Welcome to GTM Vault, trusted by 27,000+ GTM leaders building the future of revenue.

This week’s guest is Rafael Broshi, co-founder & CEO of Notch — the company pioneering resolution-first AI agents. Notch isn’t about smarter chatbots or reply-only wrappers. It’s about agents that solve customer problems end-to-end: issuing refunds, verifying accounts, processing claims, and more.

The results? 40x faster response times, 80% lower cost to serve, and double-digit lifts in conversion rates. Enterprises like Mixtiles and Guardio are already replacing bloated tier-one teams with lean, AI-first operations powered by Notch.

Before founding Notch, Rafael spent a decade in the Air Force, led product strategy at AGT International, and invested with Foundation Partners. His career across defense, product, and investing shaped how he builds AI for mission-critical production environments — not just slick demos.

“Reply-only AI caps at 70% accuracy.

Resolution-first means orchestrating agents and deterministic rules until you’re at 95–99% — good enough for production, not just POCs.”

This one’s for founders designing AI-native orgs, operators buried under CX costs, and execs rethinking how support turns from a cost center into a growth engine.

This episode in the Library: the written playbook, the stack and the full transcript, free and open. Open it at pro.gtmvault.co.

Listen & Subscribe

Listen on YouTube

Listen on Spotify

Listen on Apple

In This Episode

Why reply-only bots are a dead end — and what resolution-first unlocks
How autonomous agents are already closing millions of tickets end-to-end
The GTM playbook for replacing bloated support orgs with lean AI-first operations
Why agentic architecture = rules + workflows + LLMs (not just “throw it to a model”)
How Notch drives both cost savings and revenue growth
Which industries are adopting fastest (finance & telecom may surprise you)
Why CX orgs in 2027 could be 70% smaller — and more strategic than ever

5 GTM Takeaways to Steal

Reply-Only Is Dead – Bots that just answer FAQs cap at 70–80% accuracy. Production requires 95–99%.
Orchestration Beats One Model – Agentic workflows + deterministic rules scale where single LLMs fail.
Cost Is the Wedge, Growth Is the Hook – Notch clients triple VIP campaign response rates with AI sales assistants.
Managed Service = GTM Channel – Every successful implementation drives 3–5 warm referrals.
Process Discipline Is Underrated – Clear SOPs make AI adoption 10x faster and easier to scale.

Episode Highlights

00:00 – Intro: Meet Rafael Broshi (Notch)

02:36 – Why CX was broken and ripe for reinvention

03:10 – From chatbots to resolution-first AI

04:28 – Why LLMs cap at 70–80% accuracy in production

06:57 – Inside agentic architecture: rules and LLM workflows

10:09 – Replicating tone, policies, and company voice

12:04 – Case study: 3x revenue lift with AI sales assistant Ella

14:29 – Metrics that matter: CSAT, cost-to-serve, conversion

15:48 – Early ICPs: e-commerce, finance, telecom

16:59 – Why managed service beats ads as GTM motion

18:26 – Balancing product and service in enterprise AI

19:50 – When autonomous agents go mainstream

20:45 – Overrated vs underrated AI trends in CX

22:00 – What a CX org looks like in 2027

24:31 – When founders should start designing AI-first orgs

25:59 – Lessons from Air Force and VC shaping GTM

26:49 – Selling AI in production vs demo environments

28:01 – Rapid Fire

Further Reading

Want to go deeper? Related GTM Vault playbooks & frameworks:

Sponsor Spotlight: ZoomInfo

You might know ZoomInfo for contacts and emails — the fuel behind SDR teams. But ZoomInfo has quietly rebuilt itself into something much more powerful: a system that automates your entire GTM motion.
Signals come in. Outreach goes out. The list gets worked. Plays get triggered. Automatically.
ZoomInfo calls it GTM Intelligence. It’s not enrichment. It’s execution.

See a demo today

Thanks to our sponsors who help keep this newsletter free and high-signal. Want to reach 27,000+ GTM leaders? Explore sponsorship options here.

Connect

Follow Rafael Broshi: LinkedIn // Notch
Follow Rick Koleta: LinkedIn // RiteGTM

Resolution-first AI isn’t just fixing support. It’s redefining what lean, AI-native GTM looks like.
Catch up on all GTM Vault episodes →

Full transcript, searchable to the timestamp: pro.gtmvault.co/episodes/gtm-28.

GTM 28 | Resolution-First AI: Support as a…