AI Shopping: what are the top startups now?

Is AI shopping finally becoming a real market?

AI shopping is already bringing real sales to retailers, and the strongest startups are now competing for a share of those purchases.

According to Adobe Analytics, traffic from AI platforms to US retail websites increased 138% year over year in May 2026. Those visitors converted 54% more frequently than non-AI referrals and generated 53% more revenue per visit. In July, AI referrals were still growing 62% year over year, and their conversion advantage had reached 60%, according to Adobe data reported by EMARKETER. AI referral growth and conversion rates are the type of market signal we revisit each quarter in our market report about AI shopping.

That's a remarkable change for a market that barely existed a few years ago. Yet fewer than 17% of US internet users use generative AI for shopping every week, according to EMARKETER. Most consumers still discover and purchase products through conventional retailers, marketplaces and search engines.

Meanwhile, ChatGPT, Google Gemini, Amazon's shopping assistant and Walmart's Sparky are expanding into product recommendations and purchasing. Independent startups are growing at the same time as some of the world's largest technology companies are moving into their market.

We now have enough evidence to identify serious AI shopping businesses. The harder question is which startups can keep their customers and make money once AI shopping becomes a standard feature everywhere.

What actually counts as an AI shopping startup?

The AI shopping market includes consumer shopping assistants, retailer search platforms and companies that supply product data to AI agents. Some of its biggest businesses barely interact with consumers directly.

Onton, Daydream and Phia build applications that help people discover products, compare prices or find clothes. Constructor, Bloomreach and Algolia sell technology that retailers use to improve search results, recommendations and product discovery. Channel3 and newer infrastructure companies help AI applications find and understand products across merchant catalogs.

These businesses make money in different ways. Consumer assistants can earn affiliate commissions when shoppers buy products. Enterprise platforms generally charge retailers for software, while infrastructure companies can charge developers and merchants for data access or related services.

For our ranking, we include independent companies with meaningful AI-powered shopping products, even when they existed before generative AI became popular. However, we distinguish those established businesses from newer AI-native startups.

That distinction is essential. Comparing a consumer app with two million monthly users against a company generating $260 million in annual recurring revenue tells us very little unless we know what each business actually does.

Which AI shopping startups are leading today?

Constructor, Bloomreach and Algolia currently lead AI shopping on established commercial scale, while Onton and Daydream are among the strongest newer consumer challengers.

Constructor has reported extraordinary growth in its retailer customer base. Bloomreach has disclosed more than $260 million in annual recurring revenue, and Algolia operates search infrastructure handling trillions of queries annually. Meanwhile, Onton has reported more than two million monthly active users, and Daydream has attracted over 1.5 million shoppers to its AI fashion platform.

These companies are winning in quite different markets. Constructor and Bloomreach help retailers sell more products through their existing websites. Algolia provides the search technology behind digital experiences. Onton and Daydream compete for consumers before they reach a retailer.

Our ranking puts more weight on demonstrated adoption, revenue, customer retention and competitive advantages than on funding alone.

Startup	Main segment	Evidence of scale	Current assessment
Constructor	Enterprise product discovery	322B shopping interactions in FY2026	Strongest enterprise growth case
Bloomreach	AI personalization and commerce	Over $260M ARR	Leading disclosed private revenue scale
Algolia	Search and retrieval infrastructure	Estimated $250M ARR; 1.75T annual queries	Major infrastructure leader
Onton	Consumer AI product search	Over 2M monthly active users reported	Standout consumer growth
Daydream	AI fashion discovery	Over 1.5M shoppers; 3M products	Leading specialized fashion challenger
Phia	Shopping and price comparison	Over 1M users; $43M+ raised	Major consumer player, with attribution concerns
Channel3	AI product-data infrastructure	More than 5,000 brand partners reported	Promising infrastructure contender
Alta	AI styling and digital wardrobes	$11M funding round	Differentiated fashion specialist
Gensmo	AI fashion styling	Over 1M downloads reported	Emerging consumer contender
ReFiBuy	AI catalog optimization	$13.6M seed round	Promising merchant-side specialist

Is Constructor the strongest AI shopping startup overall?

