When Y Combinator batch pages go live, you can usually watch a startup's narrative snap into focus—often for the first time publicly. For Nex, a five-person team working out of San Francisco, that crystallizing moment came with a single line: "Claude Cowork for GTM Automations."
It's a deliberate stake in the ground. The go-to-market automation space has gotten crowded, quickly, and Nex's founders—Najmuzzaman Mohammad (he goes by Nazz) and Francisco Dias—are betting they can carve out territory by positioning themselves to work within Anthropic's Cowork framework.
Both spent years inside HubSpot before this. Mohammad was a senior product manager there, and later at Clipboard Health. Dias came up as an early engineer on HubSpot's CRM before shifting to the company's AI infrastructure team. Now they're building what they describe as an AI agent that "turns your GTM goals into complex GTM automations" across what the company's pricing page explicitly states are more than 1,000 integrations.
The pitch hinges on execution mechanics. Nex says its agent writes code as it works—making workflows deterministic and, crucially, cheap to re-run. The outputs aren't conversational chat responses. They're Slack messages. Opened pull requests. Live structured data tables. It's a technical distinction that matters more than it might sound, especially in a market where "agentic" has become a baseline claim but the actual delivery mechanisms vary wildly.
A Familiar Story: Ex-Operators Chasing the Next Layer
Nex isn't pioneering virgin territory here. Apollo.io launched its AI Assistant and Outbound Copilot earlier this year, billing itself as "the first AI-native all-in-one GTM platform" and rolling out a native Claude connector shortly after. Clay announced its Claygent Builder in May, then outlined a roadmap in July that included agents for most GTM jobs and integration with Anthropic's Model Context Protocol to bring Clay workflows into Claude, ChatGPT, and other surfaces. Highspot unveiled its GTM Agent in early May. Common Room has been running a Claude plugin for its RoomieAI workflows since mid-May.
What Nex is doing—deliberately, explicitly—is leaning into Anthropic's Cowork paradigm. The company's homepage calls itself an "AI GTM Engineer" and echoes the Y Combinator tagline: it's purpose-built to work inside Cowork, which became generally available on web and mobile over the summer. Cowork lets users hand Claude a task and have it work autonomously across connected tools—calendar, email, messaging, the web—until the job is done. Nex is effectively saying: we'll do that, but tailored specifically for go-to-market teams.
There's some depth to the claim. The company has published developer documentation laying out what it calls a "workflow execution and context layer," last modified in mid-July. That includes a knowledge graph, entity and relationship management, context ingestion, insights streaming. It's the kind of plumbing that suggests Nex isn't just wrapping a large language model in a cleaner interface and calling it a day.
The Product: Pre-Built Plays and Custom Workflows
Nex lists several pre-built GTM plays on its site: closed-lost re-engagement, CRM hygiene, meeting prep, deals going dark, lead scoring. Users can also describe their own workflows and generate agents from business context—at least in theory. The company says it supports Claude, ChatGPT, Cursor, LangChain, LlamaIndex, CrewAI, Slack bots, and any HTTP agent via REST API. Data sources include HubSpot, Salesforce, Gmail, Outlook, Slack, WhatsApp Business API, and data warehouses.
Pricing sits at $100 per seat per month for the Pro plan. That includes unlimited data sources, unlimited workflows, the claimed 1,000-plus integrations, and email support. There's a 14-day free trial. The billing kicks in at day 14 or $50 in AI usage, whichever comes first. Under $10 in usage? The workspace pauses and you're never charged. The Team plan offers custom pricing, unlimited seats, custom integrations, API access, and—on request—a SOC 2 report. The company's brand facts page claims SOC 2 Type 1 certification, though this remains unverified due to access restrictions on the trust portal.
Nex's stated target market, per that same brand facts page, is companies with 50 to 500 employees deploying AI agents. It's a specific band—too small for the grinding cycles of enterprise sales, too large for pure self-serve motion to work cleanly.
The Open-Source Signal
Before the Y Combinator announcement, Nex was making technical noise in developer circles. In late April, the founders posted to Hacker News about WUPHF, an open-source project they describe as "Slack for AI employees with a shared brain." The GitHub repo is active and includes architecture documentation aligned with Nex's emphasis on agent teams and context management. The Product Hunt listing shows it launched earlier this year and ranked #7 for the day with 156 upvotes—not viral, but enough to suggest some early traction in developer communities.
Between late February and early March, the company published a series of blog posts: building context-aware Cursor workflows, adding enterprise context to custom agents, building an MCP server and tools for Claude Desktop, connecting ChatGPT to CRM and email and Slack. Read as a sequence, it looks like a team thinking through the infrastructure layer in public before packaging it as a product. That's not an uncommon playbook for developer-first startups, but it does suggest the founders were building in earnest well before any formal launch.
The Inconsistencies
Company-reported data on Nex varies depending on where you look. The Y Combinator page lists a founding year of 2025 and a team size of five. The company's own brand facts page and LinkedIn profile say 2024 and 2 to 10 employees, respectively—discrepancies that remain unresolved across public sources. The company is registered as a trademark of GarageSpace, Inc., which appears to be a corporate entity Mohammad has used before—though details on that front are sparse.
Investors listed on the homepage include Dharmesh Shah, founder of HubSpot, and Girish Mathrubootham, founder of Freshworks, along with Boldcap Ventures, Together Fund, Leonis Capital, and Cur8 Capital. No funding amounts or round details are publicly disclosed. Together Fund's portfolio page lists Nex with what it labels a seed investment and marks the status as "Exited," though the company is actively operating and selling a product, suggesting this status may be outdated or incorrect.
The integration count also shifts. Some homepage graphics reference "100+" while the pricing page says "1000+." The higher number likely reflects ecosystem access through platforms like Zapier and Anthropic's remote MCP connectors, rather than native, one-to-one integrations Nex has built itself. It's a distinction that matters—particularly for customers evaluating the depth of what they're actually getting.
The Timing Bet
Leonis Capital, one of Nex's investors, published predictions earlier this year suggesting that AI labs would increasingly prioritize first-party surfaces like Claude Cowork, reshaping the developer ecosystem around those platforms. If that thesis holds—and it's still early—Nex is positioning ahead of the curve. Building on Cowork before it becomes a crowded moat. Targeting GTM workflows before every productivity tool adds "agent" to its feature list as table stakes.
Whether that timing advantage translates into real traction depends on execution details that aren't yet public. How well does the code-writing agent actually work in practice? How many of those 1,000 integrations are genuinely useful versus technically possible? Does the deterministic workflow claim hold up under the messy complexity of real-world GTM operations?
The product is live. The pricing is posted. The founders have credibility from their HubSpot years, and they've clearly been building infrastructure with some seriousness. What's missing, at this stage, is proof.
For now, Nex is a bet on platform timing, a bet on Anthropic's trajectory, and a bet that go-to-market teams are ready to hand off more work to agents than they currently do. Not all of those bets have to pay off for the company to find traction. But in a market this noisy, even two out of three might not be enough.
