Intention: The Second I of an iMarketplace

Intention is the real unit of an iMarketplace because it expresses the outcome a person wants, rather than the category in which an item happens to sit. A platform can read that intent from a sentence by identifying the need, context, constraints and implied actions, then asking a clarifying question where necessary. The result is a journey organised around purpose rather than a sequence of disconnected searches.

This is the second of the four I’s: Intelligence, Intention, Interaction and Individualisation. Together, the four I of the iMarketplace describe a category we are proposing and documenting; iMarketplace is not yet an industry-standard term.

As we define it here, intention gives intelligence something useful to understand. Without it, even sophisticated technology may simply produce a more polished catalogue.

What intention means in an iMarketplace

An intention is the practical objective behind a request. It may be explicit, as in “I need a driver to the airport at 5am on Tuesday”, or only partly expressed, as in “airport Tuesday morning”.

The words are evidence of the intention, but they are not the intention itself. The underlying need includes who or what must move, where the journey starts, when arrival is required, how much flexibility exists and whether luggage or accessibility matters.

This distinction is explored more fully in understanding intent instead of keywords. A keyword-led system chiefly asks, “Which records contain these terms?” An intent-based system asks, “What is this person trying to accomplish, and what information is still missing?”

From nouns to outcomes

Traditional catalogue structures are usually built around nouns: clothes, bicycles, flats, lessons or transport. Intentions are more often expressed as outcomes:

“I want to sell second-hand clothes before I move.”
“I need a private tutor for my daughter on Wednesday evenings.”
“Find me a second-hand bike nearby that will fit an adult commuter.”
“I am planning a weekend and need somewhere to stay.”
“Can someone collect this suitcase five kilometres away?”

Each sentence may contain an object or service, but the platform’s task is broader than recognising the noun. It must interpret the desired action, relevant conditions and likely sequence of decisions.

How a platform reads intent from a sentence

An AI marketplace does not literally read a person’s mind. It analyses the information the person has supplied and forms a working interpretation that can be checked through dialogue.

A useful process has several stages.

1. Identify the central need

The platform first distinguishes the main action from surrounding detail. In “I need someone to take me and two suitcases to the airport tomorrow morning”, the central need is passenger transport. “Two suitcases” affects suitability, but it is not the main request.

The system may represent the request as a structured set of elements:

| Intent element | Information inferred or requested | |---|---| | Desired action | Book or find transport | | Person or object | One passenger and two suitcases | | Origin | Current location or stated address | | Destination | A particular airport or terminal | | Time | Tomorrow morning | | Constraints | Luggage capacity, arrival time, budget | | Uncertainty | Exact collection time and airport |

This conversion from ordinary language into usable meaning is related to semantic search and filters, but an iMarketplace goes further by allowing the interpretation to shape the whole journey.

2. Use context carefully

Context can reduce effort. If a user says “find near me” while viewing a particular area on a map, the visible area may provide a useful local boundary. If the user has already stated that a weekend is in Bristol, a later request for pet care should not automatically be assumed to refer to Bristol; the platform may need to ask whether care is required at home or at the destination.

Context should therefore be relevant, proportionate and visible to the user. Good intent interpretation does not mean making silent assumptions from every available data point.

3. Detect what is missing

Most natural requests are incomplete. “I need gardening help” does not say whether the work involves mowing a small lawn, removing a tree or maintaining a garden every fortnight. These are materially different missions.

A conversational marketplace can ask one focused question at a time:

What kind of gardening work do you need, and roughly where is the garden?

This is why conversational search is more than a search bar that accepts longer sentences. The dialogue resolves ambiguity before presenting unsuitable results.

4. Match across universes

Once the intention is understood, the platform can consider more than one type of supply. Planning a weekend might involve a short rental, an event, local transport, pet care at home and equipment to borrow. A category-led platform would normally treat these as separate searches.

An iMarketplace is designed to connect them as parts of one intention. This is the practical significance of a multi-universe platform: goods, services, housing, mobility, missions, events, animals and skills can participate in the same journey.

5. Learn from explicit outcomes

A match is not successful merely because someone clicked it. A stronger signal may be that the user contacted a provider, booked a service, saved an option, rejected a suggestion or explained that it was unsuitable.

Over time, outcome data may improve how the platform interprets similar requests. That learning must still respect privacy, user control and the possibility that two superficially similar people want different things.

A concrete example: the suitcase five kilometres away

Suppose someone writes:

I bought a suitcase from a seller five kilometres away, but I cannot collect it tonight.

A catalogue may recognise “suitcase” and display more luggage. An intent-based system should recognise that the purchase has already happened and that the remaining need is delivery.

