The Anti-AI SDR Company Coming Out of Stealth
Rox x The Signal
If you were forwarded this newsletter, consider joining over 1,400 weekly readers—some of the smartest GTM founders and operators—by subscribing here:
Hey y’all!
I’m excited about today’s post for a few reasons.
First, the Co-founder & CEO, Ishan Mukherjee, was one of our angel investors at Groundswell (my last company); he’s insanely sharp. His first company was acquired by New Relic for 8-figures. And he was their Chief Growth Officer for the last few years. He’s a product-focused founder who is going all-in on agentic workflows (and not just because they’re the hot thing in Silicon Valley right now - he has been deep in the data and robotics world for the last decade+).
Secondly, my co-founder at Groundswell, Hari Nayak, joined Rox as a founding engineer - and given the caliber of technical talent and quality of human he is, it makes me all the more bullish on where they’ll take the company.
Third, I’ve been rooting for this team for a while, and excited that they’re coming out of stealth this week to show the world what a true AI-first company looks like.
And I’m humbled that they wanted me to write a piece for them as part of their announcement strategy. You can read my sponsored deep dive thought process here.
Here’s what we cover in today’s post:
What are account-based agents?
Software that gets smarter with time
Why Rox?
Anti-AI SDR
Supercharge your best reps
Conclusion
Alright, let’s get into it.
Building Account-Based AI Agents
What are account-based agents?
Rox's central idea is account-based agents.
This is a major shift in how we should think about supporting salespeople with agents.
Traditionally, we think about task-specific agents or assistants - eg: an email-writing assistant, a research bot, a CRM plugin, etc. Rox flips this on its head.
Instead of having agents for different tasks, Rox provides one dedicated agent per account. Each agent collects information from every relevant source:
public data
third-party data partners
first-party tools like your CRM and data warehouse
and can even leverage a team of humans who can be used to manually research and verify important data.
It then contextualizes the nuances of that account as a whole. It’s like having a highly-trained specialist (an AI team member) whose sole responsibility is to know everything there is about a single customer.
Because of the agentic nature, you’ll get value on day one.
When you first sign up and put in your target accounts, an “agent swarm” (a collection of autonomous agents) gets working - starting with a deep scrape on your website. So, Rox has the context about what product you sell and who you sell to (from things like case studies on your website, value props on your marketing website, etc.).
I know “agentic” is a buzzword right now. But this is not vaporware. You can give try it for free and see for yourself! :)
Gone are the days of signing up for a SaaS product as an “empty shell.” AI agents shift things so they can go out and do research to give context, so your very first experience with the software is relevant for you. Pretty magical.
It took me a while playing around with Rox, to understand why I would want a single agent for every account in my book of business. It turns out, using these account-based agents is something like two orders of magnitude better than using one generic agent. Wild.
Software that gets smarter with time
I’ve always joked that gtm tech gets worse the more you use it, not better (looking at you, Salesforce… after several years of usage, it starts to look more like a 59-year old Mike Tyson trying to fight a 27-year old YouTuber… tired and past its prime.) There are an overwhelming amount of fields, broken workflows no longer maintained by the person who built them, slow ui/ux as the data blooms, a graveyard of reports, the list goes on and on.
With Rox, it gets smarter the more you use it. Over time, the tuning kicks in the more context you share with it, the better it will get.
This is something my co-founders and I intended to build at Groundswell. As I mentioned before, my previous co-founder, Hari, is a founding engineer at Rox, so I’m excited to see this vision coming to life. Hari is the rare breed of someone who is exceptional, technically - can scale systems and build prod/eng/data teams. But, also, strategic, first-principled, and does not have an ego - which gives him a mindset to learn and solve real problems from a place of deep empathy.
Why Rox?
Another thing that stands out about Rox is the focus on measurable productivity gains. When you can consolidate this level of increased productivity, it allows (scaled) companies to rethink their go-to-market strategy altogether.
