Design Partnerships in Vertical AI

Vertical Collective Roundtable: Q3 2025

Jul 15, 2025

Abstract
1. Finding Initial Design Partners
2. Determining Initial ICP / Beachhead
3. Iterating on Budgets, Pricing, and ACV
4. Selecting Wedge Products to Test with Partners
5. Planning Ahead for Defensibility

Join the discussion on LinkedIn

We launched the Vertical Collective to connect the best founders & operators in Vertical AI. An invite-only organization, it serves ~150 members today, from exited CEOs to growth-stage VPs to pre-idea founders.

Each quarter, the group hosts a Roundtable bringing together ten or so members and friends for candid, closed-door conversations on key challenges and opportunities in Vertical AI. Today we summarize our last get-together in held in July in San Francisco. Founders and operators in attendance represented verticals including healthcare, veterinary, dentistry, retail, construction, real estate, auto, and insurance.

While we discussed a range of topics, a central one was best practices in design partnerships. Below, we pull out key learnings, obstacles, and bits of advice to help Vertical AI founders navigate that critical first phase of commercialization.

1. Finding Initial Design Partners

Participants highlighted effective strategies for securing early design partners. One founder emphasized a curiosity-driven approach: "You can almost play a little dumb and say, 'Hey, I'm really interested in learning about what you guys do—what’s great, what sucks?' Letting them talk is an easy way to build relationships." Especially in less competitive, legacy industries operators stressed it only took a bit of credibility to get targets to open up: "The thing that always worked for us was just asking people for advice. You'd be shocked how many people would just be like, 'Sure.' The advice thing is huge." Others affirmed traditional methods such as direct outreach remain highly effective: "Emailing, calling people just still f*ing works." Most felt there was not really an unlock and that beyond leveraging network, cold outreach and introductions were the best form of discovery.

Participants emphasized the importance of consistency: both in terms of air-time with design partners (or potential ones), but also in terms of responsiveness to their asks. "It's really easy to go back and pilot with people because if you go out on a first call and have a really bad demo, you can say, 'This is what we're thinking about right now. Can I show you something again in a week?' Because they don't feel like you're trying to sell to them, they're happy to give input. For us, it converted two execs into advisors."

We also found consensus around a mix of cold and warm relationships, in terms of design partnership cohort make-up. Some cold relationships are important to demonstrate market pull beyond personal familiarity. Warm relationships, of course, are the best place to start from Day Zero—especially with bigger brands. A product leader with experience from multiple growth-stage vertical SaaS players reflected on the importance of leveraging your whole team: "It's very valuable to hire people from the customers that you actually want to sell to...They give you insights from having done the job you're building for." She would create informal “customer councils” of employees from industry and advisors as a first-blush sounding board, prior to taking new ideas to actual design partners / customers.

One participant also stressed a balance between enough diversity to be representative of your market but enough similarity that the initial wedge would convert (and that you aren’t pulled in 50 directions). Some prospects are best to come back to. The group consensus was that the right number of design partners at inception stages typically falls in the 5-10 range—with variation mostly dependent on depth of co-build, target ACV, and complexity of wedge product.

2. Determining Initial ICP / Beachhead

Refining the initial Ideal Customer Profile (ICP) or beachhead market was stressed as crucial for early success. Nic suggested a heuristic: "One rough way to gauge your beachhead ICP is a gut-check on ability to get double-digit market share within two years... if that seems impossible, consider narrowing your beachhead ICP." Another advised being highly selective with early partners: "Think through brands people really respect within your space. We've tried to be really picky with who we work with." These are your first potential referrers, he underscored, meaning some cachet matters—for example, in construction Turner and Skanska are going to turn heads with other contractors. Or, if you have a regional “lane” strategy, look for the names everyone knows.

The group agreed, however, on the importance of ensuring diversity in your design partnerships. Beyond making sure it’s a representative body, you want to avoid selection bias toward extreme early-adopters only. Make sure you’ve had enough customer discovery calls (10-30x your number of target design partners) to ensure your uptake isn’t isolated to cases of “AI tire-kicking.” Overall, narrow, well-defined ICPs were recommended for achieving meaningful early traction.

3. Iterating on Budgets, Pricing, and ACV

Distinct strategies for low versus high Average Contract Value (ACV) businesses were discussed in detail. For enterprise-focused solutions, participants favored deeper integration and customization: "If you're truly enterprise, it makes sense to have longer design partnerships and custom features initially." Conversely, SMB-focused products require rapid scalability: "Your initial growth for SMB is, 'How can we build a wedge that catches fire quickly?' You can't afford customization." Recognize that enterprise design partnerships likely take longer to navigate product complexities or integrations—whereas SMB design partnerships may be quick but more ongoing, serving as the framework for a rolling pilot (or “proof of value”) process to demonstrate AI ROI.

