Shoppers describe a look. Your search should find it.
“Black midi dress for a wedding, size 44” is one intent, not five keywords. AACSearch turns it into filters, ranks what is in stock in that size, and completes the outfit.
Queries plain keyword search gets wrong
чёрное платье миди на свадьбу 44five constraints, one intentкросовки найк белыеtypo + Cyrillic brandrjhcjdrb“кроссовки” in the wrong layoutdress like zara satinreference brandджинсы мом высокая посадкаstyle vocabularyпуховик до 10000price in wordsWhere search loses customers today
Sizes and colours are treated as words
“44” matches a product code, “black” matches a description that says “not black”. The shopper sees the wrong items and leaves.
Out-of-stock sizes on page one
The best-matching dress is sold out in the size the visitor needs. Ranking that ignores size availability wastes the most valuable positions.
Slang, brands and typos
“кросовки”, “найк”, a query typed in the wrong keyboard layout — each one is a zero-result page unless someone curates synonyms by hand.
Visual intent with no words
Customers come with a photo or a reference brand. Keyword search has nothing to match against.
Single-item baskets
The visitor buys the dress and nothing else, because nothing showed what goes with it.
How AACSearch solves it, layer by layer
Each card names the capability behind it. Items marked “in development” are designed and scheduled, not live yet.
Finds the right item from the first letters — typos, synonyms, filters, meaning.
Attributes from the query
Size, colour and price phrases are recognised as constraints and applied as filters instead of competing as text.
Typos, slang and brand spellings
Typo tolerance, synonym sets and a zero-result ladder that relaxes the query step by step before giving up.
Hierarchical facets
Category › subcategory › style with live counts, size and colour facets that only show values still available.
Find by photo
Image search over product photos: upload a picture, get visually similar items from your catalogue.
Widens the basket: similar, frequently together, personal.
Similar items
Visually and semantically similar products for the item on screen — the fallback when a size is sold out.
Personal ranking
Repeat visitors see their preferred brands, fits and price band first, with a decaying history so last season does not dominate.
Merchandising for collections
Pin the new collection, bury end-of-line items, schedule campaign placements with start and end dates.
Answers questions a filter can't, citing your own catalogue and docs.
Fit and care answers
“Does this run small?”, “can it be machine-washed?” — answered from your product pages and size guides, with citations.
Learns how items, queries and topics connect — and leads the visitor to the goal.
Complete the outfit
Learns which items are bought together in one visit — shoes with the dress, a belt with the jeans — and suggests complements, not lookalikes.
Style missions
Will group the catalogue into occasions people actually shop for (“wedding guest”, “office capsule”) from behaviour, without manual tagging.
Wrong layout, fixed
A query typed in the wrong keyboard layout is detected and converted before search.
One visitor, from first query to goal
What the visitor does on the left, what AACSearch does on the right.
- 1
Orient
Types “платье на свадьбу” on the home page.
Shows dresses with live facets for length, colour and size, so the next click narrows the list.
Facets and filters - 2
Narrow
Adds “чёрное миди 44”.
Turns colour, length and size into filters and ranks items available in size 44 first.
Natural-language search - 3
Compare
Opens two similar dresses back and forth.
Shows close alternatives in stock in the same size.
Recommendations - 4
Ask
Asks whether the fabric stretches.
Answers from the product description and size guide, with a link to the source.
AI answers with citations - 5
Basket
Adds the dress to the cart.
Suggests shoes and a clutch that shoppers bought together with this dress.
Frequently bought together
Your first two weeks
A typical rollout for this industry. Most teams are live on day one and spend the rest of the time measuring and tuning.
Day 1
Connect the catalogue; size, colour and price fields become facets.
Day 2
Embed the widget; add synonyms for brand spellings from the first zero-result report.
Week 1
Pin the new collection, exclude end-of-line items, check “similar items” on sold-out products.
Week 2
Build a golden set from the top 50 queries and A/B test a ranking change.
What to measure
The numbers to judge a rollout by. Compare them with an A/B split or a before/after window on the same traffic.
- Zero-result rate
- The first thing slang, typos and wrong layouts break. Should fall in the first week.
- Search conversion
- Share of searches that end in a purchase — the headline number for a store.
- Items per order from search
- Shows whether complements actually reach the basket.
- Out-of-stock clicks
- Clicks on items unavailable in the chosen size; ranking by availability should push it down.
Connects to
- Shopify
- WordPress / WooCommerce
- 1C-Bitrix
- YML / CSV feed
- REST API
- Embed widget
Capabilities used
Coming next
Questions
Search for Fashion & apparel, live this week
Connect a feed or a site, embed one script, and measure the difference on your own traffic.