Case Study: How a Proptech Startup Recruited Its Founding Real-Estate Agencies with Hyper-Local LinkedIn Outreach

Julian Bagilet
Founder
The Company: An AI-First Proptech Taking On Entrenched Incumbents
AgonProp is a proptech platform built for the Argentine real-estate market. Where legacy portals are essentially classified-ad websites with a search box, AgonProp was designed from day one around AI: agencies can create a complete property listing from a handful of photos, syndicate it to multiple portals from a single dashboard, get automated property valuations (since productized as a standalone self-serve appraisal tool), and manage inbound buyers through a CRM that scores and prioritizes leads automatically. Buyers, on their side, can search in plain language — "two-bedroom apartment near Parque Sarmiento under 120k" — instead of wrestling with filter checkboxes.
It is a classic vertical SaaS story: a focused product, a clearly underserved user (the small and mid-sized real-estate agency), and a market where the dominant players have had little pressure to innovate for a decade. If you have ever built for a niche industry, you already know the shape of the problem that comes next.
Because none of it matters until real agencies are running real listings through the platform.
The Challenge: Founding Users in a Market Owned by Incumbents
Every marketplace and every vertical SaaS faces the same cold-start wall, but proptech in Latin America has a particularly brutal version of it. The incumbent portals are household names. Agencies have used them for fifteen years. Their workflows, their habits, and even their vocabulary are shaped by those tools. A new platform — even a meaningfully better one — starts from zero trust.
The AgonProp team knew that the first ten or twenty agencies would matter more than the next thousand. Founding agencies do three things no ad campaign can buy:
- They validate the workflow with real inventory. A property platform with no properties is a demo, not a product. Founding agencies bring the listings that make the platform real.
- They shape the roadmap while it is still cheap to change. The feedback of an agency owner uploading their actual portfolio is worth fifty survey responses.
- They become the social proof for everyone who comes after. In a trust-driven, relationship-driven industry like real estate, "the agency across the street already uses it" is the single most powerful sales argument that exists.
The question was how to reach them. Paid ads were the obvious option and the wrong one. Agency owners in Argentina are not scrolling Facebook waiting for a SaaS banner, and a startup burning runway on broad-targeting ads to reach a few hundred specific people in one city is lighting money on fire. The target population was small, identifiable, and professional. That is not an ads problem. That is an outreach problem.
Why 1:1 Outreach Beats Ads at the Founding-User Stage
There is a stage-fit argument here that applies to any vertical SaaS, not just proptech. At the founding-user stage you do not need volume — you need conversations. You are not optimizing cost-per-click; you are optimizing for the quality of the first twenty relationships your company will ever have.
LinkedIn outreach fits that stage almost perfectly:
- The targeting is surgical. "Owners and brokers of real-estate agencies in one specific city" is a search you can actually run on LinkedIn. No ad platform gives you that precision at that scale.
- The message arrives as a person, not a brand. A inmail-vs-connection-request">connection request from a founder reads completely differently than a sponsored post. At this stage, the founder's face is the brand.
- Every reply is a discovery call waiting to happen. Ads generate clicks; outreach generates conversations. Founding users come from conversations.
So AgonProp used nuph.ai to run the campaign that would recruit its founding agencies.
Building the Hyper-Local ICP: Agency Owners in Córdoba
The first strategic decision was geographic focus, and it is the one most founders get wrong. The temptation is always to go wide — "real-estate agencies in Argentina" sounds like a bigger opportunity than "real-estate agencies in Córdoba." It is not. It is a shallower one.
AgonProp picked Córdoba, Argentina's second-largest city, and concentrated the entire founding-user effort there. The campaign was literally named after its ICP: "Dueños de Inmobiliarias en Córdoba" — owners of real-estate agencies in Córdoba. The targeting logic:
- Role: owners and brokers, not employees. Adopting a new platform is a business decision. Agents can love a tool; only the owner can commit the agency to it. Targeting decision-makers directly removed an entire layer of internal selling.
- One city, deliberately. Density creates network effects that dispersion cannot. Twenty agencies in one city talk to each other, compete with each other, and notice each other's listings. Twenty agencies scattered across a country are twenty isolated accounts.
- Local signals in every profile. Because the pool was city-specific, every prospect shared context — the same neighborhoods, the same local market conditions, the same professional associations. That shared context is raw material for personalization that generic national campaigns simply do not have.
