Which AI shopping startup is growing the fastest?
In our AI shopping market deck, you will find everything you need to understand the market
SUMMARY
Constructor is the fastest-growing AI shopping startup overall today, while Onton has the fastest demonstrated consumer-user growth.
The race is split between two very different businesses. Onton, Phia and Daydream are competing for the shopper directly, while Constructor is becoming infrastructure inside large retailers, so raw growth percentages cannot be compared mechanically.
Onton's consumer trajectory is still the hardest number to beat. Its monthly active audience rose from roughly 50,000 to more than 2 million, a roughly 40-fold increase achieved with only about $10 million of disclosed funding.
Phia has reached similar mass-market territory with around 1.5 million shoppers and thousands of direct brand partners. The affiliate-attribution controversy does not erase that adoption, but it makes historical sales and monetization claims much less useful for deciding whether Phia is actually ahead.
Daydream has quietly become a much stronger challenger. Its 1.5 million-plus shoppers now sit alongside 325 retailers, more than 10,000 brands and roughly 3 million products, while Powered by Daydream gives the company a second distribution channel directly through retailer websites.
Constructor has the strongest evidence that its growth is durable rather than a launch spike. Customer count rose 82%, EMEA revenue increased 116%, the platform handled 322 billion product-discovery interactions and gross revenue retention remained at 96%.
The retailer side of AI shopping is also producing cleaner economic proof than the consumer-app side. Constructor customers have reported measurable revenue, conversion and order-value improvements, while the newest consumer startups still disclose far less about the dollars generated by each additional shopper.
Bloomreach shows how large this category can eventually become rather than who is growing fastest. Its more than $260 million of ARR and rapidly increasing adoption of newer AI tools suggest that AI shopping infrastructure can support a very large software business once it becomes embedded in retailer workflows.
Consumers appear much more comfortable asking AI to find, filter and compare products than letting an autonomous agent make the purchase for them. That helps explain why discovery tools are scaling faster than fully agentic shopping products.
The longer-term risk is that conversational shopping becomes standard inside ChatGPT, Gemini, Amazon and other huge platforms. The startups with the best chance of surviving that shift are the ones that own retailer relationships, proprietary product intelligence, transaction data or distribution beyond a standalone chat interface.
This market map, featured in our AI shopping market deck, highlights top companies and startups in the AI shopping market
Why are AI shopping startups growing so fast right now?
AI shopping startups are growing unusually fast right now because consumers have started using conversational AI for real product discovery, while retailers are simultaneously replacing old search tools with AI systems that can lift sales.
The change is visible from several directions. Onton grew from roughly 50,000 monthly active users to more than 2 million. Phia says 1.5 million shoppers now use its fashion-shopping product. Daydream has also passed 1.5 million shoppers and recently expanded to 325 retailers, more than 10,000 brands and roughly 3 million products.
Retailers are moving at the same time. Constructor said its customer base grew 82% in its latest fiscal year while its platform handled 322 billion product-discovery interactions. Bloomreach, a much older commerce-software company, surpassed $260 million in annual recurring revenue and said the share of customers using one of its newer AI agents or AI tools more than doubled in a year.
Those numbers describe two markets growing together. Consumers are becoming comfortable asking AI what to buy, while merchants are becoming comfortable letting AI decide what products shoppers see.
That overlap is what makes the current AI-shopping wave different from the earlier generation of recommendation engines. AI shopping is becoming both a consumer interface and part of the retailer's core ecommerce infrastructure.
What does “fastest-growing AI shopping startup” actually mean?
For this comparison, the fastest-growing AI shopping startup should be the company expanding real usage and commercial activity fastest, rather than whichever startup can show the biggest funding round or the most flattering percentage.
That definition immediately creates a problem: Onton, Phia, Daydream and Constructor sell very different things.
Onton is mainly a consumer product-discovery platform, so monthly active users tell us a lot. Phia also owns the consumer relationship but monetizes through shopping activity and retailer partnerships. Daydream started as an AI fashion destination and is now selling its technology directly to retailers. Constructor mostly works behind the scenes for large ecommerce companies, where customer growth and retailer revenue matter far more than app downloads.
