The automated sales desk

The sales development representative (SDR) role has long been defined by a specific kind of grind: scraping lists, writing hundreds of emails, and hoping for a one-percent hit rate. It is a role that has historically served as a gateway into the technology industry for human workers but is increasingly being reimagined as a job for software. Novum, a San Francisco-based startup in the Y Combinator Summer 2024 cohort, is a primary example of this shift. They are not building a tool for SDRs to use; they are building a software agent that aims to replace the core functions of the role itself.

Founded by Aditya Bhardwaj and Sarthak Sahoo, Novum focuses on the "top of the funnel"—the most repetitive and research-intensive part of the sales process. The platform is designed to take a high-level description of a target customer—such as "CTOs at Series B fintech companies in New York"—and turn it into a list of verified leads. This involves more than just a database query. The agent scrapes the open web, checks social profiles, and verifies email deliverability before any outreach occurs.

Research as a differentiator

Once a list is built, the research phase begins. While traditional sales automation tools like Apollo or Outreach have existed for years, they primarily function as static databases or mail-merge engines. Novum's approach is distinct because it uses LLMs to perform the research step that previously required human intuition. It can browse a prospect's website, read their latest company news, and synthesize a reason for reaching out that goes beyond the generic greeting. This is the difference between a sequence and an agent; the former follows a fixed script, while the latter adapts its output based on the data it discovers about each individual prospect.

The core value proposition here is autonomy. In a traditional setup, a human has to approve every list and often every email. Novum's agents are designed to operate with a higher degree of independence. They handle the search, the verification, the drafting, and the initial response handling. This moves the human's role from "doer" to "manager," where they oversee the agent's performance and handle only the high-value conversations once a prospect has expressed interest.

The competitive landscape for sales agents

Novum is entering one of the most competitive segments of the current AI boom. They face established incumbents like Apollo.io, which is aggressively adding AI features to its massive database, and fellow startups like 11x.ai and Artisan. The differentiator for Novum appears to be the depth of its agentic reasoning. By focusing on the end-to-end automation of the workflow rather than just a single point of friction, they are betting that companies will prefer a specialized agent over a general-purpose CRM extension.

However, the category faces a structural challenge: the potential for inbox saturation. If every startup has an agent sending perfectly personalized emails at infinite scale, the signal-to-noise ratio in professional communication may collapse. Novum’s success depends on its ability to keep the quality of outreach high enough that it doesn't just get caught in spam filters. For now, the company is betting that better research and more human-like reasoning will keep their agents effective in a world of increasing automation.