Data study · Patents behind AI search
Conversational Shopping Patent Filings Are Flat Since 2018, but 43.3% of Those Published in 2026 Use Generative AI
We found 536 patent families on product recommendations in chat and assistant settings. Yearly counts barely moved (74 in 2018, 75 in 2025), while the generative AI share went from 5.7% in 2023 to 43.3% in 2026 so far.
Analysis of 536 conversational shopping patent families, 59 using generative AI · Google Patents Public Data, October 2026
Key findings
- 1
536 patent families published since 2018 describe product recommendations or shopping inside a conversation, chatbot or assistant.
- 2
Yearly publications are flat: 74 in 2018 and 75 in 2025, a 1.4% change.
- 3
The generative AI share rose from 3 of 53 (5.7%) in 2023 to 25 of 75 (33.3%) in 2025 and 13 of 30 (43.3%) in 2026 so far.
- 4
151 of the 536 (28.2%) describe personalization or user history, and 47 (8.8%) describe adding to a cart or completing an order.
- 5
299 families (55.8%) include a US publication and 239 (44.6%) a Chinese one. Microsoft (20), IBM (15) and Amazon (11) hold the most.
- 6
A patent shows what was filed, not what is live in any product.
Summary
Picture a shopper typing into a chat window: I need hiking boots for wide feet under $150. A few product cards come back with a short explanation. Conversational shopping like this has been patented for years, and the record shows a shift in how it is built rather than how often it is filed.
We found 536 patent families since 2018 that describe shopping or product recommendations in a conversation. The yearly count barely changed, from 74 in 2018 to 75 in 2025. The method did change. In 2026 so far, 13 of 30 such families (43.3%) describe generative AI, compared with 3 of 53 in 2023.
Personalization is the most common feature, in 151 families. A patent shows what was filed, not what is live in any product.
What we measured
We pulled every publication in Google Patents Public Data (patents-public-data.patents.publications) published from January 1, 2018 to September 17, 2026 that carries at least one CPC code in G06F16 (information retrieval), G06F40 (natural language processing) or G06N (AI and machine learning) and has an English abstract. That is 2,170,848 publications in 1,214,491 patent families. For each publication we kept the English title, the English abstract and up to the first 20,000 characters of the English claims. English claims exist for 488,406 publications, mostly US, European and international (WO) filings, so text matching is deeper for those offices than for Chinese, Korean or Japanese filings, which we matched on title and abstract.
A family counts as conversational shopping when its text mentions product or item recommendations, shopping, e-commerce, an online store, a shopping cart or purchasing a product, and also a conversation, chatbot, dialog, chat, virtual or digital assistant, or a natural language request. It counts as generative when the text also names generative AI (large language models, LLMs, generative AI or generative models, GPT, foundation models or pre-trained language models). Features come from keyword groups in the same text.
Filings are flat, but the method is changing
We counted conversational shopping families by first publication year and the subset that describes generative AI.
Families
| Year | All conversational shopping families | With generative AI | Share (%) |
|---|---|---|---|
| 2,018 | 74 | 0 | 0 |
| 2,019 | 54 | 0 | 0 |
| 2,020 | 50 | 0 | 0 |
| 2,021 | 79 | 0 | 0 |
| 2,022 | 53 | 0 | 0 |
| 2,023 | 53 | 3 | 5.7 |
| 2,024 | 68 | 18 | 26.5 |
| 2,025 | 75 | 25 | 33.3 |
| 2,026 | 30 | 13 | 43.3 |
Source: aisearchoptimization.net analysis of 536 conversational shopping patent families, 59 using generative AI · Google Patents Public Data, October 2026
The 2026 total covers only January 1 to September 17. By first filing year the count ranges from 43 to 74 a year between 2019 and 2024, so there is no clear growth in volume.
Personalization leads the feature list
A family can describe several features. We counted them for all 536 families and for the 59 generative ones.
