Why Online Marketplaces Are Evolving

Online marketplaces are evolving because more choice no longer automatically produces a better experience. Saturated catalogues, fatigue with filters, changing user expectations and generative AI are pushing platforms to understand what people want, not merely match the words they type.

The direction is from catalogue navigation towards assisted discovery. That does not mean conventional marketplaces are disappearing: their scale, liquidity, familiar interfaces, specialist tools and established trust remain considerable strengths.

One possible next step is the iMarketplace, a category we are proposing and defining publicly. As described in the difference between a marketplace and an iMarketplace, it is organised around intentions, native intelligence, conversation and connected universes rather than categories alone.

Four forces changing marketplaces

Marketplace evolution is not driven by one technology. It reflects the interaction of catalogue growth, interface fatigue, wider digital habits and newly practical forms of artificial intelligence.

| Force | What users experience | Likely platform response | |---|---|---| | Saturation of listings | More choice, duplication and uneven descriptions | Better interpretation, ranking and contextual matching | | Filter fatigue | Repeated decisions before relevant results appear | Natural-language requests and clarifying questions | | New expectations | Demand for speed, relevance and continuity | More individualised, assistant-led journeys | | Generative AI | Dialogue and content generation become practical | Native conversational search and listing assistance |

1. Saturation makes discovery harder

A marketplace benefits from a broad supply. More sellers can mean greater choice, competitive pricing and a better chance that somebody has exactly what a buyer needs. This network effect is one reason established platforms are so useful.

Yet catalogue growth also creates a discovery problem. Similar items may have inconsistent titles, incomplete descriptions or different category labels. Search results can contain thousands of technically relevant listings without making the decision easier. The structural issue is explored further in why catalogues are no longer enough.

Consider somebody looking for a suitcase five kilometres away, suitable for cabin luggage and available for collection this evening. A category-based journey may require choices about luggage type, dimensions, condition, price, location and delivery. The person’s actual intention is simpler: obtain an acceptable suitcase nearby before tomorrow’s journey.

An intent-based search system can treat distance, timing and practical suitability as parts of one request. It may still use categories and structured attributes behind the scenes, but the user does not have to know the catalogue’s internal organisation.

2. Filters transfer work to the user

Filters remain valuable when people know precisely what attributes matter. They are predictable, transparent and efficient for comparing standardised products. Their limitation is that users must translate a real-world need into the platform’s vocabulary.

That translation becomes tiring when filters are numerous, unclear or repeated across several services. A person arranging a driver to the airport may care about collection time, passenger count, luggage, child seating and reliability. If the request is split between separate text boxes, filter panels and message threads, the user becomes the journey coordinator.

A conversational marketplace can instead ask one useful follow-up question: “How many passengers and suitcases will there be?” This is the practical distinction between keyword matching and understanding intent instead of keywords. The aim is not to eliminate filters, but to use them as optional controls rather than the only route to relevance.

3. User expectations now come from assistants

People increasingly encounter digital systems that complete sentences, interpret natural language, generate summaries and maintain context across a conversation. These experiences influence what they expect elsewhere.

A marketplace search such as “I need a second-hand bike for commuting, I’m 1.75 metres tall and can collect near me this weekend” now feels like a reasonable request. Users may expect the platform to distinguish a commuter bike from a child’s bike, interpret location and availability, and ask about budget if it is missing.

This expectation is broader than personalisation based on previous clicks. It concerns individualisation: adapting to the stated situation, constraints and desired outcome. The distinction is part of the four I of the iMarketplace: Intelligence, Intention, Interaction and Individualisation.

Users also expect continuity between related tasks. Planning a weekend may involve accommodation, an event, local transport and pet care. Conventional specialist services can handle each component particularly well, but the user must often repeat dates, location and preferences. A multi-universe platform is designed to preserve that context across connected needs.

4. Generative AI changes the interface

Earlier marketplace automation typically concentrated on ranking, recommendations, fraud signals and customer support. These remain important. Generative AI adds an interface capable of receiving open-ended requests, asking questions and producing useful drafts.

In an AI marketplace, this can support both sides of an exchange. A buyer can describe a need conversationally. A seller who wants to sell second-hand clothes can upload a photograph and receive suggested titles, descriptions or attributes, subject to review. The wider implications are covered in why AI changes everything for a marketplace.

The architectural distinction matters. Native AI means intelligence is designed into discovery, listing creation, moderation and cross-universe matching from the outset. It is not simply a chatbot placed over an unchanged catalogue. This difference between foundational design and later integration is examined in native AI versus added AI.

AI also introduces responsibilities. Generated descriptions can be wrong, conversational systems can misunderstand constraints, and automated moderation can make mistakes. An effective AI assistant marketplace therefore needs confirmation steps, visible listing details, reporting mechanisms and routes for human judgement. Conversation should reduce effort without concealing uncertainty.

