BILLIONS

Today on BILLIONS, I’m sitting down with Daniel Chait, co-founder and CEO of Greenhouse, the hiring platform used by more than 7,000 companies.
Daniel has spent more than 27 years as an entrepreneur and built Greenhouse from an idea VCs initially dismissed into a $100M+ ARR company.

Now, hiring is being disrupted again.
Candidates are using AI to apply to more jobs and make every resume look better.
Companies are responding by using AI to filter thousands of increasingly similar applications.
Daniel calls it the “AI doom loop.”
And Greenhouse is betting that the answer to AI breaking hiring… might actually be better AI.

In this masterclass, we break down:
→ The 25-customer test Greenhouse used to prove product-market fit
→ The fundraising hack that turned investor rejection into demand
→ Why Greenhouse cut 30% of its team during COVID — only for hypergrowth to return weeks later
→ How the company crossed $100M ARR and brought in TPG as a majority investor
→ Why AI has made hiring worse for both candidates and recruiters
→ Why resumes are becoming a worse signal for identifying great talent
→ Greenhouse’s acquisition of Ezra AI Labs and its bet on Voice AI interviews
→ Why Daniel believes every candidate could eventually get a first conversation
→ How AI could actually remove some of the most biased parts of traditional hiring
→ Why SaaS pricing is moving from seats, to usage, and eventually outcomes
→ Why founders shouldn’t hire a sales leader before figuring out how to sell
→ How to hire great executives for jobs you’ve never done yourself

We also discuss startup hustle, acquisitions, surviving market cycles, the future of resumes, whether candidates will accept being interviewed by AI, and what the next generation of hiring software could look like.

If you want to understand how AI is changing hiring — and how a $100M+ SaaS company is trying to reinvent itself before someone else does — this is the episode.

TIMELINE:
00:45 – Building Greenhouse and why hiring was broken
03:41 – Getting rejected by VCs and ignoring startup hype cycles
06:34 – The fundraising hack that got investors to say yes
09:40 – The job fair hustle that gave Greenhouse its first insights
11:29 – The 25-customer product-market-fit test
13:14 – Cutting 30% during COVID — then hypergrowth came back
16:45 – Buy vs. build: how Greenhouse thinks about acquisitions
18:14 – Acquiring Ezra AI Labs and betting on Voice AI
20:21 – The “AI doom loop” breaking hiring
23:25 – Do we still need resumes and job applications?
25:35 – 3 bets made by Greenhouse
27:45 – Will candidates accept being interviewed by AI?
31:55 – From SaaS seats to usage and outcome-based pricing
33:55 – Bigger mistake : why founders hire sales leaders too early
35:45 – How to hire for a job you’ve never done yourself

Today on BILLIONS, I’m sitting down with Kieran Flanagan, SVP of Agentic GTM & Systems at HubSpot.
Kieran spent 10 years helping build one of the most copied growth engines in SaaS, left to become CMO of Zapier, and then came back to HubSpot to help rebuild its entire go-to-market playbook around AI.

And they had no choice but to rethink it.
In just one year, HubSpot lost 140 million visits — around 80% of its organic traffic — as Google and AI completely changed how people search.
Yet the company kept growing.

Some of HubSpot’s AI-powered experiments increased conversion rates by 3x, 5x, and even 10x.

Today, HubSpot serves more than 300,000 customers, generates more than $3B in revenue, and is rebuilding the growth engine that made it famous for an AI-first world.

In this masterclass, we break down:
→ How HubSpot lost 80% of its organic traffic and still kept growing
→ The AI experiments that generated up to 10x conversion improvements
→ Why Kieran believes informational SEO is being killed by AI
→ How HubSpot is optimizing for ChatGPT, Gemini and answer engines
→ Why “taste and judgment” are becoming more valuable as AI gets better
→ How to build context layers that make AI agents smarter over time
→ Why junior employees may actually struggle more with AI than experts
→ How HubSpot built a network of 350 creators to replace part of its search-driven growth
→ Why some shows inside that network grew 5x
→ How more than half of HubSpot’s revenue now comes from customers who started as free users
→ The new distribution channels and ecosystems HubSpot is betting on next

We also discuss why AI is creating a “sea of sameness”, why perfectly polished content can now be a disadvantage, how creators can influence visibility inside AI engines, and what happens to Google when users no longer need to click on links.

