Travel has a training problem.

The global travel industry is in a strange position. Consumer demand has held up far better than many forecast, even against a backdrop of geopolitical instability and regional volatility. Destinations are more complex and faster-moving than ever. Visa rules shift overnight, border policies change without warning, and travellers arrive better informed and less tolerant of generic itineraries than at any point in the last twenty years.

And yet the primary mechanism that destination marketing organisations, airlines, and hotel groups use to educate the trade is rooted in the digital architecture of the early 2000s. The Learning Management System. The one-and-done certification. The 45-minute webinar. The static destination specialist programme that the advisor enrols in, completes, and receives a badge for.

This is not a content problem. The content is often excellent. The model the content lives inside is the problem.

The operational reality of an advisor’s day

The first thing to understand is how training fits into an advisor’s actual working life. The answer is: barely. The average employee in any sector has roughly twenty-four minutes a week available for formal learning — around 1% of a full-time job. It is a corporate average, not a travel-trade specific number, but every advisor I have spoken to has confirmed it sounds about right, and many would argue it is generous.

Those minutes live in the cracks: the moment between calls, the gap before a meeting, the train ride home. The suppliers asking advisors to spend them on a 45-minute webinar are not competing with idleness. They are competing with the moment the advisor could be using to actually breathe.

There is a deeper issue underneath the maths. Human working memory has limited capacity, and when that capacity is full — which it is for most advisors for most of their day — new information does not get encoded. It bounces off. A webinar that tries to teach an entire destination strategy in one session is not failing because the content is bad. It is failing because the brain it is being delivered to is already at capacity. This is not a motivation problem. The advisors are not lazy and they are not disengaged. They are full. And the response to a full vessel is not a bigger jug.

The other forty hours

The twenty-four-minute figure is a working-week statistic. Take eight hours for sleep and eight for work, and there are still roughly forty waking hours every week that sit outside the office — the commute, the evening on the sofa, the Sunday morning scroll. Almost none of those minutes belong to anyone in the travel industry. They belong to the platforms that have spent fifteen years and billions of dollars perfecting the art of fitting useful, interesting, addictive content into the spaces between everything else. That is where habit forms, where loyalty is built, and where the next generation of trade engagement will actually be won.

Even short content, delivered in the right moments, is not enough on its own. There is a deeper biological problem underneath every certification campaign the industry runs — one that has been quietly destroying the return on training investment for decades, and that almost nobody on the supplier side is talking about honestly. It is called the Forgetting Curve.

The science nobody disputes

In the late 1880s, a German psychologist called Hermann Ebbinghaus ran one of the most important experiments in the history of cognitive science. Without active reinforcement, he found, memory decays exponentially. By the end of the first 24 hours, retention has dropped to roughly 30 to 50% of the original. By the end of the first week, around 90% of the new information is gone.

Ebbinghaus published this in 1885. The research has been replicated continuously for over 140 years, including by modern neuroscience using techniques he could not have imagined. It is one of the most settled findings in the entire field of memory research.

What this means in money

A destination specialist programme costs anywhere from a few thousand pounds for a small market to several hundred thousand for a flagship multi-region campaign. The content is usually excellent and the messaging polished. Within seven days of completion, by the science we have known for over a century, roughly 90% of the specific selling points that programme was designed to embed are no longer accessible to the advisors who completed it. The certificate has been awarded. The badge appears on a profile. But the actual product knowledge the campaign existed to install is gone.

Nobody is measuring the decay, because the metrics the industry uses — the number of certified agents, the completion rate, the course satisfaction score — are all captured at the moment of completion, before the curve has done its work.

This is the model failing, not the advisors

It is tempting to interpret this as a problem with advisors: they were not engaged enough, not committed to learning. This interpretation is wrong, and it has cost the industry millions. The Forgetting Curve applies to every human brain — to doctors memorising drug interactions, to pilots learning new airframes. If the design of a campaign produces 90% knowledge loss within a week, the failure is in the campaign design, not in the people who completed it.

