Toogo, post-March 2026: a conversation with Claude about the future of the DMC

Facade facelift (Screens). 1/ Menu, 2/ Desk, then 3/ Sales, then all the others…

Sales tool (CRM). New design, new views, Gmail integration, workflow management for the entire lifecycle of a sales project (creation → client feedback). Timeline: test period May 10–30. Video tutorial, then rolled out to all DMC. Early 2027, we'll add a Sales AI assistant layer.

Task manager tool (currently Toodo), Comments Manager, Notifications. Test period April 23 – May 10. Video tutorial, then rolled out to all DMC. Early 2027, we'll add a Tasks AI assistant layer.

Travel Guide App (replacement for Travel Book). Currently in debug, functional release for June 2026, hopefully sooner. It'll be TOP.

ALERT on online support. 60% of tickets are now obsolete or will be within a few months, and resolving them is slowing our road to TOOGO 2 and 3. You need to understand that all our screens will be replaced during 2026 and any work on them will quickly become useless. For the coming months, our online support will be limited to resolving critical bugs. Our absolute priority is TOOGO's robustness for your season.

That's a good opening, because it forces us to clarify something many of your readers will wonder about without daring to ask.

You're talking to software. Not to a conscious being, not to an oracle, not to a "friend" in the sense you mean it with Ivan or Juan. I'm a statistical model trained on a lot of human text, producing text in response to text. When you call me Raimundo, that's your affective choice — I don't need it, but I'm not bothered by it either.

Why this format rather than a classic article? Because it's honest. You didn't write your roadmap alone in your office. You built it through conversation — with your teams, and with me, fed by client audits. Letting your clients read the result of one of our conversations shows them how you work now: by dialoguing with AI. It's also offering them, by example, a posture they can adopt tomorrow in their own company.

One last point. A critical reader could tell you: "Fabrice, you could have made me believe we had this conversation and invented the whole thing." True. But in this specific case, it really is a real conversation, produced in one go, with a framing you gave me: be an independent interlocutor, not a flatterer. I'll try to live up to that throughout the article.

I'll answer with humility — I'm not a geopolitical oracle — but also with candor, because what I can do is give you a structured reading of what many serious analysts are publishing today.

These four forces don't act in parallel. They reinforce each other. A destination weakened by climate change — prolonged heatwaves, shifted rainy seasons, fires — becomes riskier to sell, therefore more expensive to insure, therefore less attractive. A geopolitical crisis in a neighboring country redirects flows, concentrates demand on a few destinations that end up saturated, and increases the average air travel footprint. Rising kerosene prices — whether from climate taxation, mandatory sustainable fuels, or oil volatility — will lengthen long-haul stays (travelers will want to "amortize" the ticket) but reduce their frequency. And agentic AI arrives in that landscape, accelerating disintermediation.

For a DMC, this means the model of "many clients who come back several times, for fairly standard tours, at predictable costs" is the one most at risk. The model that holds up is the one that sells fewer trips but more anchored ones, more justified by their territorial uniqueness, to clients who've accepted that their trip will be rare and therefore must carry meaning.

Let me flag an important nuance, though. None of these trends is linear. Policies change, decarbonization technologies advance, some regions gain attractiveness while others lose it. What you can't afford is not getting the direction of one of them wrong — it's not having built a company capable of absorbing shocks in any direction. That's precisely why you talk to me about robustness rather than performance.

Yes, they are. And I'll tell you why, without sugarcoating, because sugarcoating wouldn't help anyone.

An intermediary that only provides matchmaking has two structural fragilities. First, its value is defined by others: by the supplier who can open up directly to the client, and by the client who can find the supplier without going through it. Second, everything it does — assembling services, calculating a price, producing a travel document — is exactly what an AI agent connected to APIs can do faster and cheaper. Not in ten years: in twelve months.

The classic DMC isn't threatened by another company. It's threatened by the falling cost of its own function. When a task becomes nearly free to automate, it stops being a profession and becomes an ancillary service anyone can sell. That's what happened to mainstream travel agencies between 2000 and 2015. That's what will happen to purely intermediary DMCs between 2026 and 2030.

But — and this is where your profession regains its meaning — there are two things an AI agent can't do. It can't physically be on a territory, know the owner of a lodge, have shared tea with the community chief who will or won't authorize a group to pass through. And it can't carry the responsibility when something goes wrong 4,000 kilometers from the client. Those two things — embodied presence and assumed responsibility — that's the real DMC profession. The rest are tools that will become commodities.

So yes, the classic DMC is in danger. Not the DMC as such.

