How to use this page:
Find your selected Jira issues in the table below. Select the expand to use them as your source of truth to write release notes.
|
Release |
https://cheitgroup.atlassian.net/projects/SCLB/versions/11277 |
|
Date |
|
|
Version |
26.08 Release |
|
Description |
|
|
Contributors |
How to use this page:
Find your selected Jira issues in the table below. Select the expand to use them as your source of truth to write release notes.
|
Release |
https://cheitgroup.atlassian.net/projects/SCLB/versions/11344 |
|
Date |
|
|
Version |
18.08 Release |
|
Description |
|
|
Contributors |
We want non-EU customers (e.g., Switzerland, UK, US) to successfully complete checkout for exotic-leather items by shipping to an EU address (e.g., France). The flow should guide them clearly, reduce friction, and increase conversion when EU shipping is possible.
Current struggle
At the moment, non-EU customers who want to buy exotic-leather bags face a confusing checkout flow. After the pop-up, they must first tick “Ship to a different address,” manually complete the shipping section, and then go back to re-enter their billing details. This creates unnecessary friction, increases the chance of abandonment, and lowers the likelihood that they complete the purchase even if EU shipping would actually be possible.
As a customer outside the EU,
I want a clear prompt and a guided path to enter an EU shipping address (while keeping my non-EU billing address),
so that I can place the order successfully when shipping to the EU is allowed.
|
# |
Scenario / Gherkin |
Notes / Visual |
|---|---|---|
|
01 |
Scenario: Pop-up appears for exotic leather + non-EU billing Given the cart contains at least one exotic-leather item When the user lands on the checkout Then a pop-up appears with title “Please note,”… |
already existing
|
|
02 |
Scenario: “Change shipping address” guides the user Given the pop-up is shown When the user clicks Change shipping address Then the system leaves the non-EU billing country in the billing address dropdown |
Billing country remains unchanged (non-EU).https://app.box.com/s/owoeqh8m72ts5l7bm5ccm3qm51oqdz2d https://app.box.com/s/929u6pom5e02cxgk1tlxlbeon7tcaus7 |
|
03 |
Scenario: Non-EU countries disabled in Shipping Given an exotic-leather item is in the cart When the user opens Shipping › Country / Region Then only EU countries are selectable |
Keep full list for clarity; disable instead of removing. |
|
04 |
Scenario: Server validation enforces EU shipping Given an exotic-leather item is in the cart When the user clicks “Place order” Then the order is blocked with the inline error |
function already existing |
|
05 |
Scenario: Billing outside EU is allowed Given an exotic-leather item is in the cart When the user completes checkout with an EU shipping country Then the order proceeds normally. |
function already existing |
|
06 |
Scenario: Mixed carts Given the cart contains any exotic-leather item When the user checks out Then the EU-only shipping rule applies to the entire order. |
function already existing |
|
07 |
Scenario: “Shop other bags” exits flow Given the pop-up is shown When the user clicks Shop other bags Then the user is redirected to the bags listing page. |
function already existing
|
Do you also want me to add a “Visual design” section (like in your other Confluence stories) with the pop-up copy and inline helper text summarized, so design and dev have it in one place?
User Story
As an admin, when I edit a page and leave all fields of a pop-up empty, the pop-up should not be displayed at all.
Scenario: Admin leaves all fields in a pop-up empty
Given I am logged in as an admin
And I edit a page that has a pop-up (e.g saclab.com/sell)
When I leave all fields of the pop-up empty
And I save the page
Then the pop-up should not be shown on the page
How to use this page:
Find your selected Jira issues in the table below. Select the expand to use them as your source of truth to write release notes.
|
Release |
https://cheitgroup.atlassian.net/projects/SCLB/versions/11309 |
|
Date |
|
|
Version |
12.08 Release |
|
Description |
|
|
Contributors |
How to use this page:
Find your selected Jira issues in the table below. Select the expand to use them as your source of truth to write release notes.
|
Release |
https://cheitgroup.atlassian.net/projects/SCLB/versions/11276 |
|
Date |
|
|
Version |
07.08 Release |
|
Description |
|
|
Contributors |
Users enter a free-text Product Title (e.g. “Hermès Birkin 30 Etoupe Gold”). We use rule-based logic against our Airtable lists to auto-fill Brand, Model+Size, Color, and Material fields. We prioritize exact matches – no fuzzy corrections. Any attribute not matched stays blank for manual selection; all dropdowns include an Other option.
Our Airtable source includes multilingual variants (English, German, French) from our Weglot data, so titles in any of these languages map back to a single English canonical value.
This story covers the parsing logic only; dropdown UI behavior is in a separate story.
Exact Matching with Exploration
Normalize title (lowercase, strip punctuation, normalize accents).
By default, match only full tokens or exact phrases from Airtable.
Explore optional high-confidence fuzzy or proximity matching (e.g., Levenshtein distance) for edge cases – but avoid guessing at the cost of accuracy.
