PropTech · B2B2C · Founder-Led · AI Product · WIP

Vesta

Built from 10 years inside the problem — an AI deal intelligence platform for residential real estate agents.

Founder & PM — Discovery, Strategy, Validation Design

Product

Vesta

COMPASS LANDING PAGE

vesta-compass

Type

B2B2C PropTech SaaS

Wedge Feature

Compass — AI offer strategy

In Queue

Edge — negotiation intelligence (in discovery)

Methods

JTBD, Value Prop Canvas, Business Model Canvas, Opportunity Hypothesis, Reference Customer Program, Landing Page Validation

Status

WIP — Active, In Validation

The brief

Most AI real estate tools are built for the listing side. Compass is built for the moment that actually breaks a buyer's agent.

I've been in residential real estate since 2016. I founded Magnetic Residential, ran it for seven years, and closed more than 200 transactions. That background means I didn't need market research to find this problem — I lived it from both sides.

Here's what I watched happen hundreds of times: a buyer's agent in years two or three of their career walks their client into a competitive multiple-offer situation. The offer window is closing. They don't have a mentor free to call. They don't have data on what comparable offers looked like. They have their gut, a client who's anxious and trusting them completely, and a blank offer form.

Most of the time, they guess.

Sometimes they overbid because the client can't take another loss. Sometimes they underbid because they're afraid to push. And more often than not, they lose — not because their client wasn't competitive, but because the agent had no playbook for that specific situation.

Vesta is an AI deal intelligence platform built to close that gap. The wedge is Compass: AI-powered offer strategy and negotiation coaching for buyer's agents in multi-offer situations. Sold as a seat-based SaaS through brokerages, not directly to agents.

The opportunity hypothesis

I believe that buyer's agents in years one through four at mid-market brokerages — working without senior mentorship or a data-backed playbook — are losing competitive offers not because their clients aren't qualified, but because the agent has no structured framework for offer strategy in multi-offer situations.

I can address this by building Compass: an AI tool that delivers real-time, data-backed offer strategy and negotiation coaching calibrated to the specific property, market conditions, and competitive context — embedded inside Vesta and sold to brokerages as a seat-based team tool.

I will know this is working when the Compass landing page converts 15% or more of visitors to waitlist signups, indicating the pain is acute enough to motivate action. Conversion below 5% is a signal to reassess whether this pain point is real or whether I'm solving for the wrong moment.

The users

Two personas. One product motion.

Primary Persona

Marcus Rivera

Year 3 Buyer's Agent · 6–10 deals per year · Sub-cap

Marcus is three years in and hustling. He's past the point of needing hand-holding on the basics, but he's not far enough along to have a real playbook for the hard situations. He closes six to ten transactions a year — enough to stay licensed and generate some income, not enough to hit his cap or build the referral flywheel he needs to grow.

Multi-offer situations are where Marcus bleeds clients. He's lost three buyers in the last year because his offers came in too low or too late. He knows the data exists somewhere — comparable offers, days on market, list-to-sale ratios — but he doesn't have the time or the tools to synthesize it into an offer recommendation under pressure. When it matters most, Marcus is guessing.

"I know what I should do in theory. I just don't know what the right number is right now, for this house, in this market."

Goals

Win more competitive offers. Build client trust. Generate referrals. Hit his cap.

Fears

Losing a client's trust on a winnable deal. Being outmaneuvered by a more experienced agent. Having no explanation for why his offer lost.

Secondary Persona

Diana Chen

Year 7 Dual-Side Agent · 22–30 deals per year · Above-cap · Team Coordinator

Diana is past the early-career scramble. She's above cap, runs both sides of most transactions, and has a team coordinator handling her admin. She doesn't need help on offer strategy — she's built her own playbook through repetition.

What Diana represents for Vesta is a different value signal: brokerage-level proof. Brokerages recruit agents like Diana. They retain them with tools and culture. If Diana adopts Vesta and talks about it with her team, it becomes a recruiting argument for the brokerage. She's the word-of-mouth engine inside the Reference Customer Program.

"If this makes my newer agents better without me having to babysit every offer, it's worth it."

Jobs to be done

Anchored to outcomes, not features.

The JTBD canvas anchors Vesta's design to the actual outcomes Marcus is trying to achieve — not just the features he might click.

Functional job

When I'm preparing a competitive offer in a multi-offer situation, help me build a data-backed offer strategy that gives my client the best shot at winning without overpaying.

Emotional job

Make me feel like a professional who knows what they're doing — not someone who's winging it under pressure in front of a client who's counting on me.

Social job

Help me be the agent my clients talk about because I had a plan, not the one they left because I didn't.

Secondary functional job

Give me a repeatable framework I can use across different markets, price points, and competitive conditions — so I stop building strategy from scratch every time.

Value proposition

What Compass relieves and what it creates.

