While Jordan was busy eating waffles in Belgium (seriously, dude, they have them at the farmers market!), we hosted Yash Tekriwal from Clay for some show and tell.
For those who don't know Yash, he's literally the person who coined "GTM Engineer", back when Clay was a 7-person team and he was building their entire sales process from scratch. Former high school CS teacher turned automation wizard, Seth Godin altMBA alum, and the guy who built 50+ company-wide automations that saved Clay thousands of hours before "AI" was even a buzzword.
In other words: when Yash talks about the future of go-to-market, you listen.
If you missed it, grab some coffee because we're about to walk through three features that will be playing at a theater near you soon.
The Framework: Capability, Intelligence, Personalization
Before we dive into the demos, here's how Yash thinks about AI evolution, and why it matters for every growth hacker reading this:
Capability: AI doing things it couldn't do before (like agents, voice, video) Intelligence: AI doing tasks better than before (smarter models, better outputs)
Personalization: Making AI custom-tailored to your specific context
These three pillars aren't just academic, they're the foundation for everything Clay just shipped.
Feature #1: Clay Navigator - Web Actions That Actually Work
Remember the old days of scraping? You'd hit a website that required form submissions or pagination and just... give up?
Navigator changes that completely.
What it does: Autonomous web actions that can navigate, input data, and extract results from any website - even ones that require complex interactions.
The demo that broke our brains: Yash showed Navigator automatically:
Going to the FINRA website
Inputting individual names into search forms
Processing results to verify broker credentials
Providing a complete audit trail (you can literally watch a replay of what the AI did)
Why this matters: Every website is now scrapeable. Every database is now accessible. Every manual research task just became automatable.
The auditability piece is huge here. You can see exactly what went wrong (or right) instead of playing AI detective at 2 AM.
Feature #2: Image Generation That Doesn't Suck (but might bite)
Let's be honest, Clay's old DALL-E integration was... questionable. Nightmare fuel, really.
The new image generation through the core "Use AI" action? Big Improvement.
What we saw: Yash created consistent, brand-aligned images using:
Company hex codes extracted from logos
Dynamic prompting with multiple data sources
Multiple model options (GPT-4O, Gemini Imagen, etc.)
The use cases that got us excited:
Recruitment: Creating images of candidates holding company swag
Marketing assets: Generating visual content at scale for blog posts and documentation
Ad personalization: Brand-specific creative that would have taken weeks
Sure, Yash showing off cats biting well known logos (said cat produced an extra paw when biting down on the Google logo but seemed fine with all the others……Google, what say you?) might seem gimmicky, but imagine creating 500 personalized ad variations in an hour and feeling confident the images won’t embarrass you.
Feature #3: Document Upload - The Context Revolution
This one's still in development, but we got a sneak peek as Yash fussed over the normal kinks of any new feature.
The concept: Upload PDFs directly as context for your AI prompts. Think instant few-shot learning without the copy-paste nightmare.
Why our community should care: Remember Leo's comment about pain-based segments? This is exactly how you'd implement that at scale.
Imagine uploading all your best-performing email templates, case studies, or PVP examples, then having Clay generate new content that actually matches your voice and converts like your best stuff.
No more "go ask your BDRs for their best emails, then manually format them into prompts, then hope it works."
The MCP Server Bombshell
Buried in the livestream was the announcement we've been waiting for: Clay is building MCP server integration.
What this means in English: Soon, Claygent will be able to connect directly to:
HubSpot
Salesforce
Notion
Gmail
Google Drive
Boom.
And, buh bye business intelligence platforms (it’s just a matter of time Yo). Instead of copy-pasting prompts into ChatGPT company by company, you'll run strategic queries across your entire CRM at scale.
"What are the five most recent customers we closed in the software industry? Who are the most similar customers to X company I'm talking to right now?"
All automated. All in Clay. All at the speed of thought.
The Strategic Shift: From Middleware to Data Engine
Here's what really caught our attention in the conversation with Yash: Clay isn't trying to cross the chasm by training 200 people per company.
Instead, they're betting on the RevOps Trojan Horse strategy.
Train 2-3 talented people who can achieve 100x results across the entire organization. This isn't just about adoption, it's about creating a new category of professional: the GTM Engineer. The Cannonball GTM community is chock-a-block with them but they’re a rarity out in the wild, and Clay is betting on them to bridge from the early adopter segment into the ‘what color is your Lambo and can I borrow some fuel for my jet’ windfall that will come when they crush the early majority segment.
And Yash should know, he literally created this role when Clay was still figuring out enterprise sales. As he put it: "Clay is basically programming. You need to know the same programming concepts... but you don't have to know how to use all the different functions, packages, NumPy libraries."
Coming from someone who taught high school computer science before revolutionizing go-to-market at one of the fastest-growing SaaS companies, that's not theory, that's pattern recognition.
What This Means for Cannonball GTM Users
These features directly accelerate our core methodology:
Focus: Navigator makes competitive intelligence and market research trivial Investigate: Document upload lets you apply proven frameworks at scale
Narrate: Image generation creates compelling visual PVPs in minutes
But here's the bigger picture: We're watching the birth of the data engine that will power the next generation of go-to-market.
The Bottom Line
Navigator eliminates research bottlenecks. Image generation scales creative production. Document upload turns your best practices into systematic advantages.
And MCP servers? That's the bridge from innovator/early adopter to early majority.
What feature has you most excited? Drop a comment - we're already planning our next Clay builds around these capabilities.
Share this post