Attribute Dependencies
Model depends on detected Brand.
Size depends on detected Model (e.g., Birkin→sizes 25,30,35).
Color scoped to Brand (brand‑specific names first, then generic).
Material scoped to Brand (Hermès leathers vs. generic types).
Hardware parsing: detect terms like “Palladium”, “Gold‑plated”, “Gold Tone” – but be cautious since “Gold” may refer to either color or metal finish.
Translations & Canonical Values
Airtable stores EN/DE/FR variants from Weglot; match any variant and save as the English canonical term.
Other Fallback
If an attribute isn’t matched (or match confidence is low), leave it blank and let the user select manually. Always display Other as dropdown option.
Analytics & Continuous Improvement
Log every title and parser result to an analytics table.
Review common misses and update Airtable data or matching logic over time.
User Story
As a seller,
I want the system to auto-fill form fields from my title,
So that I can list my bag faster and with fewer clicks.
Desktop & Mobile:
Step 1: Title input + Continue button. (will be new design)
Step 2: Detailed form with: (existing design)
Pre-filled fields for matched attributes
Blank dropdowns (with Other) for unmatched ones
Clear notice: “We’ve auto-filled these fields based on your title. Please review.”
|
# |
Scenario |
|---|---|
|
01 |
Given a title with “Hermès” When parsed Then Brand = Hermès |
|
02 |
Given title includes “Birkin 30” When parsed Then Model/Size = Birkin 30 and Brand=Hermes |
|
03 |
Given title includes “Etoupe” When parsed Then Color = Etoupe and Brand=Hermes, as Etoupe is only a Hermes Color |
|
04 |
Given title includes “Togo” When parsed Then Material = Togo |
|
05 |
Given title in German (“Hermès Birkin 30 Weiß”) When parsed Then Color = White (English) |
|
06 |
Given title term not in lists When parsed Then corresponding field remains blank and Other is available |
|
07 |
Given title contains model but no brand When model uniquely matches a brand Then Brand inferred by model |
|
08 |
Given parsing runs When completed Then record title + matched results to analytics for later review |
Future: Evaluate AI-enhanced parsing once we have sufficient analytics data.
We are exploring a radically simplified flow for registering bags using AI. The idea is to let users upload images first (instead of filling dropdowns manually) and receive suggested bag data based on the uploaded content.
Once images are uploaded, an AI process is triggered in the background. After approximately 10 seconds, the AI returns model suggestions, brand, size, material, and potentially serial number information. These suggestions are injected into our form fields as pre-filled values, which the user can then confirm or correct.
The AI request may fail or be incomplete based on image quality or missing views. In those cases, the user receives a simple message asking them to improve their upload. We do not flag individual empty fields — we simply recommend better image quality or more images.
The long-term goal is to reduce user friction, speed up submissions, and match bags accurately with our internal dropdown data. This story is focused on the AI-based upload flow and interaction, not on dropdown logic or data validation — that is handled in a separate story.
Estimated request time: ~10 seconds per set of images. Costs and tokens per request should be evaluated.
As a user
I want to upload photos of my bag and get suggestions automatically
So that I don’t have to manually fill out long dropdown forms
Example User Images:
Example Prompt:
You are an expert in luxury handbags, especially Hermès, Chanel, Dior, and Bottega Veneta. You will be shown a set of 3–6 images of a single handbag.
Your task is to identify the handbag and return your best structured guess in the following JSON format:
{
"brand": "",
"brand_confidence": 0.0,
"model": "",
"model_confidence": 0.0,
"size": "",
"size_confidence": 0.0,
"material": "",
"material_confidence": 0.0,
"hardware": "",
"hardware_confidence": 0.0,
"color": "",
"color_confidence": 0.0,
"serial_stamp_location": "",
"stamp_confidence": 0.0,
"notices": []
}
Confidence values must range from 0.0 to 1.0 and reflect your certainty in each prediction.
Only fill in values where you are reasonably confident (e.g. above 0.5). If unsure, leave the value blank but still include a confidence score (e.g. 0.2). Avoid guesses.
Model names should follow structured naming like:
- `"Birkin 25"`
- `"Picotin Lock 18"`
- `"Classic Flap Medium"`
- `"Saddle Bag"`
For Hermès, if the stamp is visible, describe its location (e.g. “Interior base”) and include that in `serial_stamp_location`.
The output is meant to pre-fill dropdown fields for a resale listing. It should be short, structured, and help reduce manual work.
If any part of the bag isn’t clearly visible (e.g. hardware is covered, interior is dark), include a note in `notices`.
Return only the JSON and nothing else.
Example AI output gpt 4o vision:
{
"brand": "Hermès",
"model": "Picotin Lock",
"size": "18",
"combined_model": "Picotin Lock 18",
"material": "Clemence",
"color": "Craie",
"hardware": "Gold",
"stamp_or_serial_hint": "Interior embossing visible (HERMÈS PARIS MADE IN FRANCE)",
"confidence_scores": {
"brand": 1.0,
"model": 0.95,
"size": 0.9,
"material": 0.85,
"color": 0.8,
"hardware": 0.95
},
"feedback": "If any detail seems incorrect, you can update it manually before submission."
