Founders on Hiring in the AI Era
And other founder-founder advice from our Q3 Vertical Roundtable
Abstract
Learnings from Euclid’s Q3 Vertical AI Roundtable in NYC, including:
1. How AI-Era Hiring is Different
2. When Forward Deployment Works
3. Keep the Early Team Hyper-Focused
4. How Product Teams View Commoditization
5. People on Cap Table Matter Too
6. AI Customer Success
7. Post-LLM Pricing
8. AI FatigueWe held our latest quarterly Vertical Collective Roundtable a few weeks ago in NYC. From multi-exit growth-stage CEOs to idea-stage founders—from real estate, to legal, to banks, to healthcare—the group represented the full spectrum of Vertical AI perspectives.
In today’s roundup, we compile key learnings from our Roundtable discussion. We find it’s a great real-time barometer for the questions Vertical AI operators are grappling with today—and solutions that might work for your startup.
Top 3 Quotes
On the value of proclivity to experiment: “We actually are devaluing expertise in our interviews and we’re basically trying to find people that are more experimentative.”
On the value of forward-deployment: “That's actually where [our FDE] motion originally came from… it was essentially a GTM compression to get to usage, which then leads to lock in, which then leads to pricing power.”
On the mindset of embracing creative destruction: “[Many] think that race to the bottom is a bad thing… we think it’s the opposite… the real unlock is [creating new value for those customers & taking advantage of the vacuum that ensues.]”
Roundtable Learnings
1) How AI-Era Hiring is Different
Why it matters: Internal AI edge emanates from pace of experimentation and cross-functional usage.
What we heard
Changing backgrounds: De-emphasize fixed credentials. Hire the experimenters. Many teams see outsized experimentation from AI-native hires.
New critical skills: Prompt quality drives output quality. Testing tool use cross-domain often needs to be a dedicated responsibility.
Look outside big tech: Engineers from big tech can be dismissive of emerging AI tools, thinking they can do it better and used to ingrained practices.
New goals / tactics: weekly AI experiments (5-day limit, must deploy with measurable KPIs). Function-specific AI Slack channels. Goal to maintain ~2:1 revenue/employee.
Quotes from the room
“We actually are devaluing expertise in our interviews and we’re basically trying to find people that are more experimentative.”
“We have certain engineers, especially those that came from like the big corps, who were like, nah, I can do this way better [and were slow to adopt AI].”
“The bulk of value [in AI experimentation] is driven by people who are sub-26.”
2) When Forward Deployment Works
Why it matters: FDE (forward-deployed engineering) can win implementation speed and workflow “ownership,” but only if the unit economics hold. Product-Led Growth (PLG) shines when ACV is constrained.
What we heard
Pros of FDE: faster implementation, deeper relationships, tighter workflow control. May even be seen as a cost of implementation for enterprise.
Cons: scaling org design is hard; being embedded changes dynamics.
Rule: FDE services should be break-even, not a profit center.
Alternative: At lower ACVs, solve complexity in product (not services). Enterprise tiers can layer services (e.g., “Oracle engineers” mode).
Quotes from the room
“One of the other benefits of forward deploy actually is just about crunching your implementation time down as founders in the early stage… you’re probably just doing it yourself.”
“That’s actually where [our FDE] motion originally came from… it was essentially a GTM compression to get to usage, which then leads to lock in, which then leads to pricing power.”
3) Keep the Early Team Hyper-Focused
Why it matters: Sequencing second (and third) verticals is mostly about portability (tech + GTM expertise), bandwidth, and—ultimately—founder creativity.
What we heard
Stay focused: Stick to your beachhead until you can port core tech and repeat workflows where you have expertise / advantages.
Path of least resistance: Recognize that horizontal expansion is fundamentally hard. Selling new stuff into the same audience is typically way, way easier than selling the same stuff into fundamentally new audiences.
Overfitting risk: Avoid early overfitting in new verticals, just as you would with a net-new startup: you need 30–40+ real conversations/design partners before you can begin trust the signal.
Expansion timing: Consider market size, internal bandwidth, and fundraising / investor messaging. Don’t ignore investors (important signal) but external pressures shouldn’t derail the roadmap.
Creativity: One founder felt strongly that continual market expansion is a function first and foremost of founder vision & creativity.
Quotes from the room
[My company was] “operating for, I want to say six and a half years in the [first] vertical until it picked the second vertical to kind of go into.”