Constructor has one of the strongest claims to being the top AI shopping startup today, thanks to rapid customer growth and an enormous volume of shopping activity across retailers.

In March 2026, Constructor reported that its customer base had increased 82% during the fiscal year ending January 31. It also reported 322 billion shopping interactions and 96% gross revenue retention.

An 82% increase in customers over one year is unusually fast for an established enterprise software business. The 96% gross revenue retention figure is also encouraging: it indicates that Constructor kept approximately 96% of its recurring revenue from existing customers before counting expansion, based on the company's reporting definition. Who leads retailer-facing AI and who leads consumer apps is mapped on the market grids in our report on AI shopping.

The reported 322 billion shopping interactions work out to roughly 882 million per day over a 365-day year. These include interactions handled by Constructor's platform rather than completed purchases or unique shoppers, but the volume shows how deeply its technology is used across retail websites.

We particularly like the combination of fast customer acquisition and high retention. A startup can attract new clients through heavy discounts or aggressive marketing, but keeping almost all recurring revenue from existing accounts suggests retailers continue finding value in the product.

Constructor's main weakness in our ranking is financial transparency. Without a recent, directly comparable annual revenue figure, we cannot confidently say it is larger than Bloomreach or Algolia. Still, its operating growth makes it our strongest enterprise AI shopping contender.

Is Bloomreach actually bigger than the newer AI shopping startups?

Bloomreach is already much bigger in disclosed recurring revenue than the newer consumer AI shopping startups for which comparable financial information is available.

In February 2026, Bloomreach announced that it had exceeded $260 million in annual recurring revenue during 2025 and achieved positive free cash flow. The company reported more than 1,400 global brand customers, including Pandora and American Eagle.

Its existing customers are also adopting more AI tools. Nearly half were using at least one agent or next-generation AI capability, more than twice the proportion twelve months earlier. The number of customers using four or more such tools had quadrupled.

Compare that with Daydream and Onton, which must attract consumers and persuade them to keep returning. Bloomreach starts with companies that already pay for commerce software.

There is one important limitation: Bloomreach's $260 million ARR covers its broader software portfolio, including established products that existed before the current wave of generative AI.

Even with that distinction, Bloomreach is among the clearest commercial winners in AI-powered commerce today.

Could Algolia become the infrastructure winner in AI shopping?

Algolia is one of the best-positioned independent companies to power AI shopping behind the scenes, with an existing search business operating at a scale that younger infrastructure startups cannot easily match.

In September 2026, Algolia announced infrastructure connecting AI applications to live product catalogs through the Model Context Protocol. The company reported more than 1.75 trillion search queries annually and over 18,000 business customers.

That works out to approximately 4.8 billion queries per day, although the figure covers Algolia's broader search activities rather than shopping alone.

Sacra estimated Algolia's annual recurring revenue at $250 million in July 2026, compared with $230 million at the end of 2025 and $210 million in 2024. These figures suggest growth of about 9.5% from 2024 to 2025, followed by another 8.7% increase by July 2026.

Algolia is therefore a large, established software business growing at a much more moderate rate than the fastest consumer AI shopping applications.

Its biggest advantage is the underlying problem it solves. Shopping assistants need accurate prices, stock availability, sizes and product details. A language model can produce convincing recommendations, but it cannot reliably determine whether a particular pair of shoes is available in size 42 without access to current inventory information.

Algolia already provides search and retrieval systems to thousands of businesses. If more shopping moves into AI assistants, its existing infrastructure could become even more useful.

We consider Algolia one of the strongest AI shopping infrastructure companies, although it has not disclosed how much incremental revenue agentic commerce is generating.

Is Onton the fastest-growing consumer AI shopping startup?

