Key takeaways
- Ecommerce site search is built for shoppers who already know what they are looking for.
- It struggles with vague needs, fit and compatibility, comparisons and words your catalogue doesn’t use.
- AI site search improves the list of results; AI chat can ask a follow-up question, explain and recommend.
- The two work together: the search bar for shoppers who know, a conversation for the ones who don’t.
What is ecommerce site search?
Ecommerce site search is the search bar on an online store: the shopper types a word or phrase and gets a list of matching products. Most site search engines match the query against product names, descriptions and attributes, rank the results and add autocomplete and filters on the side.
It does one job very well: getting a shopper who knows what they want to that product fast. A model number, a brand, a product name, a code copied from a review – search finds them in a second, and filters narrow the list from there.
Where a search bar runs out
The trouble starts when the shopper doesn’t know the right word – or knows exactly what they need but not which product delivers it. Four kinds of query keep ending in the wrong results, or none at all:
- Vague needs – “a gift for a 7-year-old who likes science”, “something warm for autumn hiking”. No product is called that, so the results are random or empty.
- Fit and compatibility – “will this water filter fit my coffee machine?”, “which screen protector fits my phone?” The answer depends on two products at once, and a list of results can’t confirm it.
- Comparisons – “which of these two is quieter?” Search shows both; the shopper still has to read two spec sheets.
- Words your catalogue doesn’t use – synonyms, slang, typos, another language. The shopper types “couch”, the catalogue says “sofa”, and the page says “No results found”.
A no-results page is the most expensive dead end in a store. The shopper has told you exactly what they want, and the site has answered that you don’t have it – often when you do. And these questions aren’t rare: across the stores AskSpot runs on, about half of chat conversations are product questions like these.
AI site search vs AI chat: what’s the difference?
“AI site search” usually means a smarter search bar. Instead of matching exact words, it matches meaning, so “couch” finds sofas – which fixes a lot of the vocabulary problem. It is still a search bar, though: one query in, one list out. AI chat works differently. It can ask what the shopper means, compare products and explain the difference.
| Keyword site search | AI site search | AI chat | |
|---|---|---|---|
| What the shopper types | Keywords, names, codes | Natural phrases | A question in their own words – or a photo |
| Typos and synonyms | Often missed | Usually understood | Understood |
| Vague needs | Random or empty results | Better guesses | Asks a follow-up question, then recommends |
| Fit and compatibility | Can’t confirm | Can’t confirm | Answers from the product data |
| Comparisons | Lists both products | Lists both products | Explains the difference and points to one |
| What comes back | A list | A better-ranked list | A recommendation, with product cards and add to cart |
| Best for | Shoppers who know what they want | The same, with fewer dead ends | Shoppers who need help choosing |
AskSpot is the third column: an AI chat agent for e-commerce that searches your own catalogue from inside the conversation. It doesn’t replace your search bar – it picks up where the search bar stops.
Where the conversation takes over
The handover from search to conversation should happen at the moment the search bar runs out – not after the shopper has given up. There are four places it can happen.
1. On the no-results page
When search comes back empty, the chat opens with a short proactive message – “Try searching with me” – and lets the shopper describe what they need in their own words. The dead end becomes a second chance.
2. On the search results page
A shopper scrolling through a long list gets the same offer – “Need help finding a product?” A sentence or two about what they need turns a page of results into a short list of the products that fit.
3. With a photo
Some searches can’t be put into words at all. The shopper attaches a photo in the chat and the agent finds matching or look-alike products in your catalogue.
4. On the page itself: the AI sets the filters
Some stores go one step further. Instead of listing results in the chat, the agent works on the store’s own page: it opens the right category, ticks the filters and points out the best product on the list, narrating each step so the shopper only has to confirm. It is a dedicated integration built for each store – see it on morele.net in The AI assistant that clicks for your customer.
In the chat, every route ends the same way: product cards from your catalogue with the price and an add-to-cart button, so the shopper can decide without leaving the conversation. More on how it works on the AI product search page.
spotSee it on your own catalogue
Book a short demo – we’ll run the Agent on searches from your store.
What happens when the AI recommends a product
A good answer is one the shopper acts on. In published AskSpot case studies, shoppers click the product the agent recommends in most of the conversations where it recommends one:
78%
of chats with a recommendation end in a click on the suggested shoe
AskSpot Case Studies
57%
of chats with a recommendation end in a click on one of the picks
AskSpot Case Studies
21.6%
of chats end in a purchase
AskSpot Case Studies
The first figure comes from Casu, a Polish footwear house, where most questions are about size and fit. The other two come from Konesso, a coffee retailer whose shoppers ask which filter or part fits their machine – and where more than 1 in 5 chats ends in an order.
How to improve your ecommerce site search
Before you add anything, fix the basics. They help every shopper who uses the search bar:
- Track zero-result searches. Export them every week – they are a list of things shoppers want and can’t find.
- Add synonyms and handle typos. Map the words shoppers use to the words in your catalogue: “couch” to “sofa”, a brand name misspelled three ways.
- Make attributes searchable, not just names. If shoppers can filter by size, colour, material or compatibility, they should be able to search for it too.
- Build filters around how people shop. Use, room or occasion often matters more to the shopper than the attributes in your product feed.
- Read your chat conversations. In chat, shoppers describe what they need in full sentences – the quickest way to find the words your catalogue is missing. Conversation analytics groups them by topic for you.
- Offer a conversation at the dead ends. Put a chat prompt on the no-results and results pages, so a search that fails turns into a question instead of an exit.
AskSpot has a 7-day free trial with no card, so you can connect your product feed and test it on the searches your store already gets.
FAQ
What is ecommerce site search?
It is the search bar on an online store. It matches what the shopper types against product names, descriptions and attributes and returns a ranked list, usually with autocomplete and filters. It works best for shoppers who already know what they want.
What is AI site search?
A search bar that matches meaning rather than exact words, so synonyms and natural phrases find the right products. It improves the list of results, but it still returns a list – it doesn’t ask questions, compare products or explain the difference.
Can AI chat replace site search?
No – they do different jobs. Search is fastest for shoppers who know the product; AI chat helps the ones who need to describe a need, check fit or compare. Most stores need both, with the chat offered where the search bar runs out.
How do I reduce zero-result searches?
Track them every week, add synonyms and typo handling, make product attributes searchable, and offer a chat on the no-results page so the shopper can describe the product in their own words.
Can an AI chat set the filters on my store’s pages?
Yes, with a hands-on assistant: it opens categories, ticks filters and adds products to the cart on the page itself, narrating each step so the shopper only confirms. It is set up as a dedicated integration for each store.
