustrating. Evgeny educates him on real estate reality, explaining that disgruntled homeowners often reject agents and that 200 dials are needed just to reach one homeowner.
Developing the First AI Voice Agent Prototype 3413 Andrew challenges Evgeny on why he did not simply scale a real estate brokerage empire with his tool instead of pivoting to software. Evgeny explains his CS background and personality mismatch with the rigid, decades-old real estate business model.
Client Acquisition and Early Low-Code Agency Architecture 4302 Andrew digs into the lean customer acquisition funnel using native Meta lead capture forms without a landing page. Evgeny describes qualifying leads via Google Meet and screening out unqualified agents despite high demand.
Voice Automation Workflows and Model Fine-Tuning 5713 Andrew maps out how an agency could build outbound voice agents for other trades like plumbing using Zapier and Vapi. Evgeny explains that true conversational salesmanship requires complex supervised fine-tuning data pipelines and LLM-as-a-judge scoring rather than basic prompt engineering.
Transition to Self-Serve SaaS and Onboarding Aiden 3201 Evgeny discusses transitioning away from client services to build software full-time and finding Aiden via YC Co-founder matching. Andrew brings Aiden in to define operational responsibilities and playfully notes Aiden's persistent outreach.
SaaS Customization, AI Coding Tools, and Company Naming 5524 Andrew presses on why Evgeny shut down paying agency clients instead of maintaining a $50k/mo agency model. Evgeny explains that custom prompt adjustments were unsustainable, and breaks down why AI coding tools still demand deep software architectural knowledge.

Statements from this episode (11)

Disclosure
Rezora charged early clients up to $2,500 setup plus $500 monthly
“So we would charge anywhere from an initial thousand to 2500 dollars setup fee, depending on the complexity and workflow, plus an ongoing 500 dollar month monthly fee, plus 20 cents per conversational minute that the AI did as well.”
Evgeny Matze Jan 8, 2026 ▶ 0:48
Disclosure
Matze: Rezora's early product was just GPT and text-to-speech
“And at the end of the day, it was just a GPT with the Texas speech model.”
Evgeny Matze Jan 8, 2026 ▶ 1:19
Insight
Matze: Expired Real Estate Listings Convert at Typical Cold Call Rates
“I would say they have a typical conversion rate as any other cold call, especially since They have, they had a very negative experience with the whole entire process, not being able to list it. I mean, not being able to sell their home with an agent before, an…”
Evgeny Matze Jan 8, 2026 ▶ 2:36
Assertion Not checkable as stated
Matze: Real Estate Agents Can Dial 200 Numbers per Live Connection
“First, you have to actually get connected to a person that picks up on you. That's okay. You can call 200 people before you actually get connected to someone that is the correct homeowner.”
Evgeny Matze Jan 8, 2026 ▶ 3:26
Assertion Not checkable as stated
Matze: Custom AI Voice Agent Booked a Listing Appointment on Day One
“And when I did that, I actually got my first listing appointment went to the first day.”
Evgeny Matze Jan 8, 2026 ▶ 5:07
Assertion Not checkable as stated
Matze: Rezora achieved over 80% sales conversion rate on demo calls
“And shockingly enough, we had above 80% conversion rate.”
Evgeny Matze Jan 8, 2026 ▶ 10:40
Assertion Not checkable as stated
Matze: No Off-The-Shelf LLM Is Designed For Sales Conversations
“The main thing is there's no LLM, there's no chat GPT, there's no cloud that is specifically meant to respond for sales or human-like conversations.”
Evgeny Matze Jan 8, 2026 ▶ 15:49
Disclosure
Matze: Rezora Fine-Tunes LLMs On Real Sales Call Transcripts
“Currently what we're doing is we take real conversations. We turned that into fine tuning data to be specific, supervised fine tuning data, and we run supervised fine tuning on a large language model, which in end results in the language model actually soundin…”
Evgeny Matze Jan 8, 2026 ▶ 16:33
Assertion Not checkable as stated
Matze: Zero Published Research On Fine-Tuning LLMs For Voice Sales
“There's literally no research out there on this topic. Zero.”
Evgeny Matze Jan 8, 2026 ▶ 18:29
Insight
Matze: AI coding tools produce errors without architectural understanding
“The thing that people have to understand is that with cloud code or any of these coding tools, they're just that they're tools. If you don't know how to swing a hammer, the nails could go in a bent. If you know how to swing a hammer, the nails to go in straigh…”
Evgeny Matze Jan 8, 2026 ▶ 30:25
Assertion Not checkable as stated
Matze: Rezora kept $34K to $35K profit on $40K agency revenue
“I want to say we kept Straight profit. Around 34,000 dollars. 35,000 dollars.”
Evgeny Matze Jan 8, 2026 ▶ 32:50
Made with StarZero

Turn any episode into a week of clips.

This entire site, about 40 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.

Try the podcast clipper Learn more