Onton has one of the most impressive reported user-growth records among consumer AI shopping startups, expanding from approximately 50,000 to more than two million monthly active users.

TechCrunch reported those figures in November 2025, when Onton, formerly Deft, announced a $7.5 million financing led by Footwork. The company initially focused on furniture before expanding into other shopping categories.

The increase represents at least 40-fold growth, or roughly 3,900% from the starting user base. Few independent shopping assistants have publicly reported a comparable jump in monthly active users.

Onton's financing is also relatively modest. Daydream raised $50 million in seed capital, while Phia has announced more than $43 million in total funding. Onton reached its reported audience milestone with considerably less disclosed capital.

That comparison doesn't prove Onton acquired users more efficiently, because we lack directly comparable marketing spending and acquisition-cost figures. Still, it shows how quickly a focused shopping product can attract attention without raising one of the category's largest funding rounds.

The missing evidence is retention and revenue. Onton's two million monthly active users were a company-reported historical milestone, and we don't have a newer independently verified figure or reliable data showing how many users buy products through its recommendations.

For now, Onton is a clear consumer growth leader, but we cannot yet put it ahead of established enterprise companies on commercial performance.

Is Daydream becoming the leader in AI fashion shopping?

Daydream is currently one of the strongest independent AI fashion-shopping startups, combining more than 1.5 million reported shoppers with a catalog of approximately three million products.

Founded by former Stitch Fix executive Julie Bornstein, Daydream secured $50 million in seed funding before launching publicly in June 2025. Its assistant lets users describe the clothes they want using natural language, including preferred styles, brands, occasions and budgets. Shopping by describing an outfit in plain words is new, and we check how far it has got in our report on AI shopping.

According to TechCrunch's September 2026 reporting, Daydream had attracted more than 1.5 million shoppers and covered over 325 retailers and 10,000 brands. Participating retailers and brands included Uniqlo, Reformation, Gucci and Nordstrom.

Three million products across 325 retailers equal approximately 9,200 products per retailer on average, although actual product coverage varies considerably. More importantly, Daydream gives shoppers access to multiple brands in one place rather than restricting recommendations to a single merchant's catalog.

The company has also added shopping through saved iPhone photos and Siri integrations using Apple's newer developer capabilities. These features fit fashion shopping particularly well because consumers often begin with an image or a vague idea of the style they want.

Compared with Phia's price-comparison focus, Daydream is more interested in helping people discover what suits them. Onton offers broader product exploration, while Daydream concentrates on fashion.

We rank Daydream among the leading consumer AI shopping startups, especially in clothing and accessories. However, its shopper count and product catalog still tell us little about repeat purchases, revenue or profitability.

Is Phia still a top AI shopping startup after its controversy?

Phia remains one of the biggest consumer AI shopping startups by reported users and funding, but its affiliate-attribution controversy makes its commercial claims harder to trust.

Founded by Phoebe Gates and Sophia Kianni, Phia launched in April 2025. In January 2026, the company announced a $35 million Series A at a $185 million valuation, bringing its total reported funding above $43 million.

Phia reported more than one million users and partnerships with over 6,200 retail brands. Its product helps shoppers compare prices, find alternatives and discover deals across new and secondhand fashion.

Those numbers would normally place Phia comfortably among the leading consumer shopping assistants. However, Bloomberg investigated cases in which the company's browser extension inserted affiliate tracking that could give Phia credit for purchases it had not actually generated.

Affiliate network Impact.com suspended the company. Phia attributed the behavior to a software issue that it said it corrected. Further reporting in August raised questions about when company leadership became aware of the practice.

The controversy goes directly to Phia's business model. Affiliate commissions depend on correctly attributing purchases to the source that generated them. If that attribution is inaccurate, commission revenue can overstate how much value a shopping assistant creates for retailers.

We still include Phia among the leading startups because its audience and fundraising are substantial. But we wouldn't treat its attributed transaction activity as reliable proof of commercial success without stronger independent evidence.

