Aug 28, 2023 · 22m · top-founders

Smith AI Breaks $22m ARR Helping Companies Manage Voice, Chat Support, $100m Cap, Burning $300k per Month

Aaron Lee · 13m spoken Nathan Latka · 6m spoken
0:00 / 0:00

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

Nathan Latka interviews Aaron Lee, co-founder and CEO of Smith.ai, exploring how the company scaled past a $20 million ARR run rate through hybrid AI and human-powered customer communication services. Lee details his journey from Google and Red Beacon to bootstrapping Smith.ai, explaining their proprietary data moats, workforce model, and fundraising milestones.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Nathan holds 30.6% of the talking time here. How this is scored →

Nathan as informed peer 5.6 Guest teaching 3.0 Guest disagreement 1.7 Nathan pushing back 4.1
05100:0010:0020:000:35–3:37 · Nathan as informed peer 5/10 Overview of Smith AI Traction and Metrics Nathan guides the conversation through Aaron's early career at Google Video and Red Beacon with rapid-fire questions about founding timelines and initial funding rounds.3:39–6:07 · Nathan as informed peer 6/10 Acquisition of Red Beacon by The Home Depot Nathan breaks down corporate M&A deal structures, pressing Aaron on cash versus earnout splits and exact exit multiples when Aaron gives vague answers.6:08–8:11 · Nathan as informed peer 4/10 Corporate Tenure at Home Depot and Genesis of Smith AI Aaron explains how observing contractor challenges at Home Depot inspired Smith AI, while Nathan questions why he left a secure corporate role.8:11–10:34 · Nathan as informed peer 5/10 Bootstrapping and Launching Smith AI Aaron details bootstrapping Smith AI without salaries for three years and explains their hybrid AI and human pricing packages.10:36–14:20 · Nathan as informed peer 7/10 Revenue Milestones and Expanding Service Lines Nathan quickly runs the math on customer counts versus reported revenue and points out an apparent mismatch, forcing Aaron to clarify higher blended SDR service tiers.14:21–19:14 · Nathan as informed peer 7/10 AI Integration, Conversational Latency, and Proprietary Data Moats Nathan scrutinizes the cap on Aaron's convertible notes and cites research on historical milestones, while Aaron explains the proprietary data moat over generic LLMs and corrects Nathan's mixup of round details.19:14–21:46 · Nathan as informed peer 5/10 Burn Rate, Runway, and Future Strategy Nathan sharply calls out Aaron when he claims not to remember his monthly burn rate, before transitioning into the standard Famous Five questions.0:35–3:37 · Guest teaching 2/10 Overview of Smith AI Traction and Metrics Nathan guides the conversation through Aaron's early career at Google Video and Red Beacon with rapid-fire questions about founding timelines and initial funding rounds.3:39–6:07 · Guest teaching 2/10 Acquisition of Red Beacon by The Home Depot Nathan breaks down corporate M&A deal structures, pressing Aaron on cash versus earnout splits and exact exit multiples when Aaron gives vague answers.6:08–8:11 · Guest teaching 3/10 Corporate Tenure at Home Depot and Genesis of Smith AI Aaron explains how observing contractor challenges at Home Depot inspired Smith AI, while Nathan questions why he left a secure corporate role.8:11–10:34 · Guest teaching 3/10 Bootstrapping and Launching Smith AI Aaron details bootstrapping Smith AI without salaries for three years and explains their hybrid AI and human pricing packages.10:36–14:20 · Guest teaching 4/10 Revenue Milestones and Expanding Service Lines Nathan quickly runs the math on customer counts versus reported revenue and points out an apparent mismatch, forcing Aaron to clarify higher blended SDR service tiers.14:21–19:14 · Guest teaching 5/10 AI Integration, Conversational Latency, and Proprietary Data Moats Nathan scrutinizes the cap on Aaron's convertible notes and cites research on historical milestones, while Aaron explains the proprietary data moat over generic LLMs and corrects Nathan's mixup of round details.19:14–21:46 · Guest teaching 2/10 Burn Rate, Runway, and Future Strategy Nathan sharply calls out Aaron when he claims not to remember his monthly burn rate, before transitioning into the standard Famous Five questions.0:35–3:37 · Guest disagreement 1/10 Overview of Smith AI Traction and Metrics Nathan guides the conversation through Aaron's early career at Google Video and Red Beacon with rapid-fire questions about founding timelines and initial funding rounds.3:39–6:07 · Guest disagreement 1/10 Acquisition of Red Beacon by The Home Depot Nathan breaks down corporate M&A deal structures, pressing Aaron on cash versus earnout splits and exact exit multiples when Aaron gives vague answers.6:08–8:11 · Guest disagreement 1/10 Corporate Tenure at Home Depot and Genesis of Smith AI Aaron explains how observing contractor challenges at Home Depot inspired Smith AI, while Nathan questions why he left a secure corporate role.8:11–10:34 · Guest disagreement 1/10 Bootstrapping and Launching Smith AI Aaron details bootstrapping Smith AI without salaries for three years and explains their hybrid AI and human pricing packages.10:36–14:20 · Guest disagreement 2/10 Revenue Milestones and Expanding Service Lines Nathan quickly runs the math on customer counts versus reported revenue and points out an apparent mismatch, forcing Aaron to clarify higher blended SDR service tiers.14:21–19:14 · Guest disagreement 3/10 AI Integration, Conversational Latency, and Proprietary Data Moats Nathan scrutinizes the cap on Aaron's convertible notes and cites research on historical milestones, while Aaron explains the proprietary data moat over generic LLMs and corrects Nathan's mixup of round details.19:14–21:46 · Guest disagreement 3/10 Burn Rate, Runway, and Future Strategy Nathan sharply calls out Aaron when he claims not to remember his monthly burn rate, before transitioning into the standard Famous Five questions.0:35–3:37 · Nathan pushing back 2/10 Overview of Smith AI Traction and Metrics Nathan guides the conversation through Aaron's early career at Google Video and Red Beacon with rapid-fire questions about founding timelines and initial funding rounds.3:39–6:07 · Nathan pushing back 4/10 Acquisition of Red Beacon by The Home Depot Nathan breaks down corporate M&A deal structures, pressing Aaron on cash versus earnout splits and exact exit multiples when Aaron gives vague answers.6:08–8:11 · Nathan pushing back 3/10 Corporate Tenure at Home Depot and Genesis of Smith AI Aaron explains how observing contractor challenges at Home Depot inspired Smith AI, while Nathan questions why he left a secure corporate role.8:11–10:34 · Nathan pushing back 2/10 Bootstrapping and Launching Smith AI Aaron details bootstrapping Smith AI without salaries for three years and explains their hybrid AI and human pricing packages.10:36–14:20 · Nathan pushing back 5/10 Revenue Milestones and Expanding Service Lines Nathan quickly runs the math on customer counts versus reported revenue and points out an apparent mismatch, forcing Aaron to clarify higher blended SDR service tiers.14:21–19:14 · Nathan pushing back 6/10 AI Integration, Conversational Latency, and Proprietary Data Moats Nathan scrutinizes the cap on Aaron's convertible notes and cites research on historical milestones, while Aaron explains the proprietary data moat over generic LLMs and corrects Nathan's mixup of round details.19:14–21:46 · Nathan pushing back 7/10 Burn Rate, Runway, and Future Strategy Nathan sharply calls out Aaron when he claims not to remember his monthly burn rate, before transitioning into the standard Famous Five questions.