The science also points at the antidote. If the same material is re-encountered at calculated intervals — the principle known as spaced repetition — the curve flattens dramatically. But this is structurally almost impossible to deliver through a traditional Learning Management System. An LMS is built around the certification event. It has no native capability to track an individual learner’s decay curve, predict when reinforcement is needed, and proactively surface the right content at the right moment. It was designed for compliance reporting.

Knowing that spaced repetition is the answer raises an immediate practical question the research itself does not address: how do you actually get a busy travel advisor to engage with reinforcement content frequently enough for it to work? You cannot legislate it, mandate it, or incentivise your way through it. The training has to become a habit, or it does not happen at all.

There is one body of work the industry can borrow from. Consumer technology has spent fifteen years and billions of dollars learning how to build daily habits. The playbook is described in detail in Nir Eyal’s 2014 book Hooked: a four-phase loop — trigger, action, variable reward, investment — that turns a behaviour into an automatic response. It is also almost entirely absent from B2B travel training.

The Hook Model, applied to the trade

Translating that loop into trade engagement requires four shifts. First, from external to internal triggers: away from ‘have you completed your training yet?’ emails toward the mental association between a professional moment and the platform — ‘I want to check if anything has changed about this destination before I draft the proposal.’ Second, the relentless reduction of friction in the action itself: open, scroll, done. Third, the embrace of variable rewards — a badge the advisor did not know existed, a public acknowledgement that lands without warning — rather than predictable points. Fourth, the investment loop: small acts of personalisation that make the platform more valuable to the user over time.

Why visibility is the actual currency

Almost every supplier-side loyalty programme in travel has been built around the wrong currency. The default architecture issues points: earn them, redeem them, and the points are the motivation. The pattern that comes back is consistent — a burst of activity, engagement dropping around month six, the programme quietly winding down by twelve months.

The points were never the motivation. They were a proxy for it. What advisors actually want is the recognition that comes with earning them: the public, named, specific acknowledgement of doing the work well, in front of their peers. Status, mastery, and belonging are all visibility mechanics. None of them is a points mechanic.

Everything argued so far — short-form content, habit-forming engagement, visibility-based motivation, spaced repetition delivered automatically — has a single underlying requirement the industry has been avoiding for two decades. None of it is deliverable through the Learning Management System architecture the industry has built around. The shift it has to make is from the LMS to the Learning Experience Platform.

The question every LMS was built to answer

Every LMS in the travel industry was designed to answer one question with high precision: did the agent complete the course? Every database table and reporting view is built around producing accurate, auditable answers to that one question. Which would be useful, if the question mattered.

Consider what a DMO, a supplier, or a hotel group is really paying for when they invest in trade training. It is not certificates or completion percentages. It is conversion. When an advisor is on a call tomorrow with a client deciding between two destinations, the question is not whether they were certified last quarter. It is whether they can confidently recall the differentiators that turn the conversation into a booking. The LMS has no way of answering that — it stops measuring at the moment of completion, which is the moment before the Forgetting Curve does its work.

What an LXP actually is

It is tempting to think the LMS could be fixed with add-ons: bolt on a quiz module, a recommendation engine, some gamification, an AI assistant. This is the instinct that produces what engineers call spaghetti code. It does not work because the underlying architecture is wrong. No amount of layering turns a filing cabinet into a habit-forming experience.

Where the LMS is administrator-centric, the LXP is user-centric. The first asks what the organisation needs the user to learn; the second asks what the user wants to learn to be successful. Where the LMS pushes mandatory courses on a schedule, the LXP pulls users into a dynamic ecosystem of content they want to spend time in. Where the LMS treats content as static, the LXP treats it as a living feed, surfaced by relevance. Where the LMS reports on completion, the LXP measures engagement, retention, and the correlation between training behaviour and commercial outcomes.

