Berlin
I run commercial teams and build the agent systems that do the work with them.
Head of Growth at SIDES. Head of Revenue at JOIN from November 2026.
15 years across business, marketing, revenue operations and product. Three companies founded in Paraguay, then commercial roles scaling startups from bootstrapped to Series B, valuations from $10M to $1B. Multiple business models, org changes and market pivots on the way to product-market fit, funding and profitability. The through line: owning a number end to end. Setting targets, forecasting against them, running the P&L, and leading the sales, account management and marketing teams that deliver it.
Old world: campaigns and dashboards. Today: agents with tools, decision logic and execution capability, running inside day-to-day commercial work. Since the start of 2026 I have been generating pipeline and revenue through agentic setups, built on Claude Code, Codex and open-weight models depending on the job.
I care about growing people as much as pipelines, and I believe the best automation is the one nobody notices.
Writing
Build logs and field notes from live agentic GTM and RevOps work.
One agent harness per context, with hard data isolation between them: one runs my personal operations, one runs commercial operations at SIDES, and the third is ready for the next role. Same pattern each time: specialist agents, shared memory, and a human approval step on anything that leaves the building.
Multi-agent system for commercial operations on Claude Code and MCP, wired into our CRM, analytics, mail and ad stack. Rolled out beyond my own team, with per-user permissions and shared memory at user, department and org level.
SDR pre-call research went from 5-10 minutes per lead to one chat message: double-digit hours back every month, for a two-figure infra bill.
Runs my life admin, research and side projects: 10 specialist agents, 29 skills, around 100 tools and 80+ scheduled jobs on top of one knowledge vault.
Measured in an A/B ablation (n=39): the full harness beat the bare model 35 to 4. The instruction file alone scored 19 to 19. The value sits in the tools and agents, not the prompt.
The open-source template of that setup. Clone it, run one script, and you have a working AI partner on your own machine in ten minutes. No signup, no telemetry.
7 pull requests from 3 external contributors in the first days.
github.com/chacosoldier/compabobCo-founded venture: a hands-free cockpit for running AI coding agents from anywhere. Live since July 2026.
app.yothere.ai
Recruiting SaaS. Joining the management team to own Revenue end to end: Sales, Account Management, RevOps and GTM.
Restaurant management SaaS used in 6,000+ locations across DACH and the Netherlands. Own marketing, outbound, RevOps and partnerships. Built the commercial agent platform.
End-to-end commercial ops across Sales and CS. Owned the commercial tech stack and built AI automations for Finance and People.
First RevOps hire. Full HubSpot rebuild, inbound/outbound processes, automated NPS and customer tiering. The pivot from growth into RevOps.
Marketing due diligence on acquisition targets for a $750M+ funded Amazon aggregator. Red flags, opportunities and growth forecasts feeding investment decisions.
First C-level title. Repositioned a consumer brand on a minimal budget and built a strategy to grow D2C revenue 20x over three years.
Four promotions in four years, IC to business-unit leadership. Ran a 20+ person P&L, powered 5M+ rides, cut 14,000+ tons of CO2, then pivoted B2C to B2B software.
Built Paraguay's first rental e-commerce platform (profitable in under six months) and its first carpooling platform. Won the CARMA AXLR8R and two World Summit Awards.
Recognition
For teams that want to run this motion themselves.
A time-capped audit of your commercial stack and workflows: the funnel end to end, where the ICP signals sit, a baseline of today's numbers and the targets I would hold a build to.
Written findings, a target architecture, a walkthrough. Typically two weeks.
Weekly working sessions with your operator, plus async review in between. Enrichment waterfalls, scoring, agentic research-and-draft loops, CRM writeback, review gates.
Ends when your team runs the motion without me.
Hands-on sessions on agentic GTM for commercial teams. No theory decks; every section is a live demo. Most recently at Revenue for Breakfast #18 by Cremanski & Company in Berlin.
Half day to full day, on site or remote.
The deck from that sessioncremanski.comA specific question with a defined end date. Second opinions on RevOps architecture, agentic tooling, and build-versus-buy calls.
Short and scoped.
Every engagement is fixed scope with a named end date, never a retainer. Out-of-scope work is agreed in writing beforehand and never a surprise on an invoice.
Availability
I run this next to a full-time operator role, so I take on a handful of projects a year, and only where my profile clearly fits the problem. Right now: a fixed-scope AI-GTM audit and build-along coaching program for a Munich B2B SaaS.
If you are building agentic GTM or rebuilding RevOps, reach out. Always happy to compare notes.
[email protected]