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

0:00 · Nathan 58.7% · guest 41.3%0:00 · Nathan 58.7% · guest 41.3%3:00 · Nathan 23.9% · guest 76.1%3:00 · Nathan 23.9% · guest 76.1%6:00 · Nathan 12.7% · guest 87.3%6:00 · Nathan 12.7% · guest 87.3%9:00 · Nathan 22% · guest 78%9:00 · Nathan 22% · guest 78%12:00 · Nathan 23.7% · guest 76.3%12:00 · Nathan 23.7% · guest 76.3%15:00 · Nathan 21.1% · guest 78.9%15:00 · Nathan 21.1% · guest 78.9%18:00 · Nathan 40.6% · guest 59.4%18:00 · Nathan 40.6% · guest 59.4%21:00 · Nathan 57% · guest 43%21:00 · Nathan 57% · guest 43%
Sharpest disagreement ▶ 19:20 Aaron brushes off monthly burn rate importance

When Nathan expresses incredulity that Aaron doesn't track burn rate, Aaron pushes back stating the burn is minor enough relative to their scale not to be top of mind.

Hardest push from Nathan ▶ 19:20 Nathan calls out burn rate deflection

Nathan refuses Aaron's attempt to dodge the question about cash burn, directly challenging him that it is the single most important metric.

Biggest teaching moment ▶ 15:34 Aaron contrasts generic LLMs with proprietary operational data

Aaron educates Nathan on why public LLM wrappers struggle in commercial call operations due to latency and zero-tolerance hallucination constraints.

Nathan holds their own ▶ 18:46 Nathan corrects revenue milestone timeline with research

When Aaron cannot recall when Smith AI crossed its first million in revenue, Nathan steps in citing his own database records to pin the date down to 2018.