Hyper-local is not a limitation — it is a personalization asset. The narrower the geography, the more every prospect has in common, and the more genuinely relevant every message can be.
Connection Notes in Spanish, Personalized by AI
The second strategic decision was language and tone, and this is where most international outreach advice quietly fails outside the English-speaking world. Argentine business culture is warm and informal. A connection note written in stiff, formal Spanish — the "usted" register — reads like a bank notice. A note translated from an English template reads worse: everyone can smell it.
AgonProp configured its campaign to generate every connection note in natural Argentine Spanish, using the informal "te" register — the way one professional in Córdoba would actually talk to another. The AI personalization drew on each prospect's profile — their agency, their role, their market — so no two notes were the same, and each one led with the prospect's world rather than the product.
Three constraints shaped the writing, and each carries a lesson:
- LinkedIn's 300-character limit on connection notes is a feature, not a bug. You cannot pitch in 300 characters, which is exactly the point — the note's only job is to earn the connection and start a conversation. AgonProp's notes introduced the founder, referenced the prospect's agency or local market, and hinted at why a conversation might be worth their time. No feature lists. No links. (Also worth knowing: the limit applies to non-premium recipients and counts characters, not words — in Spanish, where words run long, every character matters even more.)
- Sender context makes AI personalization coherent. The campaign used nuph's sender-context feature, which feeds the AI a description of who is writing and why. That is the difference between generic flattery and a note that sounds like a real founder building something real in the prospect's own market.
- The CTA was a call, not a demo. Each conversation was steered toward a short call. In a relationship-driven industry, a fifteen-minute conversation converts founding users; a self-serve signup link does not.
What Happened
We are deliberately not going to publish a wall of vanity metrics here — founding-user campaigns are about a handful of high-quality outcomes, not spreadsheet-friendly percentages. But the qualitative results were exactly what the stage demanded:
- Early replies came from the right people — agency owners and brokers, not gatekeepers — because the targeting was built around decision-makers from the start.
- Replies turned into booked calls. The informal, local-language notes consistently opened real conversations rather than polite dead-ends, and the call-first CTA converted those conversations into calendar time with the founder.
- Calls turned into founding agencies exploring the AgonProp platform with their own listings — real inventory, real workflows, real feedback flowing into the product.
Just as important is what the campaign cost: no ad budget, no agency retainer, no SDR hire. One founder, one clearly defined ICP, and a system that handled the targeting and personalization at a scale one person could never write by hand.
Five Lessons for Vertical SaaS Founders
The AgonProp campaign is a repeatable pattern. If you are building vertical SaaS and hunting for founding users, here is the playbook, condensed:
- 1. Go narrower than feels comfortable. One city, one role, one industry. Density beats reach at the founding-user stage — you can always widen the funnel after the first cohort is in.
- 2. Write in the language your market actually speaks — including register and tone, not just vocabulary. Informal Spanish for Argentine SMB owners. Whatever your market's equivalent is, an AI writing natively in it will outperform a translated English template every single time.
- 3. Respect the 300-character limit by giving the note one job. Earn the connection. The pitch comes later, in the conversation the note made possible.
- 4. Target owners, not users, when adoption is a business decision. Selling a platform to an agency means convincing the person who owns the agency.
- 5. Make the CTA a conversation, not a conversion. Founding users are recruited, not funneled. A short call with a founder closes what a landing page cannot.
The Bottom Line
Founding users are the hardest customers you will ever win and the most valuable. AgonProp's campaign shows the stage-appropriate way to win them: hyper-local targeting, native-language AI personalization within LinkedIn's constraints, and a founder showing up as a person — at a scale no founder could sustain manually.
If you are at the founding-user stage of your own vertical SaaS, you can run this exact playbook. nuph.ai's free plan includes 50 credits/month with full feature access — enough to test a hyper-local campaign in your market. Start free in under 5 minutes.
Related articles
How to 3x Your LinkedIn Reply Rates in 2026
The average LinkedIn outreach reply rate is under 5%. Here are proven strategies top sales teams use to consistently hit 15%+ reply rates.
Building Your Ideal Customer Profile: A Data-Driven Approach
Stop guessing who your best customers are. Learn how to use data from your CRM, LinkedIn, and AI to build an ICP that actually drives revenue.
Lead Scoring Best Practices for B2B SaaS Companies
A practical guide to building a lead scoring model that helps your sales team focus on the prospects most likely to convert.
Your next reply is one click away. Start free.
Free plan — 100 credits included, no credit card