5 Key Advice on Starting an AI Business

Youtube URL: https://youtu.be/CjTTSa33axg

Host(s): Syncly channel (solo talk)

Guest(s): Joseph Lee, Co-founder and CEO, Syncly; Co-founder, SUALab (acquired)

Podcast Overview and Key Segments

Overall Summary

Joseph Lee shares five core lessons from building two AI companies: SUALab, which sold for $200 million, and Syncly, an AI platform for scaling customer feedback insights. He stresses value over hype, scrappy customer discovery, and proving outcomes before product polish. He explains how his team cold-called 500 targets, offered risk-free pilots, and delivered results inside tools customers already used. He outlines how to find “hair-on-fire” problems by testing problem statements across many users, then narrowing to what repeats. He dives into co-founder alignment, timing with AI maturity, and building trust through community and social proof. The talk is direct, practical, and focused on what drives revenue and trust, not on the AI label.

Reference

AlphaGo: An AI program by DeepMind that beat human champions at Go in 2015–2016.
DeepMind: Google’s AI research lab that built AlphaGo.
YC: Y Combinator, a startup accelerator.
Hair-on-fire problem: A problem so urgent that customers will pay now to fix it.
Runway: How long a startup can operate before it runs out of cash.
LLM (Large Language Model): AI model trained on large text data for language tasks.
Looker: A business intelligence tool (now part of Google Cloud).
Tech stack: The mix of software tools a team uses.

Key Topics

Value Over Technology Hype

Joseph repeats that customers buy outcomes, not AI. Early on, a buyer told him, “I don’t care if you use AI or not.” That line shaped how he sells and builds. AI is a means to a clear result. If the product does not deliver value, AI branding will not save it. He warns that hype can raise money, but without a value prop, you risk running out of runway. The core is to define the job-to-be-done, prove it fast, and talk in the customer’s language. The technology choice matters less than solving a painful, frequent, and budget-backed problem.

Scrappy Customer Discovery and First Revenue

It took SUALab almost three years to revenue and fivefold annual growth later. The path started with manual outreach. They called about 500 fashion companies without any ties. They walked factories with laptops and a one-pager. They sold the end outcome, not fancy tech. They offered a three-month proof with “pay what you think it’s worth” if it worked. Three out of four paid. This shows how to earn trust: targeted outreach, onsite learning, and clear stakes around value, not features.

Finding Hair-On-Fire Problems

Start with five problem bullets tied to the customer’s goals for the year. Stress-test those bullets across 10 prospects. Narrow to three. After 15–100 conversations, one pattern repeats. That is your “hair-on-fire” problem. Joseph advises to push past surface complaints. Customers will not share root issues at first. You must earn trust and map daily workflows. The repeated, measurable pain with clear success criteria becomes the wedge for your product and sales narrative.

Deliver Value in Existing Workflows

Early customers do not want more tools. “Nobody wants a tech stack on top of existing tech stack.” Joseph’s team shipped results through tools customers already used, such as Looker or existing support platforms. Once users saw value, they built their own dashboards for depth. This reduces friction, speeds adoption, and keeps focus on outcomes. It also buys time to mature the product while customers get real wins.

Building Trust, Social Proof, and Community

Trust unlocks honest problems. Joseph runs meetups and training sessions to give first and build credibility. He leverages services, schools, and personal networks to seed social proof. The aim is that “people heard of you at least” before any sales call. Education and community create a safe space to share pains. That, in turn, shortens discovery cycles and supports faster pilots.

Co-founder Fit and Complementary Skills

“Finding co-founders is almost like a marriage.” Joseph stresses aligned values, especially on technology-first vs. customer-first. Neither is wrong, but misalignment is costly. Complementary skills are key so founders rely on each other. His long-run fit with his CTO enabled two companies across a decade. Shared history and clear roles reduce friction and speed execution when stakes are high.

Timing, AI Maturity, and Market Choice

He jumped into AI after AlphaGo sparked belief in real-world impact. SUALab focused on pattern recognition in manufacturing quality because AI could excel there. With Syncly, he saw LLMs finally good enough to make customer feedback analysis work at scale. He also shifted from Asia-heavy markets to a global product. New markets mean new trust-building, but also larger upside. Timing your product to tech maturity and market readiness is vital.

