Gabe Larsen, CRO of Atonom, canceled a $40KSalesforce contract with no replacement plan and built a working CRM in 3 hours using Lovable. This is the full story plus his playbook on AI SDRs and the autonomous organization.

In Episode 13 of GTM Hackers, host Charles Brun sits down with Gabe Larsen to unpack two of the most talked about shifts in go to market today. First, the rise of the autonomous organization where AI cloud employees work side by side with humans. Second, the bold decision to rip out Salesforce and vibe code a custom CRM using Lovable.

Gabe brings 20 plus years of GTM experience from Goldman Sachs, Accenture, Inside Sales.com, and Customer (acquired by Meta). He now leads revenue at Atonom, a startup building the cloud employee category across marketing, sales, customer success, support, and HR.

What you will learn in this episode:

1. Why the traditional model of scaling with headcount and tools is dead, and what is replacing it.

2. How to think about AI SDRs as employees rather than tools or cost per conversation.

3. The full 4 step onboarding process Atonom uses for every new AI cloud employee, covering identity, memory, tools, and role play.

4. Why multi agent architectures beat deterministic workflow tools like Make and N8N for complex GTM jobs.

5. The real world performance of AI SDRs on inbound lead qualification, including a case study with a 1000 percent improvement in lead response time.

6. Why cold outbound AI SDRs are failing and what signal based selling with cloud employees looks like instead.

7. The hot, warm, and cold framework for AI in GTM motion.

8. The exact moment Gabe decided to cancel Salesforce and why his whole team was relieved.

9. How his CFO and head of finance built the first version of the new CRM in 3 hours over root beers in Salt Lake City.

10. Why the maintenance argument against vibe coded CRMs is overblown.

11. The benefits of a conversational CRM over traditional reporting.

12. Why Gabe never signs SaaS contracts longer than 12 months.

Discussion topics covered:

Autonomous organization and frontier firms

Cloud employees and AI SDR architecture

Multi agent systems and supervisor agents

AI employee onboarding and coaching loops

Inbound vs outbound AI SDR performance

Signal stacking and de anonymized website visitor handoffs

Hot, cold, and warm lead handling with AI

The 3 day inbound conversion rule

Customer service AI and NPS performance data

Ripping out Salesforce at a startup

Vibe coding with Lovable

Conversational CRM as the new interface

Build vs buy decisions for modern GTM stacks

AI adoption mistakes most GTM teams are making

Guest:

Gabe Larsen is the CRO of Atonom, host of the Talk AI podcast, and a veteran GTM operator with leadership experience at Customer (Meta), Inside Sales.com, Accenture, and Goldman Sachs.

Host:

Charles Brun is the host of GTM Hackers, a B2B video podcast focused on AI powered go to market strategies. He is also the founder and CEO of Everest.

If you are a founder, CRO, VP of Sales, GTM engineer, RevOps leader, or operator thinking about AI adoption, this conversation delivers both the big picture vision and the tactical details you need.

Follow GTM Hackers, rate the show, and share this episode with a GTM operator who needs to hear it. New episodes drop weekly.

Related topics: AI in sales, AI SDR, vibe coding, replacing Salesforce, autonomous organization, GTM engineer, agentic AI, cloud employees, RevOps, conversational CRM, build vs buy, AI adoption, signal based selling, Lovable, no code CRM.

A Google AI engineer built a fully functional Slack clone in 14 days using Claude Code — without manually writing or editing a single line of code. In this episode, Nathan shares the full story.

Nathan is an AI Engineer at Google who works with the company's largest enterprise customers to build generative AI solutions. He issued himself a public challenge: clone Slack in two weeks using only vibe coding and AI agents. He did it. The project — called Slawk — is open source, already has 250 GitHub stars, and has sparked a wide debate about the future of software engineering.

In this conversation with GTM Hackers host Charles Brun, Nathan and Charles break down exactly how it was built, what went wrong, what surprised him, and what it all means for founders and GTM leaders thinking about AI-powered product development.

What we cover in this episode:

1. Why Nathan decided to clone Slack as his side project challenge

2. The tech stack he chose and how he planned the architecture with Claude

3. How he used a browser-connected MCP agent to visually compare and clone Slack's UI

4. The 3-agent system he built: QA agent, Dev agent, and CTO orchestrator

5. How the QA agent filed GitHub issues with GIF screenshots — automatically — while Nathan slept

6. Test-driven development in a vibe coding workflow

7. The real cost of the project (under $200 total)

8. Why cloud infrastructure solves most scalability concerns critics raised

9. Why MCP servers make traditional app integrations obsolete

10. Nathan's vision for an AI-native Slack and why incumbents can't build it

11. The identity crisis behind anti-AI pushback in the software engineering community

12. Nathan's next project: QA as a Service

Nathan's key argument: only 1% of people can currently build software because they need to know how to code. Vibe coding is about to change that ratio dramatically. And the companies that figure this out first will have a significant competitive advantage.

If you are a founder, a GTM leader, or a builder exploring AI tools, this episode is one of the most practical and honest accounts of what vibe coding can actually do in 2025.

Find Nathan on LinkedIn — he replies to everyone in DMs. The Loc GitHub repository is linked in the show notes.

GTM Hackers is hosted by Charles Brun, a B2B SaaS go-to-market advisor and founder with 15 years of experience scaling revenue with AI. Follow GTM Hackers on Spotify to get notified of new episodes every week.

Related topics: Claude Code, vibe coding, agentic AI, multi-agent workflows, AI software engineering, Slack alternatives, open source tools, SaaS cost reduction, AI for startups, go-to-market AI strategy, MCP server, model context protocol, AI QA testing, building in public, AI product development.

In many startups, go-to-market breaks when the team scales: more tools, more handoffs, more chaos.

In this episode of GTM Hackers, Charles Brun sits down with Christian Lim, co-founder of Youno, to explain a more durable approach: start by building GTM infrastructure that is easy to maintain, then layer automation and AI agents on top.

You will learn:

1. Why tool sprawl creates a maintenance tax that slows teams down
2. How API-first choices make your GTM stack more resilient over time
3. The two common GTM failure points: misalignment and messy CRM data
4. How CRM deduplication and enrichment improve outbound execution and forecasting
5. How simple AI classification prompts can keep data structured (for example, company size)
6. How a Slack-native assistant can help reps find prospect details faster
7. A practical workflow to automate weekly client updates using Dust and ClickUp8) How to drive AI adoption with one internal champion and one high-impact workflow

Guest: Christian Lim (Youno)

Host: Charles Brun