The Y Combinator S24 startup builds omni-channel agents that chase leads across SMS, voice, email, and webchat with one shared memory. Its bet: the follow-up that companies keep dropping is worth a lot of money.
Every company sitting on a customer database has a graveyard inside it. It goes by polite names - "cold pipeline," "aged leads," "no-response" - but everyone knows what it really is. Someone raised a hand, asked for a quote, filled out a form at 11pm, and then the follow-up never came, or came three times and stopped. The lead didn't die. The company just quit texting back. Apten, a startup from Y Combinator's Summer 2024 batch, was built on the unglamorous conviction that this graveyard is full of money.
Apten builds omni-channel AI agents that engage and nurture leads across four channels - SMS, voice, email, and webchat - and, crucially, remember the whole conversation as it moves between them. Text a lead in the morning, call them in the afternoon, and the voice on the phone already knows what the text said. For businesses that live and die by lead conversion, that continuity is the product.
B2C industries - home services, insurance, financial services, education, healthcare - generate leads in enormous volume and convert them badly. A roofing company or a university admissions office might buy tens of thousands of inbound inquiries a month, then hand them to a call center, a dialer, and a canned SMS drip that all operate on separate wavelengths. The dialer doesn't know what the text said. The text doesn't know the call went to voicemail. The lead gets contacted at the wrong time, in the wrong tone, on the wrong channel, and eventually stops answering entirely.
Apten's founders, Daniel Ho and Roshan Kumaraswamy, came at this from the operations side rather than the sales side. Ho was a product manager at Tesla building systems to automate supplier issues, with earlier stints doing supply-chain analytics at Amazon, Lucid Motors, and Walmart. Kumaraswamy spent three years at AWS building sales-planning applications. Both studied at UC Berkeley - Ho in economics and data science, Kumaraswamy in EECS. They saw a workflow problem, not a charisma problem.
The mechanics are simpler to picture than to build. A lead comes in. Apten's agent handles inbound and outbound texts and calls around the clock, follows up on a cadence tuned to how the lead actually responds rather than a fixed template, and pulls context from the company's CRM and knowledge base. If the moment calls for a phone conversation, the agent dials - using local-presence numbers that match the lead's area code, which Apten says lifts answer rates by up to 40%. When a human should take over, the agent transfers the call live, cold or warm or scheduled.
Around that core sit the pieces enterprises actually ask about before signing. Message guardrails vet every outgoing message against customizable rules before it sends. Automatic tagging classifies conversations by sentiment and intent. Native integrations connect to Salesforce, HubSpot, Microsoft Dynamics 365, and Close, and an enterprise API lets customers push and pull lead context programmatically. The company is SOC 2 Type II certified and supports HIPAA-sensitive workflows - table stakes for selling into healthcare and finance, and the reason Apten leads with compliance rather than treating it as an afterthought.
The AI sales-agent field is crowded, and Apten knows it. Its argument against the pack is narrow and specific: a robocall or a static drip trains people to hang up and unsubscribe, while a real back-and-forth earns a few more seconds of attention. Apten claims conversational calls convert several times higher than robocalls. Whether the exact multiple holds across industries is the kind of number worth auditing, but the underlying logic - responsiveness and memory beat volume - is hard to argue with.
There's a second, quieter edge: unglamorous focus. Apten isn't chasing the flashy consumer chatbot demo. It's building for the roofing installers, the community colleges, the insurance brokers - businesses where a two-week deployment and a 5% lift in booked appointments is a real budget line, not a press release. That is a less romantic market than "AI that sounds human," and a stickier one.
Point Apten at a pile of aged or no-response leads, connect your CRM, and let it re-open conversations over text and phone - qualifying, answering questions, booking appointments, and handing live calls to your reps only when it counts. For inbound, it answers instantly, 24/7, so the 11pm form fill gets a reply before the lead cools.
Ex-Tesla product manager (supplier-issue automation); supply-chain analytics at Amazon, Lucid, and Walmart. UC Berkeley, economics and data science.
Three years at AWS building sales-planning applications; early work on the Flux language at InfluxData. UC Berkeley, EECS.
The team is small - roughly four people as of its YC batch - and the approach shows it: white-glove onboarding, custom configuration per customer, and a promise of production in about two weeks. Y Combinator backed the company in its S24 cohort under partner Tom Blomfield. Reported early funding is modest, in line with a company still proving out a wedge rather than spraying spend.
Zoom out and Apten sits at the intersection of three things every B2C enterprise already pays for: the dialer, the SMS marketing tool, and the SDR headcount. Its pitch is to collapse them into one agent that behaves consistently across channels and never lets a thread go cold. The competition includes other conversational-AI and AI-SDR platforms - names like Regal, Air, Bland, Retell, and PolyAI - plus the incumbents Apten is really replacing: manual call centers and disconnected automation.
The risk in this category is sameness; a lot of companies can demo an AI that talks. Apten's answer is to compete on the seams - the memory that carries from text to call, the guardrails that make it safe for regulated buyers, the two-week install that gets a skeptical operator to yes. Those aren't the parts that make a good demo. They're the parts that make a renewal. For a company selling to industries that measure everything in cost-per-acquisition, that may be the more durable bet.
None of this is settled. The conversion multiples are the company's own, the funding is early, and the field is loud. But the core observation Apten is built on - that the most valuable lead is often the one you already paid for and then ignored - is the kind of plain truth that tends to age well.