The smartest outbound sales teams have long understood a basic truth: timing trumps almost everything else. Catch a prospect at the exact moment they're thinking about your category, and conversion rates can jump. Miss that window by a week—sometimes even a day—and you're back to cold calling.
The trick, of course, is spotting those moments before your competitors do.
That's where things get messy. A senior engineering hire announced on LinkedIn. A flurry of GitHub commits suggesting someone's rebuilding their infrastructure. An incorporation filing in Delaware that might signal fresh funding. These signals exist, scattered across the internet in real time, but most sales teams learn about them far too late to matter.
Autumn AI, a company that recently emerged from Y Combinator's Winter 2026 batch (with two team members listed on their YC profile as of early March), is betting it can solve that problem. The pitch, laid out in materials published in early February, is deceptively simple: monitor the entire internet for buying signals, filter out the noise, and deliver what's left directly to your inbox or Slack channel before anyone else sees it.
Whether the company can actually pull that off remains an open question. But the ambition is worth examining.
A Feed, Not a Dashboard
Strip away the marketing language, and Autumn's product is essentially a highly targeted alert system. Users start by defining their ideal customer profile—company size, industry, growth stage, whatever matters for their specific product. Then they specify which signals actually indicate intent: maybe it's blog posts about scaling challenges, or commits to particular open-source repositories, or new hires in the data engineering department.
Autumn's system then monitors those sources continuously. LinkedIn, GitHub, X (the platform formerly known as Twitter), personal blogs, conference registrations, even incorporation filings. When something matches the criteria, the platform pushes an alert.
The company's Launch YC post from early February describes this as a feed "filtered by intent." The idea, presumably, is to avoid becoming yet another source of information overload. Instead of forcing sales teams to wade through thousands of potential signals, Autumn aims to surface only the handful that matter for any given account or market segment.
Delivery happens through tools teams already use—email or Slack. No new dashboard to check, no additional software to log into. Just notifications that appear where you're already working.
On its website, Autumn calls itself "the first real-time signal intelligence platform for GTM teams." That's a bold claim in a category that's been around, in various forms, for years. But it hints at the company's core thesis: that existing tools are too slow, too noisy, or too narrow in what they track.
Two Columbia Grads and a Bet on Speed
Co-founders Vishnu Sampathkumar and Shiv Kampani met during orientation at Columbia, according to their YC materials. Sampathkumar's background tilts technical—he built sourcing technology for Sierra Ventures and worked as a machine learning engineer at Modulus, which focuses on AI security. Kampani comes from the research side, with work presented at NeurIPS and time spent at SandboxAQ, a quantum and AI company spun out of Alphabet.
Both announced their acceptance into YC's Winter 2026 cohort on LinkedIn late last year. The company itself was founded sometime in 2025, and the YC profile currently lists just the two of them.
That's a lean team for a product promising to monitor the entire internet, though perhaps fitting for a company in launch mode. Most YC startups at this stage are still figuring out what works.
Three Ways In
Autumn's launch materials outline three specific workflows, each targeting a different pain point in the sales process.
First: incorporation prospecting. The platform tracks newly formed companies—the kind you'd normally find through state filings or services that aggregate that data. But Autumn layers on additional signals: banking relationships, payroll setup, purchases of machine learning infrastructure. The goal is to identify not just new companies, but new companies that fit your ICP and are already taking steps that suggest near-term buying intent. This is clearly aimed at teams selling into early-stage startups.
Second: enterprise account monitoring. For larger deals, the challenge is often figuring out who to talk to and when. Autumn can track organizational changes at target accounts—new hires, departures, restructures. It builds out org trees to map departments and identify new entry points. If your champion leaves or a new VP joins the team you're trying to reach, you'd know quickly.
Third: conference preparation. Sales teams heading to industry events can monitor attendee lists (when available), cross-reference recent activity from those attendees, and arrive with a prioritized list of who to target. It's conference networking, but with a research layer baked in.
Y Combinator mentioned in a mid-February LinkedIn post that Autumn is "already powering workflows" in fintech and deep tech, though no customer logos or case studies are publicly available yet. That's standard for a company this young, but it does mean most of what we know comes from the founders' own descriptions rather than third-party validation.
An Old Problem, New Approaches
Signal-based prospecting isn't exactly a novel idea. Established players like 6sense and Bombora have built significant businesses around buyer intent data, stitching together web activity, content consumption patterns, and third-party signals to predict when companies are in-market. Clay, a newer but increasingly popular platform, has trained a generation of sales ops teams to think in terms of signals and triggers, automating outreach based on specific events.
More recently, a handful of startups have positioned specifically around real-time signals. Clearcue, Zimt, and Fundz are all attempting to compress the gap between signal and action. The category is getting crowded, and fast.
Autumn's differentiator, at least in its messaging, seems to be breadth and speed. Incorporation filings and GitHub commits aren't standard inputs for most intent platforms, which tend to focus on more traditional B2B signals like job postings and funding announcements. The implicit argument is that by casting a wider net—and by moving faster—sales teams can reach prospects before the rest of the market catches on.
Whether that actually works at scale is harder to say. Intent data has a well-documented noise problem. Spend any time in GTM-focused subreddits or Slack communities, and you'll find complaints about tools that flood teams with false positives: signals that look relevant but lead nowhere, or alerts that arrive too late to be useful.
Autumn's bet is that better filtering and cross-source synthesis can solve for that. Maybe pulling together multiple weak signals—a GitHub commit plus a LinkedIn post plus a blog mention—creates a stronger, more reliable indicator than any single data point. Or maybe the noise problem is inherent to the category, and even clever engineering can't fully escape it.
What We Still Don't Know
For a product positioning itself around transparency and speed, Autumn is holding some key details close. The website offers a "Book a Call" button, but no public pricing information. We don't know if it's seat-based, signal-based, or usage-based. We don't know how frequently data refreshes, or what guarantees the company makes around coverage and accuracy.
There's also no published information on the technical mechanics: how Autumn collects data, deduplicates it, or enriches it. For a platform selling real-time intelligence, those details matter quite a bit. Is this scraping? API access? Some combination? How does the company handle rate limits, data quality issues, or platform changes?
The launch materials include a YouTube demo video, but the exact setup and configuration process isn't spelled out publicly. Early adopters will need to hop on a call to understand what they're actually buying, which is fine for now but limits transparency.
This is all fairly typical for a company in classic YC launch mode: narrow value proposition, clear target customer, invitation to engage. Get some early users, iterate quickly, figure out pricing later. The question is whether the core insight—that sales teams need faster, broader signal intelligence—is correct, and whether Autumn's execution can match the ambition.
The next few months should clarify both. Either the product will prove it can reliably surface high-value signals ahead of the competition, or it'll run into the same challenges that have plagued intent data providers for years. For now, it's a compelling pitch in search of proof.