Over 35 of the best-performing enterprise sales teams have adopted Rox organically (more on that later in this post). The team is crazy impressive. Most of their engineers are from Stanford and MIT and have worked under famous professors like Chris Re and Matei Zaharia (Databricks). And, they just raised a $42M Series A from Sequoia, General Catalyst, Google Ventures, Elad Gil and other operator angels.
All of this stuff is impressive.
But, what impresses me the most is how Ishan responded to me when I told him congratulations.
He said: “But that doesn't really matter. What matters is: do reps find it useful? Does it make them more productive? And hopefully the productivity is significant enough where it changes the fundamental unit economics of the market. That's pretty much our sole focus.”
Anti-AI SDR
Let’s talk about the product, and where it sits in the stack.
My dumbed-down explanation of the product: Perplexity/Claude Projects + Clay-lite + a slick workflow layer (Plus… it uses dark mode — my kinda people).
It’s built for quota-carrying reps.
Rox is explicitly anti-AISDR. It's not trying to replace the human touch with automation. Instead, Rox aims to provide highly contextual, draft-level work that salespeople can use to craft thoughtful outreach, but it doesn't press send for you. This deliberate design decision ensures that the reps stay in control and that outreach quality is maintained.
Philosophically, the agents are good at doing what humans can. But, instantly. This is due to the massive technological inflection point that we’re undergoing right now with the advent of LLMs.
Here’s a direct quote from a recent conversation I had with Ishan (the man is playing 3D chess, while I’m over here trying to learn the rules of checkers) -
Before The Singularity happens, there's going to be a whole lot of people who will think about zero to automation, and there's a whole lot of people who think through supercharging and augmenting your highest value employees.
I build robots and warehouses. So I have a very visible understanding of automation across the economy. So the obsession here is every business in this world has to evolve in this kind of AI-first era.
For us, in go-to-market, AI at least opens up the opportunity to drive radical productivity gains.
The second, is the art or science of selling has changed as you're moving from a commit based revenue model to a consumption based to every model.
At the limit is usage based, but at the base layer, it's going to land and expand. With that, the art of selling changes. It became more value driven.
So as those two things happen, our existence on this planet is to supercharge your best reps to make them better.
And our approach is: we will help you literally save time on information gathering, wrangling and reasoning.
Supercharge your best reps
Where does the workflow live?
Rox’s arbitrage is around the fact that reps don’t spend time in Salesforce anymore. And, maybe more interestingly, the data is not all in Salesforce anymore (this is a very big deal).
If the product can truly supercharge quota-carrying reps (AEs and AMs) then increasing the book size by 50% next year is the goal for many of the Rox customers. The north-star question: can Rox increase the number of accounts a rep can manage?
They’re also making some big bets around their application suite - building a Mac-native App (think: SuperHuman, Notion, Slack, etc.) and an iOS app (still crazy to me that there’s no mobile app that all salespeople use… yet). And the entire interface is agentic.
Ramp is one of Rox’s early customers. They started with one rep, then the SDR Team adopted it, and eventually, it spread organically and every AE uses Rox now (it’s where they start and end their day).
Ramp has saved 200+ hours of time by automating previously manual work - things like: pre-meeting research, note-taking, follow-up reminders, and more. Ramp is re-allocating those hours into pipeline generation.
This is a great example of how software will enable teams to do more, with less. I think we’re still very early here.
Conclusion
They’re making a big bet: having the ambitious goal to build the most essential software business to run modern go-to-market. Starting with the data layer, and then the alerting layer (or, what I would probably call, The “System of Intelligence”). Eventually, they will aim to earn the right to become the system of record.
But, at the end of the day, Rox’s biggest challenge is building something that salespeople use and love (find value in). Their vision is solid and the team is a 10/10. It’s a matter of execution to help make the best reps in the world better. I’m excited to see this story unfold. It’s an optimistic view into the future of how AI agents will help augment the best reps, and not replace them. Cheers to that future!
PS - If you want to play around with it, you can sign up for free, here.
Thanks for reading, and as always, I appreciate your attention.
See you next time,
Brendan 🫡