Why user-centric architecture matters

An LMS treats expertise as something that flows downward: head office writes the curriculum, the agents consume it, the certifications are awarded. Every advisor sees the same catalogue, in the same order, regardless of who they are or what they sell. An LXP starts from the opposite premise. It tailors what each advisor sees to the parameters that actually matter to their work — their region, their role, their specialisms. An advisor in Germany who sells luxury and cruise does not see the same feed as an advisor in Brazil who sells adventure and family travel.

Two philosophies of AI are being deployed across the travel industry right now. They produce different products, serve different customers, and lead to different futures for the trade. Most operators have not stopped to notice which one they are buying.

Replace or amplify

The first philosophy says the value of AI in travel is to replace the trade: AI travel planners marketed directly to consumers, booking assistants that bypass the advisor entirely. The end-state of this philosophy is a world in which the advisor is the friction the technology is built to remove.

The second says the value of AI is to give the trade superpowers: AI that sits beside the advisor, not in front of them. AI that hands the advisor knowledge they would never have had time to acquire, anticipates the questions their client is about to ask, and turns thirty years of supplier complexity into a single conversational layer they can rely on at the point of sale. The end-state of this philosophy is a world in which the advisor is the asset the technology is built to amplify.

Why amplification needs a walled garden

Once the amplification philosophy is chosen, several design decisions follow that go against the grain of how most of the AI industry has been building consumer tools. The most important is the walled garden. An amplification platform has to be built around verified content — the DMO’s official destination information, the supplier’s actual product data, the operator’s confirmed routes and schedules. It does not draw from the open internet. It does not improvise. It does not fill in gaps with plausible-sounding inventions.

The reason is straightforward: the AI cannot, under any circumstances, hallucinate. An advisor who confidently recommends a hotel that does not exist, a visa requirement that was repealed last year, or a route no carrier operates is not amplified. They are damaged. The consumer using a general-purpose AI to plan a holiday can verify, hedge, or live with the occasional fabrication. The advisor on a call with a high-value client at 11.30am cannot.

An amplification AI also acknowledges, honestly, what AI cannot do. It cannot replicate what an experienced advisor does when things go wrong: when the flight cancels at midnight, when border policy changes three days before departure. When the system breaks, the resolution does not happen online. It happens on the phone — and no AI can pick up the phone and negotiate human-to-human with a hotelier at 11pm. Not because it is not clever enough, but because it physically cannot do the thing.

Every argument in this paper has been chipping away at the same insight. Time poverty means most training does not even land. The Forgetting Curve means most of what lands does not stick. The legacy LMS cannot deliver the reinforcement biology requires. AI without amplification philosophy hollows out the trade. Points programmes plateau because the wrong currency is being earned. Every one of those is a way of saying the same thing: the metrics the industry reports on do not correspond to the outcomes it actually cares about.

Solving the glocal paradox along the way

Every DMO and major supplier operating across multiple markets runs into the same tension. Head office wants brand consistency. Markets want local relevance. It is not a trade-off — it is an architecture problem with a structural answer. A modern platform does not distribute one piece of content across many markets; it surfaces different content to different advisors based on who they are and where they sell. The Italian feed is genuinely Italian; the Brazilian feed is genuinely Brazilian. But both sit inside the same platform, under the same brand, governed by the same editorial standards. The consistency is structural. The relevance is operational. The contradiction dissolves.

Return on Education

Most current programmes report against the wrong things: completion rates, certified-agent counts, course-satisfaction scores, badges issued, engagement minutes. These are all measurements of activity at the point of training. None is a measurement of commercial outcome at the point of sale.

Operationally, Return on Education measurement requires three things most current programmes do not collect. First, the correlation between training engagement and booking activity at the level of the individual advisor — did the agents who completed the programme actually sell more of that destination in the following quarter? Second, the retention curve for product knowledge over time — six months on, do they still recall the differentiators the programme was designed to embed? Third, the confidence signal — on a call with a client, are they recommending the destination the training was about, or an easier sell because the training did not give them the recall they needed at the moment of truth? It is the hardest of the three to measure, and the most important.