A 40-fold jump in users therefore cannot be compared mechanically with 82% enterprise customer growth. The starting bases, business models and value of each customer are completely different.
We can still make a useful judgment by asking three questions repeatedly: how quickly is adoption growing, how much commercial activity sits behind that adoption, and has the growth lasted long enough to look durable?
Funding and valuation help us understand investor expectations, but neither answers the title. Daydream raised $50 million before reaching today's scale. Onton got to more than 2 million monthly active users with roughly $10 million in disclosed funding. Phia reached a $185 million valuation extremely early. Those are interesting differences in capital efficiency and investor enthusiasm, yet actual operating growth remains the better test.
As this chart shows, and as featured in our AI shopping market deck, search interest in AI shopping has grown significantly
Is Onton the fastest-growing AI shopping app right now?
Onton has the strongest clean claim to fastest consumer growth today, because its reported monthly active users climbed from around 50,000 to more than 2 million.
That works out to roughly 40 times the earlier user base. More importantly, the company disclosed monthly active users rather than cumulative registrations, so we are measuring people who were actually using the product during a month.
The speed is striking. Forbes reported that Onton had already gone from 50,000 to more than 1 million monthly active users within roughly a year with a team of only four. Later reporting put the figure above 2 million.
Onton achieved most of that growth while concentrating heavily on furniture and home products. It is a fairly narrow starting point for a consumer shopping platform because people buy sofas, tables and lamps much less frequently than clothing or groceries.
The company is now pushing into apparel and eventually consumer electronics. That gives us the next useful test. If Onton can keep growing quickly after leaving furniture, its 2-million-user milestone will start to look like the beginning of a general shopping platform rather than the peak of a strong niche product.
We should still be careful with the word “fastest.” Onton has published much less information about revenue, transaction volume and repeat purchasing than Constructor. Its consumer growth is the clearest in the group, while its business growth remains harder to measure.
| AI shopping startup | Strongest recent consumer scale | What stands out |
|---|---|---|
| Onton | 2M+ monthly active users | Grew from ~50,000 MAUs |
| Phia | 1.5M shoppers | Reached mass-market scale in roughly a year |
| Daydream | 1.5M+ shoppers | Reached similar scale while building a large fashion catalog |
| Constructor | Enterprise product | Consumer MAUs are not the relevant metric |
Is Phia growing faster than Onton?
Phia looks stronger than Onton across the whole business, but Onton still has the cleaner evidence of faster consumer-user growth.
Phia's early expansion has been extremely quick. The company passed roughly 500,000 users within months of launch, moved beyond 1 million, and now says 1.5 million shoppers use the product. Its current App Store listing still describes Phia as trusted by 1.5 million shoppers and shows tens of thousands of U.S. ratings.
The commercial side expanded alongside that audience. Phia's merchant material currently advertises more than 5,000 direct brand partners, while its product searches span hundreds of millions of items across a far broader set of websites. The company says it drove more than $12 million in partner sales during one seven-month period.
That combination makes Phia more interesting than a consumer app growing on downloads alone. Users arrive to compare prices, find alternatives and shop across new and secondhand products, so monetization is built directly into the activity people come for.
There is one major reason we should be more cautious with Phia than we were a few months ago. Recent investigations found problems with affiliate attribution, including cookie behavior that could give Phia credit for purchases it had not actually originated. Phia removed the disputed functionality, said it would reverse affected transactions and strengthened its compliance process.
That is the awkward bit. The issue does not invalidate its user growth, but it makes historical sales and affiliate-revenue numbers less useful when we rank the company against rivals.
Phia currently has one of the strongest overall growth stories in AI shopping, while Onton's 50,000-to-2-million MAU trajectory remains the safer answer to the narrower question of who grew users fastest.
If you want more recent data on this point, please see our latest AI shopping market report.
This chart, featured in our AI shopping market deck, illustrates yearly VC funding for AI shopping startups
How fast is Daydream catching Phia in AI fashion shopping?
