Signals Wants to Reach Customers Before They Ever File a Support Ticket

Most customer support software is built around a reactive loop: a user runs into a problem, opens a ticket, waits in a queue, and eventually gets an answer. Signals, a Y Combinator Winter 2026 startup, is flipping that model on its head. The company is building an AI customer engagement platform that engages high-value customers before they reach out – answering questions they haven’t yet asked, smoothing over friction points, and opening the door to upsells in the same motion.

Signals uses large language models to initiate authentic conversations with customers at the moments they’re most likely to need help. By catching issues early, the company argues, brands can reduce inbound support volume, deepen customer relationships, and drive more revenue from the user base they already have. The company claims to be the only AI agent on the market with engagement quality high enough to be approved for outbound iMessages – a particularly notable distinction given Apple’s traditionally strict policies around business-initiated messaging on the platform.

The company was founded in 2025 by Ilya Valmianski and Alejandro Zaniolo. Valmianski, the co-founder and CEO, is a two-time founder with a physics PhD and a deep background in applied AI for consumer-facing products. He previously built AI-powered patient intake at Kaiser Permanente, developed a medical ChatGPT-style assistant at Curai Health before ChatGPT was a household name, and co-founded MDandMe. Zaniolo brings operational chops to the team: he has helped take two startups from zero to one, one of which was acquired by Blueground, and another of which grew six times year over year to $40 million in ARR and reached profitability. The pairing gives Signals an unusual combination of AI research depth and early-stage go-to-market experience.

The company is wagering that the economics of customer support are about to change dramatically. Support teams already use AI to deflect tickets and speed up response times, but those tools still sit downstream of the customer’s decision to ask for help. Signals argues that the bigger opportunity is upstream – catching confused users in the middle of onboarding, proactively reaching out to enterprise buyers who are about to churn, or nudging power users toward plan upgrades that fit their actual usage. When these interactions feel natural rather than promotional, they can double as retention, satisfaction, and revenue tools all at once.

Signals is part of Y Combinator’s Winter 2026 batch and is currently operating as a lean two-person team. For companies that live and die by customer retention – SaaS vendors, fintechs, digital health platforms, and consumer marketplaces – the pitch is straightforward: stop waiting for tickets to roll in, and start reaching out with context about who the customer is and what they’re trying to accomplish.

If Signals can deliver on its promise of authentic outbound engagement at scale, its AI customer engagement platform could redefine what customer support software looks like for the AI era – shifting the default from reactive ticketing to proactive customer intelligence. Interested teams can learn more at returnsignals.com.