CRM for a Small Travel Agency With Reps and Chats

The problem: scattered chats and replies that come too late

A small travel agency in Bogotá sells packages to San Andrés, Cartagena and the Coffee Region. It has three sales reps. Each one works from their own phone: personal WhatsApp, WhatsApp Business, and one who also answers Instagram comments when they remember to. Leads come from a Meta ad, referrals, the website and TikTok.

On Saturday night a traveler asks about availability for a long weekend. No rep is online. On Monday, when someone finally replies, the traveler has already booked with another agency. Nobody knows how many leads were lost this way, because there's no single place to see every chat: they're spread across three phones, and the agency owner only finds out about the problem once it's too late.

This is the typical pain of an agency that grew on reps and loose WhatsApp threads, not on a system. The question they ask themselves is simple: how do you get this organized without losing the close, personal touch you already have with clients?

What does a small agency with reps and hundreds of chats actually need?

Before adding any tool, it's worth separating what's really missing:

• One single place to see every chat, regardless of whether it came through WhatsApp, Instagram or TikTok.
• Fast replies after hours, especially at night and on weekends, when reps aren't online.
• Knowing which lead is hot and which one has gone cold, so the team doesn't chase someone who already said no or neglect someone ready to pay.
• Reps who keep having a say, because they know their clients and don't want a bot speaking on their behalf without review.

Draco, Skyline's AI CRM, is built exactly for this combination: several reps, several channels, and a volume of chats that no longer fits in anyone's memory.

What does the agency's inbox look like with Draco?

The first thing that changes is where the chats live. With Draco's unified inbox, WhatsApp, Instagram, Messenger and TikTok all arrive in the same place, and email also comes in as a channel. The agency's three reps stop checking four different apps: they open Draco and everything is there, sorted by conversation, no matter where the message came from.

This only solves half the problem. Having everything in one place doesn't help much if nobody is still replying at ten at night. That's where the second change comes in.

From loose chats to a funnel per destination

The agency sets up a different funnel for each product line: one for San Andrés, one for Cartagena, one for the Coffee Region. Each funnel has its own knowledge base and its own way of replying, so a lead asking about San Andrés gets San Andrés information, not a generic answer.

Who replies when no rep is online?

This is where Vega comes in, one of Draco's five AI agents. Vega is the one who handles attention, prequalifies and can close the sale in real time: it recommends from the catalog, quotes, builds packages and carries the conversation through to payment. The agency turns it on in copilot mode: Vega drafts the reply and a rep approves or corrects it before it goes out. Nothing is sent without a person reviewing it first.

Over time, once the team trusts how Vega replies in the San Andrés funnel (the one with the most volume), they move that funnel to autonomous mode. The other two funnels, newer ones, stay in copilot a bit longer. The agency can revert at any point if something doesn't feel right: it's not a one-way decision.

With this, the traveler from Saturday night no longer gets left without an answer. Vega replies, quotes the San Andrés package and puts together the booking while the rep sleeps. On Monday, the rep opens Draco, sees the full conversation, reviews what Vega proposed, and approves or adjusts the tone if something feels off.

How do hundreds of chats stay organized without losing the thread?

With three reps and several channels, volume stops being manageable by eye. This is where Draco splits the work across its team of five AI agents with distinct roles: Vega, Kai, Axel, Leo and Iris.

While Vega handles attention and quoting, Kai does the background work no rep has time to do by hand: it gives each lead a temperature (cold, warm, hot), assigns automatic tags, and moves it through the funnel stages as the conversation progresses. So when a rep opens Draco in the morning, they don't see a flat list of a hundred unsorted chats: they see the hot leads that need a call first, and can leave the ones still at the quoting stage for later.

That's the difference between "I have two hundred open chats and don't know where to start" and "I have twelve hot leads today, and the rest are in follow-up." Kai doesn't talk to customers: it organizes the work so that people and Vega talk better.

A typical day at the agency with Draco running

To see it end to end, here's the flow with the agency already using Draco:

1. Friday 8pm: a traveler writes on Instagram asking about a Cartagena plan for four people. Vega receives the message in the unified inbox, checks the Cartagena funnel, and replies with catalog options.
2. Friday 8:05pm: Kai tags the lead as warm and moves it to the quoting stage.
3. Saturday: the traveler replies with date questions. Vega continues the conversation within the same funnel, adjusts the quote, and gets it ready with the payment link.
4. Saturday afternoon: the traveler pays. Kai moves the lead to "won."
5. Monday 9am: the rep in charge of Cartagena opens Draco, sees the entire conversation from the first message through payment, and confirms everything went well. Since the Cartagena funnel is still in copilot, they review every reply Vega proposed before the system sends similar ones next time.

None of this went through anyone's personal WhatsApp, and nothing got lost between four apps. It all stayed in one place, with a history any rep can review even if they weren't the one who handled it at first.

What does the agency gain compared to before?

Is this replacing reps with a bot?

No. Draco doesn't introduce itself as a virtual assistant and doesn't sound robotic: it replies with short, warm messages in the brand's tone. Copilot mode exists precisely so the human team keeps deciding what gets sent, while the AI handles the repetitive work of replying fast and sorting leads. Reps don't disappear: they stop wasting time hunting for which phone a conversation was left on.

What to do if this is your agency

If your agency has reps working from different phones and you've already lost a sale by replying too late, the first step isn't hiring more people: it's bringing every chat into one place and letting someone reply while the team sleeps. You can see how Draco's inbox and funnels are set up at draco.skyln.co.