Daydream has caught Phia surprisingly closely on consumer scale and is now making a serious move into retailer software as well.
Daydream recently reported more than 1.5 million shoppers after coming out of beta. That puts its disclosed audience in roughly the same range as Phia's 1.5 million shoppers.
Its supply side has grown substantially too. Daydream now says its catalog covers 325 retailers, more than 10,000 brands and around 3 million products. Earlier versions of the service were already broad by fashion-startup standards, but the latest figures show a platform large enough to answer fairly specific fashion searches without relying on a tiny group of partner brands.
The more important development lately is Powered by Daydream. The company is taking the natural-language and visual-search technology built for Daydream and installing it directly on retailers' own websites.
Five retailers were already live when the program was announced, including STAUD and Alice + Olivia, while more than 25 additional brands and retailers had signed up.
That changes Daydream's growth ceiling. It can continue trying to attract consumers to Daydream itself while also earning money from retailers that already have their own traffic.
Constructor is years ahead in enterprise adoption, so Daydream has not caught it there. Against Phia, however, the gap looks much smaller now. Both have around 1.5 million reported shoppers, both sit across thousands of fashion brands, and both are trying to turn consumer shopping intent into retailer revenue.
Daydream currently has the cleaner recent operating story. Phia still appears further ahead on direct monetization, although the attribution controversy makes that advantage harder to quantify precisely.
Is Constructor actually growing faster than the consumer AI shopping apps?
Constructor has the strongest sustained commercial growth in AI shopping today, even though its percentages look less spectacular than Onton's consumer-user explosion.
Constructor said its customer base increased 82% in its latest fiscal year. The previous year, it had already reported customer growth of roughly 50%. Seeing the growth rate accelerate from an already substantial base is much more impressive than a one-off launch spike.
The scale underneath those customers has also kept expanding. Constructor processed 322 billion product-discovery interactions during the year, equivalent to more than 10,000 personalized shopping experiences per second. The company says that interaction count is 266% higher than two years earlier.
Europe has been especially strong. EMEA revenue increased 116% year over year, meaning that regional business more than doubled. Constructor also expanded global headcount by 45%.
Retention gives us an important check on the story. The company finished the year with 96% gross revenue retention. In plain English, it kept almost all of the revenue it already had before adding new customers.
Enterprise retailers do not switch search and product-discovery systems casually. The software sits close to conversion, merchandising and revenue, so replacements require integrations, testing and internal approval.
An 82% increase in customers under those conditions carries more economic weight than an app adding casual users. That is why Constructor gets the top spot overall.
| Constructor metric | Latest reported result | What it tells us |
|---|---|---|
| Customer growth | +82% | Enterprise adoption accelerated |
| EMEA revenue | +116% | Regional revenue more than doubled |
| Product-discovery interactions | 322B | Platform usage is already enormous |
| Two-year interaction growth | +266% | Usage expansion continued beyond one year |
| Gross revenue retention | 96% | Existing customer revenue largely stayed |
If you want more recent data on this point, please see our latest AI shopping market report.
This chart, featured in our AI shopping market deck, shows why Constructor is winning in AI shopping
Are Constructor's retailers actually making more money from its AI?
Constructor has unusually strong evidence that its growth translates into retailer sales, which is one reason we rank its expansion so highly.
Belk provides the clearest example. Constructor's work with the department-store chain initially produced roughly $21 million in measured incremental revenue. Continued optimization later pushed the cumulative figure to around $35 million. Belk also reported a 7.4% increase in revenue per visitor, while shoppers using Constructor's AI Shopping Agent converted at more than twice the normal rate.
Target Australia separately reported more than A$13 million in additional search revenue associated with Constructor. Rugs Direct reported a 16.5% increase in average order value.
We should not add those figures together and call the result Constructor revenue. Each retailer measured something different, over different periods and under different conditions.
The pattern is still unusually useful. Retail search software is close enough to the transaction that companies can run controlled tests and compare revenue, conversion or order value before and after deployment. Constructor has accumulated several examples where the improvement is large enough to matter financially.