What is changing, and what remains valuable

The evolution of marketplaces is additive rather than a clean replacement of one model by another.

Established players such as eBay, Vinted, Leboncoin and Facebook Marketplace benefit from scale, audience familiarity, liquidity and well-developed selling habits. Specialist platforms can provide precise terminology, focused communities and tools fitted to a particular category. Their structural limits generally follow from the design era and purpose for which they were built, not from an absence of value.

Category structures also remain useful. They support inventory management, comparison, moderation and regulatory processes. The change is that categories need not be the user’s starting point. A platform can receive an ordinary-language intention and map it to structured information internally.

| Traditional starting point | Intent-led starting point | |---|---| | Choose a category | Describe the desired outcome | | Enter keywords | Use ordinary language | | Apply filters | Answer relevant follow-up questions | | Search each service separately | Carry context across connected universes | | Review a result grid | Receive explained, controllable suggestions |

The appropriate model depends on the task. A knowledgeable collector searching for an exact product code may prefer direct catalogue search. Someone furnishing a first flat may benefit from a conversation connecting furniture, delivery, a craftsperson and short local missions.

The iMarketplace response

An iMarketplace is not yet an industry-standard term. It names a concept being proposed here for a platform centred on user intentions and built with native AI from its foundations.

Its defining characteristics are:

Intentions before categories: the user explains what they are trying to achieve.
Native intelligence: AI participates across discovery, creation, moderation and recommendations.
Conversation: the platform can clarify ambiguity rather than returning an indiscriminate result grid.
Multiple universes: goods, services, housing, mobility, missions, events, animals and skills can coexist.
Connected journeys: one intention can lead to several relevant exchanges without repeatedly entering the same context.

This is distinct from an aggregator that merely displays results from unrelated sources, or a super-app assembled from separate functions. The intended value lies in shared understanding across the journey, as explained in why universes work better together.

WEVONE is one concrete illustration, not proof that the category has prevailed. Public since 2026, it is a young platform with a few hundred registered members and is far smaller than Vinted, Leboncoin, eBay or Facebook Marketplace. That difference matters because marketplace usefulness depends heavily on available local supply and demand.

Available today in one app are Tutus for second-hand goods and fashion, Nest for housing and space rental, Mission for services and local gigs, Events, and Pilote for transport and parcel delivery. More universes are planned.

Mia, WEVONE’s built-in AI, currently supports conversational natural-language search, listing assistance through photo analysis and title or price suggestions, moderation of listings and photos, and cross-universe recommendations. Map-based discovery is filtered by the area the user is viewing. The platform’s “Earn from every action” positioning covers buying, selling, renting, booking and earning, but income is never guaranteed; it depends on demand, location, condition and pricing.

Barriers to the evolution

The technical ability to hold a conversation does not ensure a useful marketplace. Platforms must still solve several difficult problems.

First, recommendations need adequate supply. An eloquent answer is of little use if no suitable suitcase, private tutor or gardening help is available nearby. Secondly, intentions can contain sensitive information, making privacy and data minimisation important. Thirdly, users need to understand why a result appeared and how to correct it.

Trust also remains fundamental. Identity checks where appropriate, secure processes, accurate listings, moderation and dispute handling cannot be replaced by fluent text. AI may assist these operations, but marketplace governance continues to determine whether exchanges feel dependable.

Conclusion

Marketplaces are evolving because catalogue scale has created new discovery costs just as users have become less willing to navigate complex filters. Generative AI makes a conversational, context-aware interface feasible, while changing digital habits make it increasingly expected.

The likely future is not the disappearance of catalogues or specialist marketplaces. It is a broader range of interfaces: direct search when the request is exact, filters when comparison matters, and dialogue when the need is contextual or spans several areas of life.

The iMarketplace represents one proposed response. It takes a different approach by treating intention as the starting point, AI as native infrastructure and multiple everyday needs as parts of one connected journey.

FAQ

Why are traditional marketplaces changing?

Their catalogues have grown, users face more filtering work, and natural-language AI now enables new forms of discovery and assistance.

Are filters becoming obsolete?

No. Filters remain useful for transparent comparison and precise control. They may increasingly sit behind or alongside conversational intent-based search.

What does generative AI add to a marketplace?

It can interpret ordinary language, ask clarifying questions, draft listings and preserve context. Its outputs still require controls, verification and effective moderation.

What is an iMarketplace?

It is a proposed category for a native-AI, conversational and multi-universe platform organised around user intentions rather than categories alone.

Is an iMarketplace the same as a chatbot marketplace?

No. A chatbot can be added to an existing search interface. An iMarketplace, as defined here, integrates intelligence throughout discovery, listings, moderation and connected journeys.

Will large specialist marketplaces remain relevant?

Yes. Scale, liquidity, community familiarity and specialist tooling are durable advantages. Intent-led platforms offer a different model that may suit users with contextual or cross-category needs.