Outcome-Based Selling and Risk-Free Pilots

Joseph sold the end state first. He pitched business impact, not architecture. He ran three-month proof-of-value pilots with a simple rule: no value, no pay; value, pay what you think it’s worth. This forces clarity on success metrics and aligns incentives. It also gives founders room to “sell themselves” even before a full product exists, using their skills to deliver outcomes fast.

Key Themes

Customers Buy Outcomes, Not AI

Joseph makes it clear that “AI” is just a tool. Buyers want jobs done. Features and models do not matter if the problem remains. Frame all messaging around outcomes, metrics, and workflow impact. This keeps you grounded and resilient as tech shifts. It also wins trust faster because you speak to what they care about: time saved, defects reduced, revenue up.

Quotes:
- “I don’t care if you use AI or not. What I want is to get my job done.”
- “AI or any other technology is just a medium to provide the value.”

Earn Trust Through Proof and Community

Trust is built by giving value before the sale: meetups, training, and hands-on help. Then validate with low-risk pilots tied to clear goals. Use their current tools so adoption is easy. This reduces buyer anxiety, uncovers true pains, and increases conversion. It also builds your brand as a partner, not a vendor.

Quotes:
- “Give us three months to prove the value… If not, then you don’t need to pay.”
- “We try to open up a lot of meetups… to provide the value first.”

Scrappy Execution Beats Early Product Polish

Cold-calling 500 companies, walking factory floors, and pitching outcomes with a one-pager shows that hustle matters more than early product depth. Founders must sell themselves and solve a pain even if the product is not ready. This learning loop guides the product, reduces waste, and accelerates fit.

Quotes:
- “We called around 500 prospects in fashion industries.”
- “CEO, me CTO… we didn’t have a product. We just had a technology and an algorithm.”

Co-founder Alignment and Complementarity

Alignment on values and priorities keeps the team stable. Complementary skills enable speed and quality under pressure. Long-term relationships de-risk execution and improve judgment. Misaligned co-founders slow decisions and erode trust when it matters most.

Quotes:
- “Finding co-founders is almost like a marriage.”
- “Do you care the technology first or customer first?”

Timing and Market Selection

AI capability moves fast. Pick problems where current models deliver clear wins now. Syncly exists because LLMs finally make feedback analysis viable at scale. Global markets increase the prize but demand stronger trust-building and social proof. Timing matters as much as execution.

Quotes:
- “The trend is always changing… technology is just a medium.”
- “Because a large language model [is] getting much better… this is a great timing.”

Key Actionable Advise

Problem: Customers do not care about your tech.

Solution: Sell outcomes and business impact.
How to Implement: Define success metrics with the buyer. Tie demos to those. Use their language and workflows.
Risks: Overpromise without clear measurement; avoid vague ROI claims.

Problem: Hard to get first customers.

Solution: Scrappy outreach and on-site learning.
How to Implement: Build a 1-page value story. Cold-call 300–500 targets. Visit users. Observe workflows. Offer a risk-free pilot.
Risks: Time-consuming; be selective on target segments.

Problem: Unclear which problem to solve.

Solution: Find the hair-on-fire problem.
How to Implement: Start with five problem bullets from customer goals. Test across 10–50 users. Narrow to the one that repeats and has budget.
Risks: Confirmation bias; keep notes and force-rank pains by budget and urgency.

Problem: Early product adoption is slow.

Solution: Deliver value in tools customers already use.
How to Implement: Ship outputs via Looker, Google Sheets, or their support platform. Only add a new UI when value is proven.
Risks: Tool limits may cap differentiation; plan your migration path.

Problem: Low trust blocks honest feedback.

Solution: Build community and social proof first.
How to Implement: Host meetups and training. Share case studies. Leverage school, services, and network for intros.
Risks: Events without follow-up waste time; track engagement to sales.

Problem: Misaligned co-founders stall progress.

Solution: Align values and ensure complementary skills.
How to Implement: Discuss priorities (tech-first vs. customer-first). Map roles. Stress-test with small sprints before committing.
Risks: Hidden conflicts; set decision rules and escalation paths.