That makes its 82% customer growth easier to understand. Retailers are buying a tool that can be evaluated in dollars rather than an experimental AI feature whose value is hard to pin down.
For the fastest-growing-business question, this evidence gives Constructor a big advantage over consumer startups that disclose audience numbers but little about what each additional user produces economically.
Could Bloomreach be growing faster than Constructor?
Bloomreach is much larger than Constructor, but the public numbers do not show faster overall growth. They show how big AI commerce software can become once it reaches maturity.
Bloomreach surpassed $260 million in annual recurring revenue and ended its latest reported year with record net-new ARR, its strongest quarter ever and positive free cash flow.
AI adoption inside that customer base is moving quickly. Nearly half of Bloomreach customers now use at least one newer agent or next-generation AI tool, more than twice the proportion a year earlier. The number of customers using four or more of those tools quadrupled.
During the holiday shopping period, Bloomreach also reported a 113% increase in shopper engagement with its conversational shopping agent.
Those are strong numbers, but Bloomreach serves more than 1,400 brands across a much broader personalization platform. It has been building commerce software for years and now sells marketing, search, personalization and agentic products.
Constructor is the cleaner comparison for a fast-growing AI-native shopping startup. Bloomreach is more useful as a scale reference: AI commerce is already supporting a private software company above $260 million ARR, which makes Constructor's enterprise trajectory look much less niche.
Bloomreach shows the size of the opportunity. Constructor has the faster startup growth curve.
This chart, featured in our AI shopping market deck, illustrates yearly funding for AI shopping startups
Are AI shoppers actually buying things, or are they just trying the tools?
AI shopping is producing real purchases now, although the best evidence still comes from retailer deployments rather than the headline user counts of consumer apps.
Constructor's customer results already show tens of millions of dollars in measured incremental retailer revenue. Lily AI gives us another useful comparison from a different part of the shopping stack.
In four separate 28-day Google Shopping tests carried out for a luxury fashion house, a major U.S. department store, a global home retailer and an athletic-wear brand, Lily AI reported roughly $80 million in combined incremental revenue impact. The individual tests included an 8% revenue lift worth about $22 million, a 15% improvement in return on ad spend associated with roughly $50 million of additional revenue, and a 5% revenue lift worth around $8 million.
These are vendor-reported tests, so we should treat them as company evidence rather than an independent market measurement. Even with that caveat, they illustrate why retailers are moving quickly. Tiny improvements in product matching can become large dollar amounts when they are applied to enormous ecommerce catalogs and advertising budgets.
The consumer side is harder to measure cleanly. Phia advertises more than $12 million in partner sales during seven months, but the recent attribution problems mean we should not rely heavily on historical affiliate numbers. Daydream has disclosed audience and retailer growth much more clearly than transaction value. Onton provides excellent active-user data but limited public revenue information.
People are buying through AI-assisted shopping experiences. We simply have better proof of the economic impact inside retailers than inside the newest consumer apps.
| Example | Reported commercial impact |
|---|---|
| Constructor at Belk | ~$35M cumulative incremental revenue |
| Constructor at Target Australia | A$13M+ additional search revenue |
| Lily AI, four 28-day tests | ~$80M combined incremental revenue impact |
| Phia | $12M+ partner sales claimed over seven months, with attribution caveat |
Are people ready to let AI choose what they buy?
Consumers are ready to use AI for shopping help today, but full trust in AI buying decisions is still much weaker than the user-growth headlines suggest.
That gap explains why the fastest-growing products currently focus on helping shoppers search, compare and decide.
Onton lets shoppers describe or visualize what they want. Phia compares prices and surfaces alternatives across new and secondhand products. Daydream turns natural-language descriptions into fashion searches. Constructor improves the products retailers show when a shopper searches or browses.
All four reduce friction without requiring someone to hand complete control of the purchase to an autonomous agent.
