@louis · occasions · demo
Louis
The concierge who remembers what you told him.
Concierge Memory you control Occasions and gifting Companion to Sid and Felix

What it is. Louis is the Analyzer Engine configured for one moment: a customer choosing what to buy for an occasion, such as a date night, a gift or a celebration. He builds a package from the brand's own range using what the customer has chosen to share, the season and the weather, prices it from the customer's member value, and adds a surprise gift chosen for them, which they can take, share as credit with a friend, or swap for store credit.

How it's integrated. Louis reads orders and the catalogue from the brand's store, issues single-use codes through the brand's discount system, and writes zero-party data such as occasions and taste to the CRM with consent.

Live demo · try a scenario

Try Louis yourself. Different replies, different packages.

Sam, a regular at Bottle & Rind, a fictional wine and cheese shop, has an occasion coming up at the weekend. Choose how Sam replies across three turns. Before the first reply you can ask Louis to forget what he remembers. The brain panel shows the member value model building in real time. At the close, Louis builds a package and adds a surprise chosen for Sam, which Sam can take, share as credit, or swap for credit. Figures are illustrative.

Louis session · concierge
Louis brain · real-time

Conversation stage

Member value Hidden

£0
Reading…
Genuine detail —
Price focus —

Member value model

    Signals

    awaiting first reply

    Zero-party data, with consent

      Decision

      Internal record

        How Louis behaves

        Three steps. Nothing taken without asking.

        Louis is a single conversation, performed in three parts. He asks about the occasion. He uses only what the customer has shared, and shows it. Then he builds the package and adds a surprise chosen for them.

        First
        He asks about the occasion.

        Louis never opens with a product or a price. He asks what the occasion is, and lets it set the brief: a date night, a dinner for six, a gift, a quiet weekend. He adds what the brand already knows, such as the season and the weather, and keeps the questions short.

        The opener is an invitation.

        At Bottle & Rind: “Evening, Sam. What's the occasion?” The question is open on purpose. One answer tells Louis the occasion, the day and how many people.

        Each answer shapes the next question.

        An anniversary on a warm Friday leads to something lighter and chilled. A dinner for six leads to quantities. Louis acknowledges what was said before he asks again.

        Second
        He uses only what you shared.

        Louis works from what the customer has chosen to tell the brand. When he uses something he remembers, he says where it came from, and the customer can delete it.

        Every memory shows its source.

        In the demo, Louis remembers that Sam mentioned a trip to Lisbon last spring, and shows it before the first reply. If Sam asks him to forget it, Louis never mentions Lisbon and the record shows memory_used as none.

        Nothing hidden about what he knows.

        Occasions and taste go to the CRM as zero-party data, with consent. The customer always sees exactly which remembered detail Louis is using.

        Third
        He builds the package and adds a surprise.

        Louis builds a package from the brand's own range, prices it from the customer's member value, and adds a small gift chosen for them, never drawn at random. The customer can take the gift, share it as credit with a friend, or swap it for store credit.

        The surprise is chosen, not drawn.

        For an anniversary, a beeswax candle for the table. For a cheese-led package, a jar of membrillo. Louis always says why he picked it.

        Take it, share as credit, or swap.

        Taking the gift confirms the match was right. Sharing turns it into credit for the customer and a friend, through a single-use link the customer sends themselves, and marks them as a likely advocate. Swapping it for store credit tells the system the match missed, so it learns.

        What the brand actually receives
        From a single Louis conversation
        An occasion record, with consent built in.
        Below is the record from the anniversary path in the demo above, with the Lisbon memory kept and the gift shared. This is what lands in the partner's CRM when the chat closes: the occasion, the taste behind it, and the choice the customer made.
        Illustrative example, fictional names
        conversation_id #louis-b7e2
        Bottle & Rind · Thu 26 Jun · member_id #sam-1104
        9 fields · 52 seconds elapsed
        occasion
        "anniversary, Friday"
        Shared in the first reply.
        style
        "Portuguese, vinho verde"
        Confirmed after Louis asked about Lisbon.
        budget
        "up to £100"
        A special occasion, so a sparkling wine was added for a toast.
        memory_used
        "Lisbon trip, last spring"
        Shown to the customer with its source, and kept.
        weather_context
        "warm, about 23°C"
        Friday's forecast, used to lean towards chilled wines.
        package_items
        "vinho verde, São Jorge, sardines, English sparkling brut"
        Built from Bottle & Rind's own range.
        package_total
        £63
        £68 at list, with a £5 member saving.
        gift_choice
        "share"
        Beeswax table candle, shared as credit with a friend. Marks a likely advocate.
        member_value_tier
        "regular · 11 orders"
        Set the size of the member saving.
        Status
        Louis for occasions. Sid and Felix alongside.
        Three BehaviorBots, one engine. Whitelabel for partners across cultural sector, premium hospitality, membership programmes and considered-purchase D2C. Get in touch to become a founding partner, or meet Sid and Felix to see the acquisition and retention moments.