If you want to understand how one of the biggest growth companies in SaaS is rebuilding its playbook for the AI era, this is the episode.

TIMELINE:
00:50 – Building a billion-dollar company before vs. after AI03:14 – Using AI to personalize emails at scale05:40 – How HubSpot lost 140M organic visits08:26 – The new AI-first marketing playbook11:55 – Building taste and judgment into AI15:12 – Kieran’s AI Content OS18:11 – Why AI-generated content all looks the same23:48 – Replacing SEO with new distribution channels25:28 – HubSpot’s 350-creator network27:18 – How this creator network works?31:28 – Winning inside ChatGPT and answer engines34:50 – HubSpot’s next big growth bets37:10 – AI, agencies and HubSpot’s partner ecosystem39:28 – Kieran’s new book, Loop

Today on BILLIONS, I'm sitting down with Dan Rogers, CEO of Asana, to discuss AI agents, the future of SaaS, and what comes after the chatbot.

Dan took over an established software business with around $800 million in revenue. His next challenge: turning a platform built for human collaboration into an operating system for teams of humans and AI agents.

A model can generate an answer. But who needs to approve it? Which team acts next? What has already been decided? Dan's argument is that this context and coordination - not just the intelligence of the model - will determine whether AI can do useful work inside a company.

At the time of our conversation, AI products were contributing 25% of Asana's net new ARR. We get into the decisions behind that shift: charging for fulfilled requests rather than tokens, giving users an initial AI allowance at no extra cost, and buying Stack AI to accelerate the roadmap.

We also discuss keeping teams focused through market volatility, integrating an acquired company without losing its people, and why past success can become a liability when you're taking over a new business.

In this masterclass, we break down:
Beyond the Chatbot: Why generating an answer is different from coordinating work, and what changes when your team includes AI agents.
The Request-Based Pricing Bet: Why Asana is hiding token complexity from customers and charging around the fulfillment of a request.
Value Before the Sale: The decision to give users an initial AI allowance, help them discover the value, and let adoption drive further spending.
Buying Time: How the Stack AI acquisition accelerates Asana's roadmap, and why the small details of integration matter as much as the big strategic decisions.
The Model Moat Test: Does your business become more or less valuable when AI models improve? Dan explains why he believes Asana benefits.
Leading Through the Noise: Listening to customers, finding wins beyond the daily stock price, and questioning the playbook that made you successful before.

SUBSCRIBE for more episodes of BILLIONS.

Follow meLinkedIn: https://www.linkedin.com/in/-g-/X: https://twitter.com/GuillaumeMbhTikTok: https://www.tiktok.com/@guillaumemoubecheInstagram: https://www.instagram.com/guillaume_moubeche/

Today on BILLIONS, I'm sitting down with Carles Reina, the first investor and fourth employee at ElevenLabs, to unpack the AI bubble, the rise of Chinese models, and Europe's fight to stay in the race.
Carles helped build ElevenLabs' go-to-market engine on the road to over $600 million in annual recurring revenue. As the founder of Baobab Ventures, he also sees the other side of the boom: the funding rounds that may be getting ahead of reality.

Before the scale came 80 to 100 sales meetings a week, a Google Sheets CRM, and months of testing who would actually pay. Then came a sales model most teams would question: quotas set at 20 times base salary, uncapped commissions, and average quota attainment of around 167%.

But this conversation goes beyond the ElevenLabs growth story. We debate the commercial barriers Carles sees around Chinese AI models, why he believes Europe is not investing seriously enough in AI infrastructure, and what it takes to build a global company from day one.

He also explains why extraordinary growth and an AI bubble can exist at the same time - and why he expects some aggressively valued startups to face a funding reality check within six to nine months of this conversation.
In this masterclass, we break down:
The 100-Meeting Week: How Carles tested his way from creator demand to a repeatable sales motion, then hired reps who started closing deals within weeks.
The 20x Quota Rule: Why ElevenLabs set targets at 20 times base salary, rewarded overperformance with commission accelerators, and adjusted expectations when markets proved tougher.
Permission to Fail: Why Carles encouraged experiments with company money - and only needed one idea out of 100 to unlock the next stage of growth.
The AI Adoption Gap: Why strong Chinese models do not automatically translate into easy enterprise adoption, and why Europe needs more than technical talent to compete.
The Operator-Investor Playbook: Why Carles kept Baobab small despite $23M in commitments, and why he wants to return capital to investors along the way.
The Valuation Reality Check: How real AI demand can coexist with overheated funding rounds, and why a high valuation can become a liability at the next raise
Global from Day One: Why Carles challenges the one-market-at-a-time playbook, pushes founders to hire go-to-market talent early, and wants investors to do more than write checks.