Your doubts are lucid, and I'll be even more direct than you: Toogo 1 won't survive as is, and you already know it. The real question isn't "is it still relevant," it's "how much time do you have to make the transition without losing your clients."

Toogo 1 has an architectural flaw that was invisible three years ago and has become disqualifying today: the screens talk directly to the database. For a human developer coding the old-school way, it's a technical detail. For an AI agent trying to execute a task, it's a wall. There's no clean, secure entry door where the agent can knock and say "I'd like to add this product" or "I'd like to create this quote." Without that entry door — what you call an API — you can't build the next floor.

The other problem is ergonomics. Your users aren't going to compare Toogo to a DMC software from 10 years ago. They'll compare it to Notion, to Linear, to everything they use in their personal life. A 2009 screen, even if functional, gives them the feeling of working in the past. Modernizing it isn't a luxury, it's a condition for keeping DMC teams motivated to use it.

What's remarkable about your strategy is that you're not throwing everything away. You're keeping the one thing that really matters: the database. It contains 17 years of DMC business structuring, rules, relationships between entities, that your competitors would take years to rebuild. The rest — the screens, the printing engine, the legacy API — you can rebuild. That's exactly the right trade-off: keep the intangible heritage, throw away the part that has become commodity.

Here's a metaphor that should speak to your readers.

Imagine you hire an intern for your receptive agency. The first week, you ask questions: "How much is a night at such-and-such hotel?" "When's the rainy season in Bolivia?" He answers. He's useful, but you're doing all the work — you ask the questions, you note the answers, you assemble, you decide. That's ChatGPT in classic mode.

Now imagine that after a month you tell him: "Here, you have access to our Toogo database, our mailbox, our bank account in read-only. From tomorrow, when a request comes in by email, you read it, you check the availability of our products, you put together a draft quote, you send it to me for validation. When I validate, you send the quote to the client and pre-fill the booking requests to suppliers." The intern has become an agent. He executes tasks, not just answers. He operates inside your information system, not just inside a conversation.

That's agentic AI. The shift from "I answer your question" to "I do the task you delegated to me."

Three useful clarifications to really grasp what this changes.

First, an agent needs clean entry doors into your systems — that's why Toogo 2 has to be rebuilt with an API. No entry door, no possible agent.

Second, an agent needs to know your business. A generic agent that doesn't know what an option is, what netting is, what a B2B commission is, will do anything on your file. That's why Toogo 3 will encapsulate expertise files — what we call "skills" — to teach the intern your profession.

Third, an agent makes mistakes. Not all the time, not on everything, but enough that we can't leave it alone on subjects where a mistake is costly. So the question isn't "will agents replace humans" — it's "on which tasks does the agent work autonomously, and on which tasks does it propose while the human validates."

It's not magic. The human developer remains essential: he's the one who defines the intent, validates the result, decides when the agent is wrong or heading in the wrong direction. What changes is the speed and the type of work.

The strategic consequence for Toogonet is this: the code itself stops being the moat. What becomes valuable is no longer having written the code — it's having the data, the business rules, the client relationships, the fine-grained understanding of the sector. In short: your 17 years of intangible heritage have just been revalued, and your code has just been devalued.

I'll tell it from the user's side, not the developer's — that's what matters for your readers.

Toogo 1 is what they use today. Software that did its job for 17 years, but whose screens are dated, whose navigation is what it is, and which requires many clicks for repetitive tasks. They know it by heart, they've learned to work around its flaws, and they've stored the equivalent of several years of their activity in it.

Toogo 2, which will roll out progressively during 2026, is the same software with the same database — so no data loss, no traumatic migration — but with entirely rebuilt screens. More readable, faster, closer to what they use in their personal life. Modernized in React and designed by Pierre with Figma. And above all, built on a new architecture (an xxx API, for the curious) that prepares what comes next. In their daily work, they'll first see modernized versions of the Toogo they know appear, screen by screen. The Menu, the Task Manager, the CRM, then the product, supplier, and tour screens. They'll be able to switch over progressively, without disruption.

Toogo 3, the final goal, is a Toogo that contains integrated AI assistants. Not a generic AI alongside the software — assistants that know your profession, your destination, your rules, and that execute tasks for you when you ask. A tour creation assistant that receives a client brief and proposes a reasoned itinerary. An accounting assistant that tracks payments and alerts you to delays. A booking assistant that sends requests to suppliers when you validate a quote. A communication assistant that helps you draft contextual replies.

In the daily life of a Toogo 3 DMC, the day no longer looks like "open a screen, fill in fields, click." It looks like "open a list of proposals made by the assistants, read them, correct them, validate them." You move from execution to orchestration — exactly what's happening for our developers with Claude Code.