Participants emphasized the dynamic nature of pricing in the AI era, with considerations around: knowing which budget you’re accessing, tapping into services / BPO spend, and leveraging usage-based pricing. One founder explained the importance of knowing what type of spend / cost a buyer compares your product to: "Traditional firms already outsource to offshore. Now they buy employees overseas a $20 AI subscription." Another participant warned about assuming static pricing: "The mistake is assuming service prices remain the same. Eventually [with AI], costs for the whole market come down, and there's always someone willing to accept lower margins." Usage-based pricing was cited as an easy way to get into market especially for SMB buyers, whereas larger enterprises often sought more visibility on cost (though were still interested in value-aligned pricing vs. tradition flat subscriptions). The consensus was that AI-driven services businesses must rapidly iterate pricing strategies, anticipating continual margin pressure as AI becomes more accessible.

4. Selecting Wedge Products to Test with Partners

Evolving wedge products into systems of record generated extensive debate. One founder noted an indirect path to becoming integral: "We work with existing systems, extract and enrich their data, and become the primary querying interface. Without them knowing, we become the wrapper." Another emphasized strategic alignment of the wedge product as vital: "Hopefully your wedge has low friction, but if it doesn’t naturally tie to the next step in your workflow, it’s irrelevant." Ensuring a wedge product naturally progresses toward becoming a system of record is key for sustained adoption.

The group debated the potential and limitations of voice-based AI solutions as design partnership wedges. One participant voiced skepticism around defensibility: "Voice AI is difficult for durable businesses because in a world where anyone can spin up a solution, why should they still use yours?" Another participant pointed out rapid commoditization: "Companies providing call-center automation can quickly lose competitive advantage as general AI platforms easily replicate their capabilities." While voice AI offers appealing initial value propositions such as cost savings or efficiency, founders agreed it requires significant additional layers of proprietary integrations or data to become defensible.

5. Planning Ahead for Defensibility

Founders stressed the importance of building defensibility through scale, proprietary data, and strategic partnerships. One participant clearly articulated the necessity of a having a line-of-sight to economies of scale even with wedge products: "Great companies have relationships, unique data… or exclusive discounts. Even if someone else can do it, customers still prefer cheaper or better options." Another warned about commoditization risks in rapidly evolving fields: "Five years ago, scribing was hard. Now, anyone feeding transcripts into ChatGPT claims they can do it, eroding defensibility."

AI reduces the cost of delivering automatable services. But that’s true for anyone. Meaning in a competitive space with no other defensibility, there will also be someone else willing to take a point less—so that meaty 10% margin AI opened up may not last. Surplus will accrue to the buyer as providers undercut one another, engaging in a race to the bottom (or equilibrium) on pricing. Accordingly, the “market rate” for these services—what buyers are willing to pay—will come down. Founder shouldn’t assume that they will be able to command current-market pricing on a service that is clearly within the zone of competency for LLMs… without another source of defensibility, that is. Thus, founders should explicitly plan for durable advantages: e.g. industry-specific workflow, difficult integrations, incumbent partnerships, distribution network effects, key data ownership, regulatory moats, or (barring all else) first-mover economies of scale.

For Future Exploration

Shifting Control from Data to Capital Flows: Participants touched briefly on how startups might evolve from simply controlling data flows toward controlling transactions and capital. Exploring how vertical AI can capture and monetize financial or transactional data directly could be fruitful.
Impact of Regulatory Barriers on AI Adoption: Regulatory compliance (e.g., healthcare, insurance, finance) repeatedly arose as a challenge. A deeper discussion about proactively leveraging regulatory complexity as a source of defensibility rather than a barrier could yield valuable insights.
Long-Term Pricing Strategies for AI-Driven Services: With rapid commoditization of AI services, pricing sustainability emerged as a critical topic. Further exploring innovative, sustainable pricing strategies—especially value-based models—could benefit founders navigating shrinking margins.
Methods to Build Internal Expertise: Unique strategies to built “voice of customer” resources to power vertical discovery: internal subject matter experts, advisory panels, equity-incentivized strategics, “sales engineer” type G2M hires from industry—to supplement traditional, outbound customer discovery. Vertical AI sales teams should explore strategies to institutionalize and disseminate industry learnings, conversational best practices, important vertical terminology etc. to accelerate ramp for new product and G2M hires.

Thanks for reading Euclid Insights! If you know a founder who is thinking through Vertical AI design partnerships, we’d love to help. Just reach out via LinkedIn, email, or here on Substack (via comments or the DM button below).