the scores for every segment, with the reasoning behind each
ChapterTopicNathan as informed peerGuest teachingGuest disagreementNathan pushing backWhy
Overview of Smith AI Traction and Metrics 5212 Nathan guides the conversation through Aaron's early career at Google Video and Red Beacon with rapid-fire questions about founding timelines and initial funding rounds.
Acquisition of Red Beacon by The Home Depot 6214 Nathan breaks down corporate M&A deal structures, pressing Aaron on cash versus earnout splits and exact exit multiples when Aaron gives vague answers.
Corporate Tenure at Home Depot and Genesis of Smith AI 4313 Aaron explains how observing contractor challenges at Home Depot inspired Smith AI, while Nathan questions why he left a secure corporate role.
Bootstrapping and Launching Smith AI 5312 Aaron details bootstrapping Smith AI without salaries for three years and explains their hybrid AI and human pricing packages.
Revenue Milestones and Expanding Service Lines 7425 Nathan quickly runs the math on customer counts versus reported revenue and points out an apparent mismatch, forcing Aaron to clarify higher blended SDR service tiers.
AI Integration, Conversational Latency, and Proprietary Data Moats 7536 Nathan scrutinizes the cap on Aaron's convertible notes and cites research on historical milestones, while Aaron explains the proprietary data moat over generic LLMs and corrects Nathan's mixup of round details.
Burn Rate, Runway, and Future Strategy 5237 Nathan sharply calls out Aaron when he claims not to remember his monthly burn rate, before transitioning into the standard Famous Five questions.

Statements from this episode (16)

Assertion Supported
Aaron Lee: Google grew from 2,000 to 20,000 employees between 2004 and 2008
“When I joined, like Google had 2000 employees. When I left, it was 20,000, like 10 X.”
Aaron Lee Aug 28, 2023 ▶ 2:04
Assertion Not checkable as stated
Lee: Red Beacon Surpassed $10M Revenue Before Home Depot Acquisition
“Yeah, I think it was some, somewhere like 10 plus around time, but”
Aaron Lee Aug 28, 2023 ▶ 3:46
Disclosure
Aaron Lee: Smith.ai founders took no salary for first few years
“We bootstrapped the company. So in the first few years, we actually didn't take any salaries, like all the entrepreneurs out there.”
Aaron Lee Aug 28, 2023 ▶ 8:46
Disclosure
Aaron Lee: Smith.ai founders invested hundreds of thousands of personal capital
“Oh, hundreds of thousands of dollars. Yeah. Yeah. It's not, not small amount. It's significant.”
Aaron Lee Aug 28, 2023 ▶ 9:08
Assertion Not checkable as stated
Lee: Average Smith.ai customer pays close to $1,000 per month
“So it's around like I would say close to a thousand dollars. Like, of course we have like a smaller business that pay a few hundred dollars. We have bigger businesses that pay like close to 10,000 dollars per month.”
Aaron Lee Aug 28, 2023 ▶ 10:18
Assertion Not checkable as stated
Lee: Smith.ai has gross margins above 60%
“So I would say our gross margin is 60% plus.”
Aaron Lee Aug 28, 2023 ▶ 11:16
Assertion Not checkable as stated
Lee: Smith.ai currently serves 3,000 customers
“Three, 3000.”
Aaron Lee Aug 28, 2023 ▶ 11:37
Prediction Not checkable as stated
Lee: Smith.ai could reach 5,000 to 6,000 customers by year-end
“Oh, I think we can actually get to about like, if we are lucky, we probably can get a five to 6000.”
Aaron Lee Aug 28, 2023 ▶ 11:43
Assertion Not checkable as stated
Lee: Smith.ai crossed $20M ARR run rate
“We are doing, I think like a few months ago, we crossed twenty million dollar annual rate.”
Aaron Lee Aug 28, 2023 ▶ 11:58
Assertion Not checkable as stated
Aaron Lee: Smith.ai employs over 600 people across headquarters and agents
“So the team that we have about total, I would say their headquarter plus agents about more than 600 people.”
Aaron Lee Aug 28, 2023 ▶ 13:33
Assertion Not checkable as stated
Aaron Lee: All 600 Smith.ai employees and agents are in North America
“They're all in North America.”
Aaron Lee Aug 28, 2023 ▶ 14:10
Opinion
Lee: Private Business Calls Create A Data Moat Against Public LLMs
“ChatGPT, they're using all the public data, public information, but there's so much information that is not public, like all the claws in the chest for the company and for the businesses, they're not. So we have all the advantage.”
Aaron Lee Aug 28, 2023 ▶ 15:20
Disclosure
Lee: Smith.ai has raised close to $13M total
“I think together, like we raised, like with us, angel investors and the safe note, I think I was close to 13, one, three.”
Aaron Lee Aug 28, 2023 ▶ 16:21
Disclosure
Lee: Smith.ai raised on a SAFE cap near $100M
“Oh, it's much higher than that. I think it's higher than that. It was like close to more like a hundred.”
Aaron Lee Aug 28, 2023 ▶ 16:41
Opinion
Lee: Today's AI startups are unproven UI wrappers lacking revenue
“If you look at all the AI companies today, they have a nice, I call the UI wrapper on top of AI or chat GPT, but the revenue is very lacking. It's unproven.”
Aaron Lee Aug 28, 2023 ▶ 17:17
Disclosure
Lee: Smith.ai burns a few hundred thousand dollars per month
“Just a few 100,000 dollars.”
Aaron Lee Aug 28, 2023 ▶ 19:31
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 2,600 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.