How Hospitable booked 36 meetings in a month with RepX

Alan Sincich, PMM at Hospitable, had already built out Hospitable's demo library and started using hubs on the website to drive more leads when he came across RepX. His next step was obvious: create his own RepX agent, Hospi, based on the demo library and other product and sales resources he already had access to before deploying the agent on Hospitable's website.

Training RepX on real conversations, not just documentation

Alan train Hospi on a range of resources, directly in RepX's training dashboard:

• Sales call transcripts, from every angle. Alan pulled transcripts from AEs talking to small property owners, AEs talking to larger property managers, outbound cold-call reps, and customer success calls. The goal was giving RepX a full field of view on how Hospitable actually talks to different kinds of buyers, not just one script.
• The full Help Center, ingested from a single link. RepX broke the content down by section and page automatically, no manual URL uploads required.
• A simplified pricing model. Hospitable's real pricing varies by setup and gets complex fast, so the team rebuilt it as a straightforward spreadsheet RepX could actually digest and reason from.
• The complete demo library, so RepX could pull up the right feature demo live inside a conversation instead of just describing it.
• A live feedback loop. A Zapier automation was set up to pipe every RepX conversation transcript into a shared Slack channel, with Hospitable's CEO, product team, and marketing team all watching. The team read through real conversations, flagged what RepX got wrong, and fed the corrections back in. This wasn't a one-time training pass. It's a habit Alan still runs.

Why RepX felt different from a standard chatbot

The biggest differentiator, in Alan's view, was RepX pulling up real product demos mid-conversation instead of just answering in text:

"RepX gave me, if I was an outside user, the feeling of going in there and being like, okay, I can actually see the product in action. That helps build trust, even though it sounds funny. Talking to another humanoid model and seeing it back up what it's saying with a real demo builds trust."

Getting the product in front of more visitors, earlier, is something Alan's team had been chasing since he joined Hospitable. RepX gave that effort a new surface to work on.

Alan also points to the conversational nature of interactions with RepX:

"RepX was very conversational. It's looking at you, you can literally speak to it and it talks back. It felt very natural. With other agents, it can feel a bit like you're talking to just another chatbot."

‍

The numbers, after a month of RepX

• 37 leads → 36 meetings booked. Almost every single lead that came through booked a meeting.
• 6 marketing-qualified leads and 8 active opportunities currently being worked.
• 7 closed-won deals, several with a "rapid rollout" upsell attached, Hospitable's paid onboarding add-on. RepX was selling that add-on in the conversation, not just the base subscription.

How RepX qualifies and gatekeeps for the sales team

Because the whole point was freeing AEs to focus on bigger accounts, Alan needed RepX to route the right leads to the right place without human triage. RepX asks visitors for their property count directly in conversation:

"Anything over 20 properties, it'll present a link to our enterprise team and ping them immediately. It gets thrown into the sales team's workflows in HubSpot. Hospi does a pretty good job of handling it on its own"

Anything under that threshold stays with RepX and Hospitable's SMB team. RepX also picks up on buying signals inside the conversation itself, like which features a visitor asks about, and flags likely bigger accounts to sales in real time even before they hit the property-count question:

How the rollout actually happened

RepX didn't launch site-wide on day one. Hospitable started narrow and expanded as confidence grew:

• Started as CTAs on two pages, pricing and home, while the team was still validating it.
• Moved to an embedded widget at the bottom of the page once they trusted it more.
• Added a 10-second delay before RepX appears, so visitors aren't met with a pop-up immediately and have a moment to start scrolling first.
• Playbooks are next, so RepX can ask more targeted opening questions per page instead of a generic greeting, surfacing intent faster on a pricing page versus a product page.
• HubSpot integration already shipped during the pilot, giving Alan a dashboard to actually see conversion data instead of digging through raw conversations.

What's next

• Continuously feed RepX new call transcripts, on an automated cadence rather than a manual one
• Ship the page-specific playbooks
• Roll out to mobile, which Alan said was days away as of the webinar
• Use account signals like property count and on-site behavior to recommend the right plan tier automatically, instead of defaulting everyone to one
• Offer promo codes to the accounts that fit best

What to steal from this

• Train on real conversations, not just documentation. Hospitable's biggest input wasn't the Help Center, it was actual sales call transcripts across AEs, cold outbound, and customer success, giving RepX range across different buyer types.
• Build a live feedback loop and actually use it. The Zapier-to-Slack pipeline meant the whole team, including the CEO, was reading real RepX conversations and correcting them on an ongoing basis.
• Let the agent gatekeep, don't force manual triage. A simple property-count question routed bigger accounts to sales instantly and kept AEs off small deals, exactly the outcome the initiative was built for.
• Different buyers want different paths, and that's fine. Some leads want a guided conversation before they'll try the product; others want to self-serve. RepX's data made that split visible instead of forcing one funnel.
• Roll out narrow, then expand. Two CTAs, then an embed, then a delay tweak, then playbooks. Confidence was built incrementally, not all at once.
• Quality of conversation beats volume of leads. A smaller, better-qualified funnel converted at a higher rate than a bigger, shallower one, even before the head-to-head numbers came in.