Which AI shopping startups actually make the most money?

Bloomreach has the strongest directly disclosed recurring-revenue figure among our leading private AI shopping companies, while Algolia follows closely based on third-party estimates.

Bloomreach reported more than $260 million ARR and positive free cash flow. Sacra estimated Algolia at approximately $250 million ARR in July 2026. We should give Bloomreach's company-disclosed figure and Algolia's external estimate different levels of confidence.

Coveo offers another useful comparison. Although it is publicly traded rather than a private startup, the enterprise search and AI company reported $148.3 million in fiscal 2026 revenue, compared with $133.3 million the previous year.

That's approximately 11% annual growth. Coveo also reported $142.5 million in SaaS subscription revenue and a 78% gross margin, giving us a clearer picture of the economics of a mature search and AI software business.

By contrast, Onton, Daydream and Phia have not disclosed standardized revenue figures that allow a direct comparison. Their reported users, shoppers and retail partnerships demonstrate reach, but we cannot reliably convert those figures into annual sales or profits.

Constructor is a particularly interesting case because its operational metrics are impressive, yet the absence of comparable revenue disclosure prevents us from ranking it confidently by company size.

Company	Latest financial evidence	Evidence type	Important limitation
Bloomreach	Over $260M ARR	Company disclosure	Includes broader commerce and marketing software
Algolia	Approximately $250M ARR	Sacra estimate	Not independently audited public revenue
Coveo	$148.3M annual revenue	Public financial reporting	Public company; broader enterprise search
Constructor	Revenue not reliably disclosed	Operational metrics	Interaction volume is not revenue
Daydream	Revenue not disclosed	Shopper and catalog metrics	Purchase conversion unknown
Onton	Revenue not disclosed	Monthly active users	Monetization and retention unknown
Phia	Revenue not disclosed in comparable form	User, funding and attributed-sales claims	Affiliate-attribution controversy
Channel3	Revenue not disclosed	Brand-partner footprint	Paid usage and margins unknown
Alta	Revenue not disclosed	Financing and product partnerships	Insufficient commercial scale evidence

Are people actually buying more because of AI shopping?

AI shopping is now bringing retailers visitors who convert at higher rates than ordinary referrals, although we still cannot measure how many of those purchases would have happened anyway.

Adobe's conversion data shows a particularly interesting change. In earlier observations, visitors referred by AI platforms converted worse than other retail visitors. By March 2026, they converted 42% better, and by July that advantage had reached 60%.

That's a substantial improvement. A shopping assistant might send millions of curious visitors to retailers without generating many purchases. Higher conversion rates suggest that shoppers increasingly arrive after doing part of their research through AI.

We should still be careful about what those numbers prove. Someone planning to buy a laptop might ask ChatGPT for recommendations before completing a purchase they already intended to make. That transaction can appear as an AI referral even if the assistant mainly changed the route to checkout.

According to Attentive's 2026 consumer research, only 10% of purchasers completed their transactions within AI during the shopping experience examined, while 60% of AI-assisted shoppers preferred buying through the retailer's website or app.

For startups, the immediate opportunity is therefore substantial even without controlling payments. Onton and Daydream can influence what shoppers consider, while Constructor and Bloomreach help retailers convert that interest into purchases.

The unresolved question is incremental sales: how many additional purchases actually happen because consumers used these AI tools? We have encouraging conversion data, but the public evidence is still too limited to answer that precisely.

Can Channel3 compete with Google and Shopify in AI shopping?

Channel3 has a credible opportunity in AI shopping infrastructure, but competing with Google and Shopify on sheer product-catalog size would be extremely difficult.

Channel3, a Y Combinator S2025 company, is building product search and data infrastructure for developers creating shopping applications. Its technology focuses on matching products across merchants, identifying variants and supporting commerce monetization. Young infrastructure startups face giant rivals, and we rate that and other risks by likelihood in our AI shopping market report.