It might ask for the collection and delivery areas, the suitcase dimensions, the required time and whether the seller can hand it over. The relevant result could then come from mobility or a local delivery mission, not from second-hand goods. The detailed journey in I need a parcel delivered shows how an object can generate a transport intention after a transaction.

This example also illustrates why intention is the unit of the platform. The suitcase belongs to one universe, while the action needed now belongs to another. The user’s purpose connects them.

A broader example: planning a weekend

“I am planning a weekend by the coast” sounds like one sentence, but it may contain several linked needs:

Find a weekend rental within budget.
Arrange a driver to the station or book local transport.
Discover an event on Saturday evening.
Find pet care near home.
Rent equipment for an outdoor activity.

The system should not assume that every component is wanted. It could first ask what has already been arranged, then offer relevant next steps. This preserves user agency while reducing repeated searches.

A true multi-universe experience is not simply several menus under one account. As why universes work better together explains, the value comes from relationships between needs, not from the number of categories displayed.

Intention compared with catalogue search

Catalogue search remains useful when the user knows exactly what to buy. Established marketplaces have significant strengths in scale, liquidity, familiar browsing habits, trusted transaction patterns and specialist seller tooling. Their structural limits often reflect the era and category focus for which they were designed, rather than an absence of technical capability.

| Approach | Starting point | Typical input | Best suited to | Main limitation | |---|---|---|---|---| | Category browsing | Product taxonomy | Menu selection | Exploration within a known category | User must translate the need into the catalogue structure | | Keyword search | Matching terms | Short query | Known items or brands | Context and implied actions may be missed | | Filtered search | Structured attributes | Checkboxes and ranges | Precise comparison | Requires the user to know which attributes matter | | Intent-based search | Desired outcome | Ordinary-language sentence | Complex or cross-universe needs | Requires careful interpretation and clarification | | Conversational assistance | Evolving understanding | Dialogue | Ambiguous or multi-step intentions | Must avoid unnecessary questions and overconfident assumptions |

These approaches can coexist. Filters remain efficient for adjusting a known requirement, while conversation is valuable when the user does not yet know which filters to choose. The distinction between a catalogue and a conversation is therefore about the primary organising logic, not the abolition of browsing.

WEVONE as an illustration

WEVONE takes a different approach by organising one app around several universes: 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.

Available today, its built-in AI assistant Mia accepts natural-language requests, assists with listings through photo analysis and title or price suggestions, moderates listings and photos, and provides cross-universe recommendations. Local discovery is map-based and filtered by the area the user is viewing.

This makes WEVONE one concrete illustration of an AI assistant marketplace designed around intentions. It is also a young platform, public since 2026, with a few hundred registered members as of August 2026. It is far smaller than Vinted, Leboncoin, eBay or Facebook Marketplace, whose scale, supply and established user habits remain substantial advantages.

Not every multi-step intention is fully automated today. Connecting an entire move, weekend or new-flat journey is the broader design direction of the iMarketplace concept, rather than a claim that every possible combination is already complete. Platform features and terms can change, so readers should check each service directly.

What responsible intent understanding requires

Intent interpretation creates obligations as well as convenience. A platform should:

distinguish supplied facts from inferred possibilities;
ask before acting on important ambiguity;
explain why a result is relevant;
allow users to correct the interpretation;
minimise the personal information used;
avoid treating historical behaviour as a permanent preference;
apply suitable safeguards to housing, transport, care and other sensitive needs.

The objective is not to predict everything. It is to reduce avoidable effort while leaving meaningful decisions with the individual.

Conclusion

Intention is the second I of the iMarketplace because it defines what intelligence should serve. The modern platform unit is not simply a listing, keyword, category or click; it is the outcome the user is trying to achieve.

By interpreting ordinary language, identifying constraints, asking clarifying questions and connecting relevant universes, an iMarketplace can turn a sentence into a practical journey. Its success should be judged not by how confidently it guesses, but by how accurately, transparently and usefully it helps the user move from need to outcome.

FAQ

What is intention in an iMarketplace?

It is the outcome a user wants to achieve, including the action, context and constraints behind their words.

How is intent-based search different from keyword search?

Keyword search matches terms with listings. Intent-based search attempts to understand the purpose of the request and may ask for missing information.

Can an AI understand intention from one sentence?

It can form a useful interpretation, especially when the request is specific. It should ask a clarifying question when important details are uncertain.

Does intention replace categories and filters?

No. Categories and filters can remain useful tools. The difference is that the user’s purpose becomes the starting point rather than the catalogue structure.

What makes an iMarketplace multi-universe?

It connects different areas such as goods, services, housing, mobility, missions and events within one experience and, where relevant, one intention.

Is iMarketplace an established industry term?

No. It is a category WEVONE is proposing and defining publicly. It should not be treated as an already recognised industry standard.