Pains Compass addresses

  • No data-backed framework for pricing in multi-offer situations — agents rely on instinct and availability of experienced colleagues who are often unavailable
  • Generic market reports that don't speak to the specific competitive context of a live offer
  • No visibility into what comparable offers looked like — only final sale prices, which come too late
  • Lost clients and lost referrals downstream of losing winnable offers
  • The confidence gap: knowing in theory what to do but not what number to write on the form

Gains Compass creates

  • Data-backed offer strategy calibrated to the specific property and market moment — not general advice
  • Faster offer preparation when the clock is running
  • Confidence in client-facing conversations
  • Institutional knowledge that doesn't depend on a senior agent being available
  • A repeatable playbook that compounds across deals

Positioning statement

Compass gives buyer's agents a data-backed offer strategy and real-time negotiation coaching for multi-offer situations — so they stop guessing at the moment that costs them clients.

Business model & market sizing

Brokerages buy it. Agents use it.

Vesta is a B2B2C platform. Brokerages buy it. Agents use it. That distinction determines the sales motion, the pricing architecture, and the success metrics at every stage.

Why B2B, not direct-to-agent: Buyer's agents at mid-market brokerages don't evaluate and budget for their own tools the way enterprise professionals do. Brokerages do. Selling to the brokerage gets Vesta in front of 50 to 200 agents in a single contract rather than fighting for individual adoption one agent at a time.

The pricing model: Seat-based SaaS, sold to brokerages as a team tool. Seat counts scale naturally as brokerages grow.

The expansion arc: Compass is the wedge. The full Vesta platform is deal intelligence across the entire transaction lifecycle. Brokerages that land on Compass expand into the broader suite as trust builds.

LevelAmountDefinition
TAM$1.35BAll U.S. residential real estate agents paying for productivity tools
SAM$405MAgents at mid-market and independent brokerages with 50–500 agent teams
SOM$22.5M ARR25,000 seats at target pricing by end of year three

The SOM assumption is grounded in a bottom-up model: penetrating roughly 500 mid-market brokerages at an average of 50 seats per brokerage within three years — 3% of the addressable market.

Reference customer program

Six brokerages. Named references. Compounding proof.

Vesta's go-to-market is built around Marty Cagan's Reference Customer Program. The goal isn't to sell to everyone — it's to find six brokerages who will pilot the product, succeed on explicit metrics, and become the proof that makes every subsequent sale easier.

6

Brokerages

50–200

Agents per brokerage

12-month

Pilot

MetricTarget
Activation60% of licensed seats active within 45 days
Retention65% of activated agents use Compass on 2+ deals within 90 days
Renewal5 of 6 brokerages renew at end of pilot
Named reference4 of 6 brokerage principals willing to be a named reference

The reference customer program serves two functions simultaneously. It validates that Compass delivers real retention value at the agent level. And it builds the social proof infrastructure the enterprise sales motion needs to close the next 50 brokerages without a six-month decision cycle. Brokerages don't buy tools based on feature lists. They buy based on what brokerage principals they trust are already using.

Current stage — Compass validation

Live in landing page validation at vestahq.ai.

Compass is live in landing page validation mode at vestahq.ai. The landing page targets Marcus directly: buyer's agents in years one through four at mid-market brokerages. The copy speaks to the specific moment — the multi-offer situation, the offer window closing, the absence of a data-backed playbook.

The validation hypothesis

Buyer's agents in years one through four at mid-market brokerages without mentorship feel acute enough pain around multi-offer offer strategy that they will take action — joining a waitlist — when presented with a credible solution.

SignalConversion rate
Strong — pain is acute, messaging resonates, proceed to build15–20%+
Weak — investigate copy, targeting, or problem framing5–15%
Negative — pain is not acute enough or wrong momentBelow 5%

The threshold logic matters. Above 20% says the problem is real, the persona is right, and the messaging is landing. Below 5% doesn't mean the problem doesn't exist — it means either the pain isn't acute enough to motivate action right now, or I've described the solution in a way that doesn't match how the user frames their own problem. Both are information. The validation is live and being monitored against these thresholds.

What I'm learning

Insights from ten years in the problem — and a live validation loop.

The domain advantage is real — but it has to be earned, not assumed.

Ten years of real estate experience means I walked into Vesta with a strong prior about what buyer's agents need. That prior accelerated discovery significantly. But a prior isn't a substitute for validation. The hypothesis is grounded in direct experience with 200+ transactions — and it still has to be tested against behavior, not just recognition.

B2B2C creates a sequencing problem that has to be solved up front.

The brokerage buys it. The agent uses it. Those two customers have different success criteria, different anxieties, and a different definition of value. Compass was designed to satisfy the agent's job-to-be-done first — because if agents don't activate and use it, the brokerage renewal doesn't happen. The Reference Customer Program is built around that sequencing logic: prove agent-level value, then collect brokerage-level proof.

AI in a high-stakes transaction is a trust problem before it's a capability problem.

A buyer's agent using Compass to structure an offer is making a recommendation that could cost their client tens of thousands of dollars. That means the AI output has to be legible, not just accurate. If the system says "offer at $847,000" without explaining why, the agent can't defend it to their client — and they won't use it again. Compass is being designed around explainability from the start: not just what to offer, but what signals drove the recommendation, what the confidence range is, and what changes if the market context shifts.

Vesta is an active project. This case study will be updated as validation data comes in.

Interested in the full discovery artifacts? Request them here →