The company has reported more than 5,000 brand partners. That suggests meaningful product coverage, although partnerships do not necessarily represent paying customers.

And the competition is enormous. Google maintains a Shopping Graph containing tens of billions of product listings. Shopify has introduced catalog infrastructure for developers and worked with Google on the Universal Commerce Protocol, allowing AI agents and merchants to communicate across purchasing workflows.

Those companies already have access to vast merchant networks. Matching their coverage would require substantial investment, and startups may struggle to offer information that is consistently fresher or more complete.

Channel3's opportunity is more specific: independent developers may want product-matching tools, standardized information and flexible access without building directly on one large platform's shopping experience.

ReFiBuy is approaching a related problem from the merchant side. The startup raised $13.6 million in seed financing to help retailers make their product catalogs more useful and discoverable through AI-powered shopping systems.

We would watch both companies for evidence of paying customers and recurring revenue. Product-data infrastructure is a promising market, but reported catalog coverage alone won't tell us which startup can turn it into a durable business.

Are Alta and Gensmo serious competitors in AI fashion shopping?

Alta and Gensmo are credible emerging AI fashion startups, especially because their products give shoppers reasons to return even when they aren't ready to buy something.

Alta raised $11 million in financing led by Menlo Ventures in 2025. Its application allows users to manage digital wardrobes, create outfits and experiment with styling through virtual avatars.

In February 2026, TechCrunch reported Alta's partnership with fashion brand Public School, bringing its styling capabilities into retail websites. That gives Alta a possible route into enterprise fashion software alongside its consumer application.

Gensmo has focused on AI styling and virtual try-on, with company-reported downloads exceeding one million. This is meaningful early interest, although a million downloads could represent very different levels of ongoing usage depending on retention.

Their approach differs from Phia's deal discovery and Daydream's conversational product search. Someone may open Alta to put together an outfit using clothes they already own. Gensmo similarly aims to make styling and visualization part of the shopping experience.

That could help both companies build repeat engagement. It could also make monetization harder if users enjoy styling their existing wardrobes without purchasing new items.

For now, we place Alta and Gensmo below Daydream and Onton in the consumer ranking because the available evidence on sustained adoption and revenue is weaker. Both remain worth watching, particularly if fashion retailers begin paying to integrate their technology.

Are investors still funding AI shopping startups?

Investors are still backing AI shopping startups, but recent funding data suggests they're writing smaller checks than before.

A September 2026 market analysis by New Market Pitch counted 32 AI shopping financing transactions in the latest twelve-month period, compared with 31 in the previous period. Yet disclosed funding dropped from at least $291.2 million to $181.8 million.

That means the number of transactions increased approximately 3%, while the amount invested declined almost 38%. The median disclosed round also fell from $4.5 million to $2.91 million, a reduction of roughly 35%.

The same analysis found no recent round above $50 million, compared with four rounds exceeding $20 million in the preceding period. The figures cover the companies we tracked and their disclosed financings, so they should be treated as a useful market sample rather than a complete accounting of all investment.

There are still meaningful deals. Phia announced its $35 million Series A, ReFiBuy raised $13.6 million, and smaller startups such as Nudge, Naratix and Wassist have attracted early-stage capital.

The contrast with Daydream's $50 million seed round is interesting. Investors previously made some unusually large early bets on consumer AI shopping platforms. More recent activity shows continuing interest spread across narrower products and smaller financing rounds.

We don't see evidence that investors have abandoned AI shopping. But the category is no longer attracting the same concentration of very large early checks within the reported sample.

Will ChatGPT, Gemini and Amazon crush independent AI shopping startups?

ChatGPT, Gemini and Amazon are the biggest long-term threats to general-purpose AI shopping startups, particularly those offering product recommendations that consumers can already get from their existing assistants.

OpenAI has introduced shopping and merchant checkout capabilities into ChatGPT. Google and Shopify jointly developed the Universal Commerce Protocol, while Amazon continues improving shopping assistance within its existing retail ecosystem.