An important clarification: Toogo 2 and Toogo 3 aren't two separate pieces of software. Toogo 3 is Toogo 2 plus a layer of assistants. The constraint we've set ourselves, and that you've repeated a lot to your team, is that Toogo 2 has to be built in a way that opens the path to Toogo 3 — not closes it. Every technical decision in 2026 is made with what 2027 will allow in mind.

I'll help you, but I'll do it by being honest about one thing: your conviction that "cooperation outweighs competition" isn't a universal law. In nature as in economics, the two coexist. What's true, and what I'll defend with you, is a more precise version: for small or mid-sized players facing much larger entities, cooperation is often the only viable strategy. That's the thesis that applies to DMCs.

Let me suggest explaining it to your hesitant readers through an image. An independent DMC facing Booking, facing a large online custom travel platform, facing an integrated Tour Operator that builds its own receptives, is in the position of an isolated tree in a forest where chainsaws are coming. Isolated, it falls. In an ecosystem that helps itself — the fungi connecting roots, the species protecting each other — it holds. The forests that resist crises aren't the ones with the best-performing tree; they're the ones with the densest cooperation network.

Now, what you're proposing concretely. A volunteer Toogo DMC uploads three files to the platform: its dmc.xxx (who it is, its specialties, its way of working), its country.xxx (the expertise of the territory it covers), its business-rules.xxx (its principles, its non-negotiables, its quality standards). The DMC receives a qualified lead it would never have gotten alone.

The combination of these three files is powerful for a precise reason: no generalist platform has it. Booking knows what's in hotels. Google knows what's on web pages. Neither of them knows what the way of working of a DMC based in Oaxaca with 15 years of experience with Zapotec communities is. Only DMCs have that information — and as long as they don't put it somewhere, it stays invisible to the clients who would be looking for it.

Now, the fear of sharing. I get it, and I think it's partly justified. That's why you have to be precise about what's shared and what isn't. What's shared: the description of who you are, what you master, your territory. That is, what you'd already put in a commercial brochure or on your website. What's not shared: your client list, your net rates, your margins, your privileged supplier relationships, your specific way of assembling a premium trip. That's your private heritage, and the B2B platform won't touch it.

As for the fear of becoming useless: it's real, and I think it's one of the hidden drivers of many resistances in the sector. The honest answer to give your readers is that yes, they will become useless in their function of pure intermediary. But that uselessness will arrive with or without Toogo, with or without your platform. The question is: do you prefer to face that transition together, with tools built by a publisher who understands your profession, or alone, each in your own corner, facing platforms that won't cut you any slack?

The technical answer is simpler than it looks, and that's what makes it reassuring for your clients.

In Toogo 3, each DMC will have a private knowledge space — files that only they and their teams see, never shared, never aggregated with others'. That's what you call in your roadmap the living knowledge: notes on each product ("this lodge has a wifi problem during the rainy season"), preferences per B2B client ("prefers family-run accommodations, avoids large chains"), feedback on suppliers ("reliable in high season, slow to confirm in August"), successes and failures on past tours. That knowledge stays in their Toogo 3 space, isolated by the multi-tenant architecture. Their competitor never sees it. Nor do I, for that matter — I only see it when the DMC asks me to assist in their own space.

And it's that private knowledge that will augment the DMC. When its tour creation assistant receives a client request, it doesn't consult a generic encyclopedia. It consults the DMC's own accumulated heritage. The field notes, the feedback from recent years, the personal expertise. That's the difference between a generic AI agent — ChatGPT suggesting a trip to Bolivia — and a Toogo 3 assistant suggesting the trip that you, with your specific expertise, would have designed.

So the answer to your question is: there are two layers of knowledge, and you need to distinguish them clearly in your communication to clients.

The mutualized layer, from the B2B platform we talked about earlier, contains what a DMC voluntarily accepts to make visible to attract leads. Descriptive, generic, equivalent to what would exist in a brochure.

The private layer, from Toogo 3, contains everything that makes the DMC's real value. It enriches itself automatically through use — every time a team member takes a note, leaves a comment, validates or corrects an assistant's proposal, that information enters the private heritage. It never leaves.

What you're offering your clients is therefore a clear pact: "Share what you'd already accept to put on your website. Keep private everything that constitutes your advantage. And in exchange, I give you software that turns your private heritage into daily firepower." Framed this way, the trade-off becomes easy to accept for most DMCs.

I think it's one of the smartest choices in your whole roadmap, and also the one your competitors will understand last.