Survey evidence points in the same direction. Vogue Business found that general AI-chatbot usage was already common among fashion-conscious consumers, yet AI-specific fashion and beauty shopping remained far less universal. Trust in AI recommendations also lagged usage substantially.
The products asking consumers to make a small behavioral change are advancing faster than visions where an agent receives a budget and buys everything independently.
Right now, “find me better options” is a much easier product to grow than “buy this for me without asking.”
If you want more recent data on this point, please see our latest AI shopping market report.
This chart, featured in our AI shopping market deck, compares the main business model options for AI shopping assistants
Can Onton keep growing after furniture?
Onton's next move into apparel and electronics will tell us whether its remarkable 2-million-user scale can turn into a much larger shopping business.
Furniture gave Onton a smart entry point. Product discovery is genuinely difficult when someone cares about dimensions, material, style, price and how an object will look inside a room. Onton's image-generation and search tools directly attack that problem.
Apparel is much larger and purchased much more often, but competition is also brutal. Phia and Daydream are already focused on fashion, while ChatGPT, Google, Amazon and retailer-owned search systems increasingly handle conversational shopping queries. That is a rough category to enter.
Consumer electronics creates another challenge. Specifications are more structured, price comparison is already mature and retailers such as Amazon and Best Buy have huge product datasets.
Onton's expansion is therefore unusually revealing. If monthly active users keep climbing once the company enters these crowded categories, we will have much stronger evidence that people prefer the Onton interface itself.
If growth flattens, the earlier 40-fold increase will still have been impressive, but we will probably reinterpret it as exceptionally strong product-market fit in home discovery.
We do not have that answer yet. Among the major contenders in this article, Onton is the company where the next set of fresh user figures could most dramatically change our ranking.
Is Daydream building a better business than a standalone AI shopping app?
Daydream's move onto retailer websites gives it a more attractive business model than relying entirely on consumers remembering to start every fashion search on Daydream.
Consumer shopping destinations have a distribution problem. A shopper may love an AI fashion tool and still begin the next purchase on Google, Instagram, Amazon, a brand website or a general chatbot.
Powered by Daydream reduces that risk. The company can bring its conversational and visual-search technology to places where people already shop.
The initial numbers are still small compared with Constructor: five retailers live and more than 25 additional companies signed on. But the speed matters because the product was only recently introduced, and the list includes recognizable fashion brands.
Daydream also has something useful to bring into those deployments. Its consumer platform already spans 325 retailers, more than 10,000 brands and around 3 million products, while more than 1.5 million shoppers have interacted with its search experience.
That gives the company a credible chance to learn from both sides of the market. Consumer activity teaches Daydream how people describe fashion; retailer deployments give it distribution and a clearer path to recurring software revenue.
Constructor remains far ahead commercially. Daydream has nevertheless become one of the more interesting companies to watch because its latest expansion tackles one of the biggest weaknesses in consumer AI shopping: having to acquire the same shopper over and over again.
This chart, featured in our AI shopping market deck, shows how market revenue is split across customer segments in the AI shopping market
Does Phia's affiliate controversy change the growth ranking?
Phia's affiliate controversy lowers our confidence in its monetization numbers, but it does not erase the company's rapid consumer adoption.
The disputed behavior concerns who deserved credit for certain purchases.
Recent reporting described cases where affiliate cookies could be overwritten or inserted in ways that allowed Phia to receive commissions for transactions it had not truly originated. Further reporting raised questions about how early the founders knew about the behavior.
Phia has since removed the problematic features, committed to reimbursements or transaction reversals where appropriate, and said it tightened its compliance controls.
User metrics and attributed sales should therefore be treated separately.
There is no reason to infer from the attribution problem that 1.5 million reported shoppers disappear. The product still has substantial consumer adoption, a large catalog and thousands of direct brand relationships. The controversy mainly weakens the case for using affiliate sales as proof that Phia is monetizing those users better than every rival.
This changes the ranking at the margin. Phia is still growing very fast, but its historical monetization metrics no longer deserve the same weight.
If you want more recent data on this point, please see our latest AI shopping market report.
Are funding and valuation telling us who is actually winning AI shopping?
