About Achkan,
Achkan Chavoushi is Head of Growth at Nuvo, a late Series A startup that describes itself as an agentic trade network for physical goods, with more than 200,000 companies on the platform. He was the company’s second marketing hire, about two months after the Head of Marketing. Before Nuvo, he spent a year at Rippling, worked across the French and US markets at TinyClues, and helped French tech companies expand into the United States at Business France. A French native based in the US, he has spent six to seven years there since 2016.
🎙️ Listen on YouTube, Apple Podcast, Spotify
At Bulldozer, we see most marketing teams start their AI journey with content: writing more, faster. Achkan Chavoushi took a different route at Nuvo. He looked at the entire GTM machine, spotted where it was losing time and money, then built the infrastructure that lets agents do real work.
In this episode, he walks through his first 90 days: the data, the GTM brain, the agents that run every week, and how he got internal buy-in to start there. We also get into maintenance, team structure, the differences between the US and European markets, and what Rippling taught him about how to execute.
Here’s what I took away:
1) He chose to fix what already worked before launching new channels.
Nuvo was sales-led, with SDRs and account executives selling directly to target companies. Achkan found manual work, a CRM full of junk and duplicates, and SDRs losing a huge amount of time hunting for contacts. Taking a channel from 0 to 1 takes far longer than taking something that works from 1 to 10, so he wrote a one-pager quantifying the lost time and money and got a direct mandate from the CEO.
2) Data first, then a GTM brain then agents on top.
The first 30 to 60 days went into clean data: knowing exactly who Nuvo sells to, what the TAM is, and having a CRM defined well enough for proper account tiering. On top of that came a data dictionary describing every CRM field so agents know where to look, and a marketing brain holding the positioning, personas, and go-to-market approach. Since the market leaves few traces online, the TAM of roughly 140,000 to 150,000 companies is refreshed weekly.
3) The calls from the 50 biggest deals fed the brain, then agents took over.
Achkan ran the calls from his 50 largest deals through a model and got a 20 to 30 page document on personas and buying reasons. Now call transcripts flow into the CRM, where agents analyze them every week: competitor mentions, whether the pitch lands on new products, CRM updates, a one-pager per rep for manager feedback, and the forecast. That forecast combines the rep’s call, the manager’s call, and the agent’s, which can flag that a deal won’t close on time because a document was never sent. The Head of Finance, who used to rework it by hand, is happy with it.
4) Developers build the sandbox and marketers build inside it.
For the website, Achkan ruled out a vibe-coded site, too risky at Nuvo’s valuation, and a return to a developer on Webflow. An agency built a walled-off environment on Sanity and Vercel where agents work with skills hosted on GitHub, importing Figma files and shipping pages while running SEO and technical checks. The SDR tooling follows the same logic: an internal tool sends each SDR a weekly list of enriched accounts with signals and generated emails, and the in-house sequencer took one to two weeks to build on Lovable. Engineers built the backend, marketers the front end.
5) Maintenance: engineers own the architecture with versioning and tests.
Maintenance is a cost most companies never identify. Achkan had asked his boss the week before whether it would all hold up with ten to fifteen reps on it, and so far it does. Developers make the architecture decisions, GitHub handles versioning and rollback, and skills audit what has been built. He plans to add a cascade of tests on every deployment and puts the risk of breaking everything at low but real. Tools built together with the reps were adopted almost immediately.
6) The team is built around a system and the process produces the software.
The Head of Marketing hired a Head of Growth first, then a content person for brand voice, then a RevOps profile. Rippling and most companies ran with one person per channel. Achkan starts from the ideal process and asks what the system needs: data, product guidelines in the brain, AI-managed reporting, and software to visualize and act by function. He sums it up this way: the software you bought used to dictate the process, and now the process produces the software. I shared a related thought of mine: off-the-shelf software imposes its own worldview, and AI lets you build one that fits how you work.
7) From Rippling, a benchmark for what good looks like and a method for execution.
Achkan calls Rippling a school of resilience that doesn’t let go of a problem until it’s solved. It also shows what strong growth looks like, which makes real signals easier to spot. Each quarter, the company picked two or three problems, the CEO joined weekly blitz meetings, every action item had an owner and a date, and a public Slack channel followed up daily. There are three ex-Rippling people at Nuvo, and he applies the method in every new company.
Find Achkan on LinkedIn : https://www.linkedin.com/in/achkan/
Enjoy,
— Jordan
