Refty GTM Engine Case Study

01 The client

A luxury data product in the world's most competitive property market.

A non-essential product sold to 60,000 individual brokers, against cheaper limited competitors. The engine had to earn attention, not demand it.

Refty refty.ai · Dubai, UAE

Property intelligence for UAE real-estate brokers: hidden market data on hundreds of thousands of Dubai listings, verified unit-level records, live transaction comparables and price signals. Sold as a premium subscription to brokers and agencies.

CompanyRefty, real-estate market analytics, Dubai

ProductHidden market data, real-time updates on hundreds of thousands of listings

Deal size$5-50k per 6 months

Market~60,000 individual brokers in Dubai

CompetitionCheaper tools with limited functionality

Engagement8 months, semi-automated engine plus a human SDR bench

02 The problem

Ads stopped converting. The founder became the bottleneck.

Inbound from paid ads stopped paying back: competition drove conversion to deal down.

The founder personally ran the outreach team, on top of strategy, product, marketing, PR and enterprise sales.

5 SDRs worked manually, with no system deciding who to contact, when, and about what.

60,000 brokers sounds like a big market. It is not. Blast it once with templates and it is gone.

The engine had to be useful to a broker before asking for anything. So we made it a free assistant first, and a sales channel second.

03 What we installed

A semi-automated engine: machines find the moment, humans make it warm.

Deliberately semi-automated. Personalized videos, voice notes and friendly WhatsApp threads cannot be faked by full automation, and brokers can smell it when they are.

Signals as a free broker concierge

Daily enrichment from open data feeds fresh listings, emails and WhatsApp numbers into the engine. Each broker gets notifications that read like a personal assistant:

Competitor appeared on your property → alert
Price changed on a rival listing → alert
New listing in your building → alert
Every alert: market intel + a "book a demo" button

15-minute signal-to-human SLA

Demo request → HubSpot → Telegram alert to the SDR on duty, 15 minutes to pick up. At shift start every SDR finds leads auto-assigned and sorted by signal temperature, hottest first.

All of WhatsApp inside the CRM

Every WhatsApp conversation flows into HubSpot, a famously painful integration, done properly. CRM fields enrich themselves, so an SDR knows in seconds who they are talking to: solo broker sharing a subscription with friends, or an agent inside a 300-broker agency with brutal internal competition. Different mindset, different pitch, mapped in a pain matrix.

Broker tiering and live context

Brokers tiered by market experience, listing longevity and average property value. Top SDRs get top leads, newcomers still get hot ones to ramp fast. Every SDR sees the prospect's live listings and talks about the broker's actual deals, not about the product.

SDR onboarding as a product

New cohorts launched weekly: 2 days, 4 hours total. After training the engine auto-generates the account, CRM access and a pre-warmed WhatsApp number. Conversation starters and objection prompts appear directly in the CRM. That is how 5 SDRs became 25 without chaos.

Market listening, back to product

The engine collects feedback from replies and call transcripts and turns it into product recommendations the founder did not ask for. Meetings with the client stopped being status reports and became brainstorms.

04 Results

The founder got his company back. The engine kept running.

Metric	Before	After
SDR team	5	25 in 6 months
Meetings per month	~100	150-200
Reply rate	Not tracked	55-65% on signal-based touches
Reply to meeting	Not tracked	3-7%
SDR onboarding	Ad hoc	2 days, 4 hours, weekly cohorts, system-generated setup
Founder in outreach	Running it all himself	Stepped back within weeks

"I got very lucky with you."

Refty founder

"We were not an outreach team. We were a concierge for the broker: writing about their properties, their work, what they were already thinking about, or had not thought about yet."

How we would put it on the audit call

05 Installed, not rented

The proof: the engine survived a change of management.

A new leader inherited a running machine.After 8 months a new operations head joined and received a fully working pipeline: the engine, onboarding materials, processes and CRM.

It kept working after the handover.The team was slightly downsized. The engine did not notice. That is the difference between installed and rented.

Onboarding became an asset.Weekly cohorts with system-generated accounts and prompts mean the company can grow or replace SDRs without losing the playbook.

The relationship stayed warm.Meetings were brainstorms, not report reviews, because the system did the reporting by itself.

Who ran this. This engine was designed and operated by Sofia Altman, Spice GTM co-founder, in her operating role. Refty is named with permission; details and references are shared on the audit call.

60,000 brokers. One engine that feels like a concierge.