Funding is a surprisingly weak guide to who is growing fastest in AI shopping, and Onton versus Daydream makes that clear.
Daydream raised $50 million in seed financing before it had today's public consumer scale. The size of that round reflected the background of founder Julie Bornstein, the experience of the team and investor belief that AI could rebuild fashion search.
Daydream has since produced meaningful adoption, passing 1.5 million shoppers and building a catalog across 10,000-plus brands. The money therefore financed a real product rather than remaining a slide-deck story.
Onton reached more than 2 million monthly active users after raising roughly $10 million in total disclosed funding. On a crude capital-per-active-user basis, the contrast is enormous.
Phia followed another path. It reached mass consumer scale quickly and attracted a valuation around $185 million very early in its life. Investor interest rose with the consumer traction, but the recent affiliate issue is a reminder that valuation can move faster than the evidence underneath a business.
Constructor is much further along operationally and has spent years compounding enterprise adoption.
Funding is more useful here as a capital-efficiency check than as a scoreboard. Onton's growth looks especially striking because relatively little capital produced a large active audience. Daydream has converted a much larger initial investment into genuine scale. Constructor's strength comes from repeated commercial expansion rather than a recent financing headline.
This chart, featured in our AI shopping market deck, shows how AI shopping assistant technology has evolved over time
Can ChatGPT, Google and Amazon eventually crush these AI shopping startups?
General AI platforms are the biggest long-term threat to consumer AI shopping startups, because conversational product search is rapidly becoming a standard feature rather than something only startups can offer.
A shopper can already ask ChatGPT, Gemini or Amazon increasingly detailed questions about products. Those platforms start with distribution that Onton, Phia and Daydream cannot realistically match.
The startups therefore need advantages that survive after natural-language search becomes ordinary.
Onton is betting on better structured product understanding, visual exploration and ecommerce-specific search. Phia brings together price comparison, resale inventory, coupons and hundreds of millions of products. Daydream is going deep into fashion semantics and is now embedding its technology directly inside retailer websites.
Constructor has the safest position of the group. Large retailers need their own catalog data, merchandising rules, inventory, conversion objectives and shopper-behavior signals incorporated into the experience. A general chatbot can become another interface for shopping while Constructor still powers the ranking and product intelligence underneath it.
Bloomreach is already moving in that direction too. Its Loomi Connect product makes retailer search intelligence available through external AI interfaces instead of assuming shoppers will always stay on the merchant's own website.
The consumer startups still have a real platform risk. Their current growth is impressive, but the durable winners will probably be the companies that own proprietary shopping data, retailer relationships or transaction intelligence rather than simply a prettier chat interface.
Which AI shopping startup is growing the fastest today?
Constructor is the AI shopping startup we would call the strongest overall growth leader today, while Onton is clearly growing fastest on consumer usage and Daydream has become the most interesting fast-rising challenger.
Onton's growth remains the hardest consumer number to beat. Moving from roughly 50,000 to more than 2 million monthly active users means its active audience multiplied around 40 times. None of the other major independent AI-shopping startups has published a cleaner consumer-growth trajectory at that scale.
Phia has built a broader early business around roughly 1.5 million shoppers, thousands of direct brand partners and a huge searchable product universe. The affiliate-attribution controversy, however, forces us to discount some of the monetization evidence that previously made Phia look like the clearest all-around leader.
Daydream has strengthened its position lately. More than 1.5 million shoppers, 325 retailers, 10,000-plus brands and 3 million products put the consumer product at meaningful scale, while Powered by Daydream gives the company a second path through retailer websites. We would rank Daydream higher now than we would have based only on its original consumer launch.
Constructor still wins the broader question. Its latest year brought 82% customer growth, 116% EMEA revenue growth, 322 billion shopping interactions and 96% gross revenue retention. Those figures sit on top of several previous years of rapid expansion, while retailer case studies show measurable revenue gains reaching tens of millions of dollars.
That combination is unusually hard to match: fast new-customer growth, huge usage, strong retention and direct proof that customers make more money.