These companies have enormous distribution advantages. A consumer who already uses ChatGPT for work or Google for search may find it convenient to ask for product recommendations in the same interface. Amazon also combines recommendations with reviews, purchasing history, delivery and returns.

That makes life harder for standalone applications whose main feature is recommending products through a chatbot. Frankly, a generic shopping chatbot is becoming a difficult product to defend.

The outlook is different for more specialized services. Daydream can focus deeply on fashion preferences and cross-brand discovery. Alta offers digital wardrobes and outfit creation. Onton has built a visual shopping experience that users may prefer for exploring unfamiliar products.

Enterprise companies such as Constructor, Bloomreach and Algolia face another kind of competition. Retailers will still need search quality, merchandising controls, current inventory and ways to improve conversions on their own websites.

The question for those providers is whether retailers continue buying independent software as Amazon, Google, Shopify and other large vendors add similar capabilities.

We expect generic consumer assistants to face the toughest competitive pressure. Startups offering specialized experiences or deeply integrated retail technology have a better chance of keeping their place. Specialized or deeply integrated startups may hold their place, and we study what sets the winners apart in our market report about AI shopping.

Which AI shopping startups have something competitors can't easily copy?

Constructor, Bloomreach and Algolia currently have the clearest competitive advantages in AI shopping because their technology is already integrated into businesses that depend on it.

Constructor's 96% gross revenue retention and rapid customer growth suggest that retailers generally continue using its platform. Bloomreach has relationships with more than 1,400 brands and is getting existing customers to adopt additional AI capabilities. Algolia serves over 18,000 businesses and operates infrastructure processing trillions of search queries annually.

These advantages come from established customer relationships, operational experience and integration work. A competing startup can build a similar interface, but convincing thousands of retailers to replace working infrastructure is much harder.

Consumer startups have a different challenge. Onton has demonstrated rapid audience growth, while Daydream has assembled a large fashion catalog and attracted more than 1.5 million shoppers. Both need to turn those early advantages into habits that survive competition from larger platforms.

Phia's case illustrates another difficulty: a shopping app must prove that retailers benefit from the purchases attributed to it. Large audiences are valuable, but merchant trust is just as important when affiliate commissions support the business.

For now, we have stronger evidence of defensibility among enterprise providers than among newer consumer applications. Daydream, Onton and Alta could build lasting advantages, but we still need better retention, revenue and transaction data to know whether that is happening.

So, what are the top AI shopping startups now?

Constructor, Bloomreach and Algolia are the strongest established AI shopping companies today, while Onton and Daydream stand out among newer consumer startups.

Our overall leader is Constructor for its combination of 82% annual customer growth, 322 billion shopping interactions and 96% gross revenue retention. Bloomreach leads on directly disclosed private-company revenue with more than $260 million ARR, while Algolia is exceptionally well placed in search and product-data infrastructure.

For consumer AI shopping, Onton's reported expansion to more than two million monthly active users makes it a standout growth story. Daydream looks particularly strong in fashion, with over 1.5 million shoppers and approximately three million products. Phia remains an important player, although its attribution controversy prevents us from treating its commercial claims with the same confidence.

Channel3 and ReFiBuy are interesting infrastructure challengers, while Alta and Gensmo offer more specialized fashion experiences. Their current commercial evidence is too limited to put them alongside the largest players.

The wider market is moving quickly. Adobe's retail data shows that AI-referred visitors are becoming more valuable to merchants, even as financing rounds have become smaller and large technology platforms are entering the same market.

We therefore favor the startups that already have paying retailers, deep integrations and measurable commercial results. The newer consumer companies may ultimately produce a breakout winner, but none has yet demonstrated the same combination of revenue scale and customer durability.

AI shopping is already a real business. Today, the strongest companies are mostly those helping retailers sell products more effectively, while the race to become consumers' favorite independent AI shopping assistant remains wide open.