Signals wants a spot in your customers' text messages

The number that started it is not a slogan. Roughly $300 billion of clothing, shoes, and accessories bought online every year gets boxed back up and shipped to the warehouse it came from. For most brands, that reverse trip is a cost line - refunds processed, inventory restocked, an email that says "sorry it didn't work out." Signals, a Winter 2026 Y Combinator company, looked at the same event and saw something else: a customer trying to tell you something, in the only moment they were actually paying attention.

The company's pitch fits in a sentence. When a package lands on a doorstep, Signals sends a text. Not a marketing blast - a two-way iMessage that reads like a note from a store clerk. "How did the jacket fit?" If the answer is "a little tight," the software does not open a return. It offers the next size up, in the same color, arriving Thursday. The refund never happens. The customer stays.

01 / WHAT IT DOESA conversation, not a returns portal

Legally the company is Material Model, Inc. Publicly it launched as Return Signals and now goes by the shorter Signals. The distinction matters less than the product, which is a clienteling platform - retail's word for the salesperson who remembers your name, your size, and the thing you almost bought last time. That person built loyalty in stores for a century and never scaled past one shop floor. Signals is an attempt to rebuild them in software, and to put them where people actually read messages.

Everything runs inside a single iMessage thread. A pre-purchase question ("does this run small?"), a post-delivery check-in, a reorder six weeks later - one continuous conversation instead of three disconnected channels. The AI handles the routine exchanges on its own. When something unusual comes up, it hands off to a human without dropping the context, so the customer never has to repeat themselves. It connects to the tools a brand already runs - Shopify for orders, Gorgias and Zendesk for support - so the data stays in one place.

02 / THE PROBLEMThe dead zone after "card charged"

There is a stretch of the customer relationship that almost no online brand touches: the days between the moment payment clears and the moment the box arrives. In a store, that gap does not exist - you walk out with the bag. Online, it is silence. The next contact a customer often gets is a shipping notification and, if things go wrong, a returns form. Signals argues that this silence is where returns are born. A buyer who feels forgotten sends the jacket back. A buyer who gets a well-timed text feels handled.

Figures reported by Signals from early deployments. Treat as company-provided, not independently audited.

03 / WHO IT'S FORBrands that sell things people try on

The natural customer is a direct-to-consumer brand in apparel, footwear, or accessories - categories where fit is uncertain and returns are highest. Two are named publicly. Jordan Craig, a streetwear label, ran an A/B test and reported that each Signals conversation was worth an extra $30 after two weeks. Quaker Marine Supply Co., a coastal outfitter, framed it in older language.

The Jordan Craig quote is the one that will get repeated in pitch meetings, because it puts a dollar figure on a fuzzy idea. "What stands out is that the growth is both incremental and profitable," said Rob Varon, the brand's VP of Digital. "In an A/B test we saw that every Signals conversation made us an extra $30 after 2 weeks." Incremental and profitable are the two words every retention tool claims and few can prove in a controlled test.

04 / HOW IT'S DIFFERENTProactive and two-way, by design

There is no shortage of software in this neighborhood. Returns platforms like Loop and Narvar handle the mechanics once a customer has already decided to send something back. SMS marketing tools like Attentive and Postscript push one-way promotional blasts. Support suites like Gorgias and Zendesk wait for a ticket to arrive. Signals sits at an angle to all of them: it reaches out first, it expects a reply, and its goal is revenue, not a resolved complaint.

Capability	Returns portals	SMS marketing	Signals
Reaches out before a problem	No	Yes (promo)	Yes
Two-way conversation	Limited	Limited	Yes
Converts return → exchange	Sometimes	No	Core focus
Human escalation w/ context	No	No	Yes
Goal	Process return	Drive clicks	Repeat revenue

05 / THE PEOPLEA physics PhD and a retail kid

The two founders arrive from opposite ends of the problem. Ilya Valmianski, the CEO, holds a physics PhD from UC San Diego and spent his first act in medical AI - building AI patient intake at Kaiser Permanente, "medical ChatGPT before ChatGPT" at Curai Health, and co-founding MDandMe. Along the way he helped build one of the largest clinically deployed patient-facing chatbots in the country. Getting a nervous patient to answer honestly over text turns out to be useful training for getting a shopper to say the shirt didn't fit.

Alejandro Zaniolo brings the other half. His family operated apparel retail across more than 200 brick-and-mortar locations, which means he grew up inside the exact dynamic Signals is trying to digitize: the store clerk who makes a customer, not a sale. One founder knows how conversation AI behaves under real load; the other knows what good clienteling actually felt like before the internet flattened it.

06 / THE BUSINESSPaid by the revenue it creates

Signals sells to brands as software and ties its value to numbers a CFO recognizes: incremental repeat revenue, margin uplift of 5 to 10 percentage points, and a lower net return rate. That framing is deliberate. A returns tool is a cost center a brand tolerates; a revenue channel is a line item a brand wants to grow. By measuring itself in dollars added rather than tickets closed, Signals is arguing to sit on the growth side of the ledger.

There is also a quieter product underneath the messaging. Signals layers analytics with agentic search over the conversations it runs, so a brand can query thousands of chats at once - which product runs small, which complaint keeps recurring, which nudge actually lands. The messages create the revenue; the accumulated conversations create the data. For now the company is two people in San Francisco, with a $130K seed and a YC W26 stamp, which is to say the thesis is further along than the company.

07 / WHERE IT FITSRetention, on the expensive end of the funnel

Signals is placing a bet that has been slowly winning across e-commerce for a decade: that keeping a customer is worth more than finding a new one. Paid acquisition has gotten more expensive as ad platforms crowd and privacy rules tighten, which pushes brands to squeeze more lifetime value out of buyers they already have. Retention tools - subscriptions, loyalty programs, win-back flows - have multiplied as a result. What most of them share is a one-way posture: the brand pushes an offer and hopes. Signals is trying to occupy the two-way, high-intent slice of that market, the moments where a customer is already engaged and a single reply can change the outcome.

The channel choice is not incidental. Email open rates hover in the low double digits; a text from a number a customer recognizes gets read within minutes. That immediacy is why SMS marketing grew fast - and also why it burned out, as inboxes filled with coupon spam and unsubscribe rates climbed. Signals is threading a needle here: use the highest-attention channel, but reserve it for messages a customer might actually want, timed to a real event in their order rather than a marketing calendar. Whether brands can hold that discipline as they scale is the open question, and it is the same question every high-signal channel eventually faces.

08 / THE BETWill this matter in a decade?

Strip away the batch and the metrics and Signals is a wager on a simple claim: that the personal, remembered, conversational relationship shoppers had with good local retailers is not gone, just un-built online, and that the phone in everyone's pocket is where it gets rebuilt. If that is right, the interesting part is not the returns-to-exchange conversion - it is the standing thread. A brand that owns a live message channel to its best customers owns something durable, harder to disrupt than an ad account and stickier than an email list. If it is wrong, Signals will be one more well-timed nudge in a channel people learn to ignore. The two-week A/B test at Jordan Craig is a small piece of evidence for the first story. The rest is still being written, one blue bubble at a time.