The hierarchy is fairly clear. Onton has the fastest demonstrated consumer-growth curve. Daydream currently has the strongest upward momentum among the fashion challengers. Phia remains one of the biggest consumer winners but deserves more skepticism around monetization. Constructor has the best evidence of sustained business growth at meaningful commercial scale.
If we have to choose one company for the title, Constructor is the answer today.
If you want more recent data on this point, please see our latest AI shopping market report.
In our AI shopping market deck, we identify pain points entrepreneurs should prioritize
OUR METHODOLOGY
This analysis asks which AI shopping startup is growing fastest by comparing the operating evidence available across consumer shopping products and enterprise commerce platforms. Because those businesses expose very different metrics, we assess adoption, commercial expansion, usage at scale, retention, economic impact and the persistence of growth rather than ranking companies on a single headline percentage.
For consumer businesses, we prioritize active-use figures over cumulative registrations when both are available. Onton's monthly active users are therefore particularly useful because they let us reconstruct a growth curve from roughly 50,000 to more than 2 million active users rather than relying on downloads or sign-ups alone. Phia and Daydream are assessed using their reported shopper counts alongside retailer, catalog and partnership expansion.
For enterprise businesses, we give more weight to customer growth, revenue expansion, platform usage and retention. Constructor's 82% customer growth, 116% EMEA revenue growth, 322 billion product-discovery interactions and 96% gross revenue retention are evaluated together rather than treating any one of those figures as decisive by itself.
We do not mechanically compare percentages across business models. A 40-fold increase in monthly active users and an 82% increase in enterprise customers represent different kinds of growth, so we consider the scale of the starting base, what each unit represents economically, whether other operating indicators are moving in the same direction and whether the growth appears to be continuing beyond an initial launch period.
Commercial impact is kept separate from company revenue. Retailer case studies from Constructor and Lily AI help show whether AI product discovery can produce measurable economic value, but incremental retailer sales, conversion gains and improvements in average order value are not treated as revenue earned by the AI vendor itself.
We also separate adoption from disputed monetization. Recent reporting about affiliate attribution at Phia reduces the weight we give to historical affiliate-sales evidence, but we do not automatically discard separate user-growth figures. The underlying question is whether a controversy affects the metric being used, rather than whether it should invalidate every operating number associated with the company.
Funding and valuation are supporting evidence only. Daydream's $50 million seed round, Onton's roughly $10 million of disclosed funding and Phia's reported $185 million valuation are useful for understanding investor expectations and capital efficiency, but none of them establishes who is growing fastest operationally.
The final ranking comes from assessing the recent evidence across these dimensions and looking for metrics that reinforce one another. We give the most weight to growth that is measurable, economically meaningful, sustained beyond a short launch spike and supported by more than one operating indicator.
Key sources include TechCrunch on Onton's financing, user growth and expansion beyond furniture, Forbes on Onton's earlier 50,000-to-1-million MAU trajectory, Phia's current product information, Phia's merchant and partnership figures, Phia's Series A announcement, TechCrunch's reporting on the Phia affiliate-attribution controversy, and TechCrunch's follow-up reporting on the issue.
We also use Vogue's analysis of affiliate attribution and Phia's response, Daydream's latest operating figures and Powered by Daydream launch, Vogue's reporting on Daydream's consumer scale, TechCrunch on Daydream's original $50 million seed financing, Constructor's latest customer, revenue, retention and platform-usage figures, and Constructor's retailer case studies.
Additional context comes from Bloomreach's $260 million-plus ARR and AI-adoption update, Bloomreach's conversational-shopping engagement data, Lily AI's reported Google Shopping tests, Vogue's consumer survey on AI-shopping adoption and trust, OpenAI on product discovery in ChatGPT, Google on its AI-powered shopping features, and Amazon on AI-assisted product research and shopping.
This chart, featured in our AI shopping market deck, shows how market revenue is split across Europe, Asia, North America, Africa, and South America in the AI shopping market
Related blog posts
- Which AI shopping startups are making the most money today?