TIMELINE
00:00 - Intro - from first investor to fourth employee at ElevenLabs
03:25 - 80-100 sales meetings a week: finding what actually sells
06:46 - Hiring "icebreakers" vs. scaling the sales team
09:36 - Selling the future: from developers to enterprise
12:25 - Building a sales team that experiments with AI
17:05 - The 20x quota rule, uncapped commissions & 167% attainment
21:46 - The incentive mistake that held back enterprise sales
24:13 - Chinese AI models: great technology, harder enterprise adoption
34:01 - Europe's AI infrastructure problem
34:38 - Hugging Face, Nvidia & the European funding gap
37:37 - Building Baobab Ventures: fundraising and early results
43:37 - AI bubble or real growth? The next funding reality check
47:53 - Three lessons for founders building globally

SUBSCRIBE👇https://www.youtube.com/c/GuillaumeMoubeche?sub_confirmation=1

Join the lemlist family 🚀
https://community.lemlistfamily.com/join?invitation_token=9db20d2c4707b75ebb065462261c4665a3c35a59-aad63a7f-90a5-452e-a628-4ccbcbaa84ca
Follow me
https://www.linkedin.com/in/-g-/
https://twitter.com/GuillaumeMbh
https://www.tiktok.com/@guillaumemoubeche
https://www.instagram.com/guillaume_moubeche/
Learn how to book meetings with 30%+ of your prospects (for free): https://www.lemlistfamily.com/multichannel-masterclass?mtm_campaign=5008&mtm_source=organic&mtm_medium=youtube

Today on BILLIONS, I'm sitting down with Aaron Levie, the co-founder and CEO of Box, who has rapidly transitioned his enterprise platform from a pure SaaS model into a cutting-edge playground for autonomous AI agents.
Aaron has a masterclass view of the data infrastructure that legacy tech giants wish they controlled. In this conversation, we pull back the curtain on the high-stakes battle for sovereign AI.
We unpack the real fallout of the U.S. government's unprecedented export controls on Anthropic's frontier models, why data platforms like Snowflake are posting blockbuster quarters amidst the AI boom, and how specialized tools like Cursor show that the future of intelligence is multifaceted, not winner-take-all.
If you want to understand where the real economic value of applied AI resides over the next decade, this is the blueprint.
In this masterclass, we break down:
The Export-Control Precedent: Inside the unprecedented restriction of Anthropic's frontier model from non-US users and why Aaron calls it a brand-new moment in AI regulation.
The Safety-Rhetoric Boomerang: How AI safety messaging scared the government into a model-approval pipeline — and why some safety advocates may quietly prefer that outcome.
China's $50B Scenario: Why Aaron believes China can simply throw $50B at compute to stay in the race and why France, Japan, the UK, and Germany may be forced to build their own sovereign models.
The Multifaceted Intelligence Future: Why the AI market won't be "winner-take-all," and how Cursor's applied-layer harness (routing tasks across cheap and premium models) became the template.
Building for Machine Users: How Box adapted its file system MCP server, CLI, Markdown editing, HTML compatibility so agents and people work off the same data.
Why More Agents Make SaaS More Valuable: Why deploying 100x more agents than employees increases the value of the underlying CRM, ERP, and content systems instead of killing them.
The CS-Grad Dislocation: A grounded look at the shifting job market away from Big Tech layoffs and into AI startups and industries like life sciences and manufacturing.
TIMELINE :
00:00 – Turning Box from SaaS into an AI agent platform
03:00 – Sovereign AI: why countries will build their own models
05:51 – Can open and Chinese models catch up to the frontier?
08:51 – The chip embargo debate and Jensen Huang's argument
14:09 – AI safety, regulation, and government model approval
18:29 – Why AI won't be winner-take-all (the Cursor case study)
21:33 – How Box built a file system for AI agents
27:17 – Is SaaS dead? Why agents make software more valuable
30:48 – Will AI replace jobs? The truth about CS grads