Travel-tech
From WhatsApp and Excel to one system: the Notion architecture behind a travel experiences platform
Notion discovery and architecture for a travel experiences platform in Mexico: 40 agents, 3 countries, and a 3-phase plan to leave Excel behind.
40
sales agents mapped onto a single commercial pipeline
3
countries operating on the same workspace design
12
business processes prioritized into a 3-phase plan
Discovery and plan validated with the client's leadership.
A custom travel experiences platform — 40 sales agents and teams across Mexico, Guatemala, and Peru — ran its entire commercial flow on WhatsApp and Excel. La Agencia AI ran the full discovery, mapped 12 business processes, and designed the Notion architecture to hold them: a lead CRM, sales registration, commission tracking, and role-based dashboards, prioritized into a 3-phase plan that the client’s leadership validated meeting by meeting.
Where things stood
Leads arrived over WhatsApp. Sales went into one spreadsheet; commissions lived in another. At the end of every month, each agent scanned 300 to 400 rows hunting for their code, just to confirm what they had earned — “a fairly archaic system,” in the team’s own words.
Itineraries had their own version of the problem. “We almost always start from scratch again,” the head of operations told us. Twenty years of trips designed, and every new proposal began on a blank page. The company’s know-how — which hotel, which guide, which restaurant works in each destination — was scattered across a WhatsApp group, destination spreadsheets, and Word guides.
None of this was a cheap-tooling problem. It was an architecture problem: critical information with no single place to live.
The solution, in three acts
Here is the flow we designed.
Act 1 — Leads stop slipping away. A WhatsApp message lands at 3 p.m.: “I want a trip to Peru in July.” The agent logs it in the CRM — destination, budget, dates — the system assigns it by workload and moves it through a clear pipeline: incoming, proposal, booked, completed. Leadership’s dashboard answers the questions that used to take phone calls: how many leads came in this week, who owns them, what each agent’s close rate looks like.
Act 2 — A closed sale becomes data. The deal amount already lives in the system, so the commission calculates itself: amount times the agent’s rate. Every agent opens a personal dashboard — theirs alone — showing pending, paid, and the per-deal breakdown. Nobody combs through 400 spreadsheet rows again.
Act 3 — Knowledge moves out of the chat. Recommendations get captured by destination, with who made them, when, and for which kind of client. When an agent builds a London proposal, the best options show up where the work happens — not three screens away, in a chat message from a year ago.
How it was built — and what we chose not to build
Discovery ran as a series of working sessions with leadership and operations. Every process was scored on urgency versus impact and placed into one of three phases: Phase 1 holds the revenue engine — CRM, sales registration, commissions, dashboards; Phase 2, integrated accounting and the knowledge base; Phase 3, compliance and sensitive-data protection.
Two decisions shaped the project. The first came from the client: integrated accounting jumped up a phase because it was their biggest complaint, and the plan was reordered that same week. The second came from us: we recommended against replacing their itinerary platform — it connects to a GDS and holds years of historical data — and turned that piece into a feasibility study instead. You don’t perform open-heart surgery on a system that is running in production.
Where it stands
What we can say truthfully today: discovery is done, 12 processes are documented and prioritized, and a 3-phase architecture has been validated by the client’s leadership. The feedback that mattered most to us wasn’t technical: “you explain it in a very simple way.” That’s the house standard — if the plan needs an engineer sitting next to it to be understood, the plan is wrong.
Under NDA, the client’s name stays out. The sector, the problem, and the decisions are real.
Sound familiar?
If your commercial operation lives between WhatsApp and Excel, the problem isn’t your team — it’s that nobody designed where the information should live. Thirty minutes is enough to find out whether this applies to your case.