Skip to content
Free audit (opens in a new tab)

Case study · AI guest memory & follow-up

A tasting room that remembers every guest and follows up for you.

OKAERI (おかえり, “welcome back”) turns each tasting-room visit into a lead. Guests scan a QR code and note what they taste; when they come back, the team already knows them, and a personal follow-up is waiting for the manager’s yes.

Watercolor of a sake bottle and glass beside four linked circles: past tastings, preferences, memories and context, future opportunities
Illustration: what OKAERI remembers about each guest.
For
A Japanese sake tasting room in the US
Built
QR guest web app, guest memory, check-in alerts, AI lead scoring
Status
Pilot in progress. Results have not been measured yet.

Tasting rooms forget every guest at the door.

Guests tell the bar what they love, what they need and what they’re planning. All of it is gone when they walk out, so follow-up is generic, or never happens.

What one guest shared in a single visit (from the demo):

  • Loved

    “N-end’s umami. REI was too light.”

  • Asked

    “Does N-end ship to Oregon?”

  • Planned

    “Team offsite for 12 in December.”

How it works

Small AI agents pass the guest along, and a person approves before anything goes out.

Diagram: past tastings, memories and context, and preferences connect to future opportunities for each guest
Illustration: one memory per guest, from past tastings to what to offer next.
  1. 01Built

    The guest scans a QR code

    A short web page, no app to download. Guests say who they are, agree to be remembered, and log each pour as loved, liked or not for me.

  2. 02Built

    One memory per guest

    Tasting notes and check-ins are saved to that guest’s own memory, so recall never mixes people up.

  3. 03Built

    The team gets a check-in alert

    When a returning guest checks in, the team sees what they loved and what to follow up on, before they reach the bar.

  4. 04Built

    AI scores the lead

    An agent reads the guest’s memory and scores them as a lead: a gift, an offsite, wholesale, a club sign-up.

  5. 05Next

    A personal follow-up, drafted

    An agent drafts the email and a sake pick, written from what this guest actually said.

  6. 06Next

    The manager says yes

    Nothing sends until a person approves or edits it.

John checks in. The room already knows him.

A tablet at a sake bar showing a returning-guest alert with the guest’s past preferences and a suggested greeting
Illustration: the check-in alert on the bar’s tablet.
Check-in alert

Returning guest just checked in. Likes N-end (5/5) and Hizo Koshu. Follow up: a bottle for their sister in Portland, and the 12-person offsite in December.

A laptop showing an AI-drafted follow-up email with Edit and Approve buttons
Illustration: an AI-drafted follow-up, waiting for approval.
Draft follow-up · waiting for approval

“Hi John, lovely to see you again! Shall we check N-end shipping for your sister, and hold a December tasting for your team?”

From the pilot demo. John is a sample guest.

Why it matters for a Japanese brand in the US

A tasting-room manager reading an AI follow-up suggestion on a tablet
Illustration: the manager reviews the suggestion before anything is sent.

The same setup works wherever you pour or sample in the US: retail tastings, pop-ups and trade-show booths.

Want this for your tasting room or US launch?

Tell us how you meet US customers today. We’ll show you where a QR experience and AI follow-up would help, in Japanese or English.