“Are there specific features that are relevant to other segments of our market or is this just a kind of ‘we're to shift and focus on this customer base over here now’?”
4) How Product Teams View Commoditization
Why it matters: Competing on per-task pricing is a race to the bottom. Durable advantage comes from owning workflows, capturing compounding data, and inventing better business models, not cents-per-call.
What we heard
Embrace commoditization: “destroy” existing pricing to reveal workflow value.
Center a pain-point: The “raise big and find PMF later” strategy (e.g., Harvey) is high-risk versus bottom-up workflow mastery.
Don’t reinvent the wheel: Invest to build proprietary workflow datasets but avoid building your own infrastructure unless it’s truly net-new.
Quotes from the room
“There's going to be another company that says no, we're going to do that work for $1, right. And it becomes this race to the bottom to where like you, you kind of lose leverage.”
“[Many] think of that race to the bottom is a bad thing… we think it's the opposite… real unlock is like how you take the commoditized task with AI that's happening today and then leverage that data to harness… new value creation.”
5) People on Cap Table Matter Too
Why it matters: The “big brands” aren’t what they used to be. In services-heavy and trust-sensitive markets, raises have value in signaling longevity.
What we heard
Longevity: Buyers in trust-sensitive verticals often need confidence you’ll be around for many more years; a Series A / B can go a long way in this regard.
Brand Equity: Potential to leverage strategics to drive trust with buyers.
VC Selection: Large venture brands vary widely in their actual value. Make sure an investor truly understands your business and incentives are aligned.
Quotes from the room
“Why did we raise money? Every raise was predicated around the idea of brand equity creation.”
“We took money from [redacted strategic] in this latest round specifically because we want to basically take advantage of this entire market that they serve.”
6) AI Customer Success
Why it matters: Prevent overfitting by diversifying input but ensure you are able to go deep. Earn trust by documenting the customer’s world and saying “no” with principle.
What we heard
Sample size: Multiple design partners across the spectrum allow you to ensure your product roadmap isn’t over-pivoted to one player. One-to-many partners (agencies / consultants / distributors) can help widen the lens.
Relationship craft: Never “correct” customers—they live in their business—but it’s OK to say no. Iterate wireframes / prototypes weekly to make design partners feel heard.
Saying no: Exhaust workaround options before building. Be willing to lose customers over product roadmap.
Quotes from the room
“Only so many customers will let you go deep and get everything you need.”
“We have a chatbot we built that essentially allows our customer success managers to talk to data so they stop bothering our data scientists for dumbass queries that they don't understand how to do in Looker.”
“Try to solve it every possible way other than building it first.”
7) Post-LLM Pricing
Why it matters: Use the familiar pricing modalities your market already understands. Outcome-based models likely require proving outsized ROI first.
What we heard
Pricing: Don’t invent new pricing. Often best to start with industry-standard models aligned with how the customers already buy or make revenue.
Long-term: Consider giving away today’s product to monetize the proprietary workflow data it collects.
Models: Important to consider in your pricing. Constantly re-evaluate frontier models (OpenAI, 11 Labs, etc.) and watch cost-performance.
Quotes from the room
“Figure out the business model [before deciding] how to monetize.”
“It's really about not inventing a new monetization model.”
8) AI Fatigue
Why it matters: Many buyers have AI fatigue. Lead with outcomes, not “AI” (even if your pricing isn’t outcome-based). Distinguish yourself with UI/UX. Seek compounding “superset data.”
What we heard
AI messaging: In some legacy verticals, actively avoid “AI” branding. Vertical customers care about results. Many have been burned by half-baked AI tools and are hounded by growing vendors.
UI/UX: Differentiation matters—too many tools ship the same LLM-generated interfaces.
Datasets: Partner with incumbents and pursue integrations, but be sure your product creates data they don’t have (“superset data”).
Implementation as CS: In many verticals, you’ll know the customer’s IT systems better than they do. Produce schemas / documentation of their systems and share with customer IT teams. They don’t have time for it (and limited expertise)—use them as leave-behinds to build brand & trust.
Quotes from the room
“For every integration, I always make sure my product generates more rows out than it takes in.”
“UI differentiation, UX differentiation, early stage. Like, this is kind of what I'm gambling on right now.”
Thanks for reading Euclid Insights! Euclid is a VC partnering with Vertical AI founders at inception. If anyone in your network is working on an idea in the space, we’d love to be helpful. Just drop us a line via DM or in the comments below.