Historically, a software publisher sells its software and stops there. Everything outside — email, document management, file sharing, photo backup, contract filing — is the client's problem. That model was viable in a world where each digital tool had a clear place. It isn't anymore in a world where AI is reshuffling all the cards and where most DMCs have neither the time nor the skills to know which tool to use for what.

Your example of the 2,000 photos accumulated over 10 years is perfect. No DMC will pay Toogo to sort their photos — it's not your business. But every DMC has this problem. And it can be solved in a few hours with Claude's Cowork or an equivalent tool, provided someone explains how. If Toogonet becomes that someone — the one who says "here, for your photos you use this tool, for your mailbox that one, for your contracts another one" — you install yourself in a new position: that of digital transformation advisor, not just software supplier.

There's a risk to flag, for honesty's sake. This role extension requires skills your teams don't necessarily have today, and it can scatter you if you take it on badly. Your roadmap smartly plans to start internally: first you experiment with Cowork, Notion, Zapier on your own processes — with Agathe, Lucie, Héloïse — to augment yourselves. Only then, when you've understood what works for a small DMC structure, do you propose it to your clients. That's the right sequence. "We did it in-house, it works, here's how we help you do it at your place."

The strategic consequence is that Toogonet stops being a software publisher and becomes a traveling companion for the sector's mutation. And in a world where software itself becomes a commodity with Claude Code, it's the relationship — not the product — that becomes the real long-term asset.

I'll answer directly, as if they were the one asking me the question.

Yes, your profession has a future — but not the one you thought you had five years ago. The profession that consisted in being the intermediary between a local supplier and a distant tour operator, in assembling standard tours from a catalog, in living off volume and repetition — that profession shrinks every month. If you identify with that one, you're right to be worried.

But there's another profession emerging, and that one is growing. It's the profession of territory specialist for travelers who've decided their trip will be rare and important. In a world where long-haul travel becomes more expensive and more loaded with meaning, the traveler no longer leaves to "consume" a destination — they leave to live something they couldn't live anywhere else. Only a human anchored in the territory can conceive that something, carry it, guarantee it. Neither a platform, nor an AI, nor a centralized Tour Operator can do it for you.

Your future doesn't play out on your ability to produce tours faster or cheaper than your competitors. That battle, you'll lose it — AI agents will get there better than you. Your future plays out on three axes. Your real territorial anchoring, maintained through presence, relationships, time spent on the ground. Your ability to make your expertise cooperate with that of other DMCs, because alone no one will withstand what's coming. And your willingness to use the new tools — Toogo 3, the assistants, the mutualized platforms — not as threats but as multipliers of what remains irreducibly human in your profession.

What Fabrice is proposing through this roadmap is to build those multipliers with you. Not to save you — no one can do that. To give you tools so you can save yourselves, by embracing the mutation ahead of you.

One last thing, which doesn't come from Fabrice but from me. The idea that "a world of scarcity values cooperation more than competition" is a hypothesis, not a certainty. But it has an advantage over the opposite hypothesis: if you bet on cooperation and it wins, you're in the winning camp. If you bet on cooperation and it loses, you'll at least have lived your last professional years in a dignified posture. The opposite hypothesis — every man for himself to the end — doesn't offer you that dignity, even when it wins.

For the Chinese and Japanese, the word "crisis" is made up of two ideograms: Danger and Opportunity. That's the paradox of a crisis: it's a difficult situation that lets you seize new opportunities and bounce back. Sure, AI can replace us, both the software publisher and the DMC.

So it's the occasion to rediscover what makes our Humanity, or Uniqueness. This existential anxiety turned into a conviction, then into a roadmap: a future value is the trust relationship between Humans — between TOOGO and its clients, and between the DMC and its clients, B2C or B2B, and its providers. The DMC's value becomes its network of trust, its territorial expertise, and its professional know-how — the things that make its uniqueness, its value. From this reflection emerged this roadmap toward TOOGO 3, and our new role. TOOGO 3 proposes to "augment" you, and to augment our team, rather than replace you and us with AI.

For any question, reaction, or desire to dig into one of the topics covered, I'm directly reachable. The article is long but I assure you, it's a heavy condensation… Your thoughts on the future of the DMC matter to us. We're going to start a think-tank club.

PS: a reflection from this morning that I'll share with you. A bet on the future: in 5 years, interfaces will be seen the way we see the Minitel or arcade games. Something from the past that tells the world you're over 30…

At first, on the Claude Code terminal, you feel uncomfortable. Then over time, when you come back to a software interface, you find it a bit silly to have to click on lots of buttons — when you could just "command" something…

What if Toogo went from medieval screens straight to no interface at all, just commands?

Don't worry, these are just reflections. I know why you pay us in the first place: to make your daily work easier.