Families
| Feature | All 536 families | Share of 536 (%) | Generative 59 families |
|---|---|---|---|
| Personalization or user history | 151 | 28.2 | 21 |
| Cart, checkout or placing an order | 47 | 8.8 | 3 |
| Inventory or availability | 38 | 7.1 | 5 |
| Price | 36 | 6.7 | 7 |
| Purchase history | 23 | 4.3 | 2 |
| Product images | 14 | 2.6 | 3 |
| Clarifying or follow-up questions | 4 | 0.7 | 3 |
| Product comparison | 3 | 0.6 | 2 |
Source: aisearchoptimization.net analysis of 536 conversational shopping patent families, 59 using generative AI · Google Patents Public Data, October 2026
Checkout and stock questions come up far more than side-by-side product comparison, which appears in only 3 families. Among the 59 generative families, 21 (35.6%) describe personalization.
Who files conversational shopping patents
Owners and publishing offices for the 536 families:
| Owner or office | Families |
|---|---|
| Microsoft | 20 |
| IBM | 15 |
| Amazon | 11 |
| eBay | 9 |
| 8 | |
| Include a US publication | 299 |
| Include a Chinese publication | 239 |
| Include a Korean publication | 49 |
Source: aisearchoptimization.net analysis of 536 conversational shopping patent families, 59 using generative AI · Google Patents Public Data, October 2026
Ownership is spread out. Besides large platforms, insurers and a number of small commerce software firms appear with a handful of families each.
What this means
If a shopping assistant recommends products from a description of need, your product pages have to describe needs as well as specs. Fit, use case, sizing and who the product is for give a conversational system something to match. Rewriting product pages around those details is work an AI search optimization agency can do without touching your specs.
Keep prices and stock status current on the page and in your product feeds, since checkout and availability appear often in these filings. Our free AI visibility audit shows whether AI assistants already recommend your products, which is the clearest test of AI brand visibility for a store.
Limits
A patent shows what was filed, not what is live in any product.
Patents usually publish about 18 months after the first filing, so families first filed in 2025 and 2026 are heavily undercounted, and 2026 publication counts cover January 1 to September 17, 2026 only.
Only families tagged G06F16, G06F40 or G06N are included. Commerce patents classified only under business-method codes are not counted, so the true number of conversational shopping filings is higher.
Feature groups are keyword based and broad. 'Availability' can refer to things other than stock.
Owner names come from harmonized assignee data; individual owners are not named.
Embed this research
Add the charts from this study to your page. The embed links back here.
<iframe src="https://aisearchoptimization.net/embed/data-studies/conversational-shopping-patents" title="Chat Shopping Patents: 43.3% Now Use Generative AI" width="100%" height="720" style="border:0" loading="lazy"></iframe>
Cite this study
APA
Grenyo, K. (2026). Chat Shopping Patents: 43.3% Now Use Generative AI. AI Search Optimization. https://aisearchoptimization.net/resources/data-studies/conversational-shopping-patents
Plain link
Chat Shopping Patents: 43.3% Now Use Generative AI - AI Search Optimization - https://aisearchoptimization.net/resources/data-studies/conversational-shopping-patents
Frequently asked questions
Are conversational shopping patents growing?
Not in volume. We found 74 families published in 2018 and 75 in 2025. What grew is the generative AI share, from 5.7% in 2023 to 43.3% in 2026 so far.
What features do these patents describe?
Personalization or user history appears in 151 of 536 families (28.2%). Cart or checkout appears in 47 and inventory or availability in 38.
Who files the most?
Microsoft holds 20 families, IBM 15 and Amazon 11. A patent shows what was filed, not what is live in any product.
What should an online store do?
Describe who each product is for and keep price and stock data current. Those details match the features most often named in these filings.
Related studies
- Patents behind AI searchFreshness Shows Up in Only 3.6% of AI Ranking Patents
- Patents behind AI searchQuery Fan-Out Patents: 44.1% Now Use Generative AI
- Patents behind AI searchSearch Patents 2026: 39.9% Now Mention Generative AI
- Patents behind AI searchAI Content Detection Patents Up 237.5% Since 2022
See where AI search leaves you out
Run the free AI visibility audit. If you want the fixes done for you, the plan is a flat $299/month.