Problem: Pilots drag on without clear outcomes.

Solution: Outcome-based, time-boxed pilots.
How to Implement: Three-month proof. Define metrics upfront. “No value, no pay; if value, pay what it’s worth.”
Risks: Loose scope; fix success criteria and data access on day one.

Problem: Building the wrong product.

Solution: Sell yourself before the product.
How to Implement: Use services to deliver early wins. Convert wins into product features after value is clear.
Risks: Services creep; set a clear services-to-product transition plan.

Noteworthy Observations and Unique Perspective

Early revenue can take years, then growth can be explosive.
- “It took us almost… three years to generate the revenue… Every year we grew five times.”
Price after proving value reduces friction and creates trust.
- “If not, then you don’t need to pay… If you think there’s a value, pay whatever you think is valuable.”
Use AI when the timing is right, not just because it is hot.
- “The trend is always changing… technology is just a medium.”
Deliver value through existing tools to speed adoption.
- “Nobody wants a tech stack on top of existing tech stack.”
Co-founder fit is strategic, not cosmetic.
- “Finding co-founders is almost like a marriage.”

Companies, Tool and Entities Mentioned

Syncly
SUALab
DeepMind
AlphaGo
Y Combinator (YC)
Looker (Google Cloud)
Manufacturing and fashion industries
Consumer apps and e-commerce
Korea (market context)

Linkedin Ideas

Title: Customers Don’t Buy AI. They Buy Outcomes.

Main Point: Frame every pitch around the job-to-be-done and measurable impact.
Core Argument: AI is a medium. Value wins.
Key Quotes: “I don’t care if you use AI… I want my job done.” “Technology is just a medium.”

Title: The 500-Call Rule for Your First 10 Customers

Main Point: Scrappy outreach and on-site learning beat theoretical research.
Core Argument: Call widely, visit users, offer risk-free pilots.
Key Quotes: “We called around 500 prospects.” “Give us three months to prove the value.”

Title: Ship Value Inside the Tools They Already Use

Main Point: Reduce friction by delivering through the customer’s current stack.
Core Argument: Adoption grows when you meet users where they work.
Key Quotes: “Nobody wants a tech stack on top of existing tech stack.”

Title: How to Find Hair-on-Fire Problems in 30 Days

Main Point: Start with five problem bullets, test, and narrow by repetition and budget.
Core Argument: Pattern-finding in customer pains guides product fit.
Key Quotes: “Once you talk to… customers, you will hear the one common bullet point.”

Title: Co-founder Fit: Tech-First vs. Customer-First

Main Point: Align values and ensure complementary skills before you scale.
Core Argument: Misalignment is costly; set priorities early.
Key Quotes: “Finding co-founders is almost like a marriage.”

Blog Ideas

Title: Outcome-Based Pilots: A Playbook for Early AI Startups

Main Point: Time-boxed, “no value, no pay” pilots align incentives and speed trust.
Core Argument: Define metrics, deliver fast, then price on proven value.
Key Quotes: “Give us three months to prove the value.” “Pay whatever you think is valuable.”

Title: Selling Without a Product: Lessons from Factory Floors

Main Point: How to win early deals with a one-pager and on-site discovery.
Core Argument: Sell yourself and outcomes; productize after proof.
Key Quotes: “We didn’t have a product… We just tried to show them the end goal.”

Title: When AI Maturity Meets Market Need: Why Now for Feedback Intelligence

Main Point: LLMs now make feedback analysis useful at scale.
Core Argument: Timing matters; build when tech can deliver clear wins.
Key Quotes: “Large language model [is] getting much better… great timing.”

Title: Build Trust Before You Sell: Community as a GTM Strategy

Main Point: Meetups and training create social proof and uncover real pains.
Core Argument: Give first to earn honest feedback and faster adoption.
Key Quotes: “We try to open up a lot of meetups… to provide the value first.”

Title: The Co-founder Blueprint: Values, Roles, and Decade-Long Fit

Main Point: Align on priorities and complement each other to scale.
Core Argument: Long-term trust and clear roles beat credentials alone.
Key Quotes: “Finding co-founders is almost like a marriage.”