Desktop:
Full-width upload box with drag-and-drop & preview
Feedback animation and result step
Mobile:
Large tap zone to add files
Loader animation and compact result fields
|
# |
Gherkin scenario |
|---|---|
|
01 |
Given a user is on the bag registration page |
|
02 |
Given images are uploaded |
|
03 |
Given AI analysis takes up to 10 seconds |
|
04 |
Given the user receives pre-filled results |
|
05 |
Given confidence is too low or image quality is poor |
|
06 |
Given brand/model/size are suggested |
|
07 |
Given color/hardware/material are suggested |
|
08 |
Given a serial number or stamp appears in image |
|
09 |
Given the user continues to the next step |
|
10 |
Given image upload comes first |
AI response format should be JSON, containing keys for: brand, model, size, hardware, color, material, optionally serial_number, and a confidence_score
Match to internal dropdown values must be exact or fuzzy (e.g. Birkin 30 vs. Birkin 30cm)
No user-generated values here — the system always suggests known internal matches, or asks the user to improve image quality
Expected model: GPT-4 Vision or similar (test in playground if needed)
Cost & latency TBD (est. ~10 seconds per request, ~2000–3000 tokens, priced by OpenAI plan)
Future exploration: combine vision with live web lookup or product matching — not yet in scope
Optimize dropdowns on the Sell Your Bag form by enhancing existing inline dropdowns:
Search-as-you-type within dropdowns; matches highlighted in bold (e.g., typing “handle” shows Timeless Handle)
Substring matching functionality – match any part of the option text
Option ordering: brand-specific entries first, then general entries, and always leave Other at bottom (remove redundant “I don’t know” option)
Replace “I don’t know” with “Other” and keep existing function to always show it when there is no match
Cascading logic (already in place): Model+Size filtered by Brand; Color and Material filtered by Brand
Optional: Combine Model+Size into one dropdown (entries like “Birkin”, “Birkin 25”, “Birkin 30”)
New Hardware dropdown with brand-specific finishes (Palladium, Gold Tone, Silver Tone, Ruthenium)
As a seller,
I want dropdowns that filter instantly and highlight matches,
So that I can fill fields quickly without unnecessary scrolling or extra options.
Tap a dropdown field—list expands inline beneath the input.
Type in the search box at top of dropdown; keyboard opens only when typing.
Options filter by substring; matching portion of each entry is bolded for clarity.
Ordering: brand-specific options appear first, followed by general options, then Other.
Hardware field: brand-specific finishes listed.
(Optional) Model+Size field combines model and size entries, e.g., “Birkin”, “Birkin 25”, “Birkin 30”.
Select an option; dropdown closes and field populates.
|
# |
Scenario |
|---|---|
|
01 |
Given I type “handle” in any dropdown When filtering Then “Timeless Handle” appears with handle bolded |
|
06 |
Given no options match searc When filtering Then only Other remains visible |
Future: Use analytics to refine option lists and consider AI enhancements.
We’re overhauling the Sell Your Bag flow to boost submissions, reduce friction, and prepare for future AI‑enhanced matching. This Epic covers three parallel stories:
49.1 Updated Dropdown Enhancements
Improve existing dropdowns with real‑time substring search, bolded matches, proper ordering, and a new hardware field.
49.2 Smart‑Match A: AI Image Upload & Data Extraction
Radical variant: user uploads bag images first, AI analyzes photos (brand, model, size, serial) and pre‑fills form fields for quick confirmation.
49.3 Smart‑Match B: Title Parser & Auto‑Fill
Rule‑based variant: user types a single free‑text product title; the system parses and auto‑sets Brand, Model+Size, Color, Material.
We can implement either Smart‑Match A or B – or combine both in future – with dropdown fallbacks to ensure price calculations and user confirmations remain accurate.
As a bag owner
I want a fast, intuitive listing process – whether through improved dropdowns, AI‑assisted image parsing, or smart title parsing –
So that I can register or sell my bag with minimal effort and high confidence in the details.
Dropdown enhancements are live, reducing average form‑completion time by ≥30%. Note: Can we track that in GA4?
Smart‑Match A: AI prototype processes images and pre‑fills ≥70% of fields correctly in user testing.
Smart‑Match B prototype parses free‑text titles and pre‑fills ≥70% of fields correctly in user testing.
Any unmatched fields default to the enhanced dropdown UI (Story 49.1).
All relevant UX changes validated on both desktop and iOS mobile.
(Data from both Smart‑Match variants logs mismatches for analytics and iterative improvement.)
Notes & future exploration:
We must retain dropdown fallbacks to support price-calculator dependencies.
Recording user interactions and mis‑matches will guide AI‑enhanced refinements.