Skip to content
Neurally

AI for marketplaces and e-commerce.Aimed at conversion, retention and unit economics.

Marketplaces already run on matching, ranking and personalisation. The question now is which parts of the customer journey and the operation AI should change, and whether the margin moves when it does.

Book a free discovery call (opens in a new tab) Where to start, and what it costs

If you run a marketplace or an online retailer, you may recognise this.

  • “We've added AI features. Conversion hasn't noticed.”
  • “Search, recommendations and support each have their own AI project. They don't talk to each other.”
  • “Customer service costs keep rising with volume. We can't hire our way out.”
  • “We don't know whether to build this into the product or buy it.”

Marketplaces already run on matching

A marketplace exists to put the right option in front of the right person. That has been a data and ranking problem for as long as marketplaces have been online, which is why AI tends to arrive as a series of features: a smarter search box, a recommendation strip, a support assistant.

The return comes when the journey itself is rethought, the operation behind it changes shape, and someone is accountable for the commercial result. That is a product and operating-model decision before it is a technology one, and it is where we start.

The decisions that matter here.

  1. 01

    Where AI changes the customer's journey

    Search, matching, recommendations and merchandising are where a marketplace earns its conversion. Decide which of them AI should change, and what result would justify the work.

  2. 02

    Which side of the marketplace to help first

    Buyers, sellers and your own operation each offer a different return. A value case for each, with assumptions stated, makes the choice arguable.

  3. 03

    What happens to operational cost

    AI-assisted service and operations release capacity. That becomes a saving only when spending is reduced or avoided, so plan what the capacity is for.

  4. 04

    Build, buy or partner

    Some capabilities belong in your product. Others are better bought. We sell neither, and help you choose and brief the partners who build.

What tends to go wrong.

  • Features without a measure

    An AI feature ships because it can. Nobody agreed beforehand what change in conversion or retention would count as success.

  • Local wins that don't add up

    Each team improves its own metric, the customer journey stays disjointed, and the commercial result is hard to find.

  • Automating a journey that needed redesigning

    Automating a poor process preserves the wrong work, only faster.

How the four stages apply to a marketplace.

  1. Stage 01

    Decide

    Pick the few opportunities across buyer experience, seller tools and operations that could move conversion, retention or cost, each with a value case.

  2. Stage 02

    Design

    Redesign the journey and the workflows behind it, including which customer conversations stay with a person.

  3. Stage 03

    Deliver

    Test the riskiest assumption in a small, time-boxed pilot with agreed measures, then scale what works.

  4. Stage 04

    Embed

    Give each result a named owner and track it against the measures agreed at the start.

How the four stages work

Where this comes from.

28%
lift in marketplace conversion from a merchandisable app experienceClearScore, before Neurally
45%
new-user growth, with an 8% conversion uplift and 130% growth in app monthly active usersNot On The High Street, before Neurally
⅓
of UK revenue came from a new Jobs vertical as it grewGumtree, before Neurally

These are outcomes from roles Stuart held before founding Neurally. They are not Neurally client results, and the wider commercial figures reflect the work of whole teams, not AI alone.

Start with one or two areas where the return is suspected but unsized.

The Diagnostic takes a week. It looks at one or two priority areas, such as search and matching or customer operations, and gives you prioritised opportunities, an initial value hypothesis with its assumptions, a view of readiness, risk and dependencies, and a recommended next step.

Talk it through first, free (opens in a new tab) Full scope on the services page

Engagement 02

AI Advantage Diagnostic

£6,0001 week

  • Prioritised opportunities for the areas in scope
  • An initial value hypothesis, with its assumptions stated
  • A readiness, risk and dependency view
  • A recommended next step

Also worth a look: AI Opportunity Workshop (£2,500, 1 day).

Questions we hear.From this sector.

Have you worked in marketplaces?

Yes, as an operator. Stuart led product at Gumtree, Not On The High Street and ClearScore, and worked on search at HouseTrip, a holiday-rental marketplace. Those are roles held before Neurally, not Neurally client engagements.

Do you build recommendation or search systems?

We do not write the software ourselves, but we stay with the work until it is live. We help you decide what is worth building, specify pilots and acceptance criteria, help you choose and brief the teams or partners who build, and work alongside them on the first releases.

We are smaller than the companies you worked at. Is that a problem?

Not if the business has real operational complexity and a leadership team able to act. We work with businesses of roughly 50 to 1,000 people.

What will we have at the end of a Diagnostic?

Prioritised opportunities for the areas in scope, an initial value hypothesis with its assumptions stated, a readiness, risk and dependency view, and a recommended next step. They are yours to keep whether or not you continue.

Related reading

Start with a conversation.Thirty minutes, free.

Tell us what you are trying to decide. If we can help, we will say how and what it would cost. If we can't, we will say that too.

Book a free discovery call (opens in a new tab) Or send a message