Darwin AI bets 'AI workers' win mid-market sales: 300+ clients, $2M ARR, Base10 seed

Darwin AI is a Sao Paulo startup whose Darwin 2.0 platform lets small and mid-sized businesses create and train 'AI workers' for inbound sales, outbound campaigns, collections and support — handling WhatsApp, Instagram and voice calls in Spanish and Portuguese, integrated with the client's CRM, priced per processed conversation.

Mid-market sellers ignored by enterprise AI and micro-tools would pay per conversation for trainable 'AI workers'; if they won't automate, Darwin AI has no wedge.

US$2.5M seed in 2024 (investors include Canary, FJ Labs, Latitud, Dalus Capital, Bridge Latam); US$4.5M follow-on seed led by Base10 Partners (Aug 2025); total ~US$7M (R$37M+).

After building and selling Sirena, a contact-center AI company acquired by Zenvia, Argentine co-founders Lautaro Schiaffino (CEO) and Ezequiel Sculli (CPO/CTO) — partners for 11 years — founded Darwin AI in Sao Paulo in 2023. Their thesis: Big Tech serves the extremes (enterprise and micro-businesses), while the mid-market, whose high-ticket products demand consultative selling, is being left behind.

Darwin 2.0 lets clients create role-specific AI workers — inbound sales, outbound campaigns, collections, support — integrated with CRMs, calendars and knowledge bases, with human handoff and hallucination controls built in. After a US$2.5M seed in 2024 the company reported 300+ clients in 20 countries and $2M+ annual recurring revenue by August 2025, when Base10 Partners led a US$4.5M follow-on seed (total ~US$7M); 60% of the round was earmarked for Brazil, with Mexico next, and the founders said revenue had more than doubled in the first half of 2025.

Scaling: by March 2026 Darwin AI reported more than $3M in annual sales across 20+ countries and ~$7M raised, with named AI workers per role (qualifying, reactivation, satisfaction, collections, post-sales) and expansion into Spain and Portugal planned.

Schiaffino and Sculli had 11 years of partnership and one exit (Sirena to Zenvia), giving them credibility with the mid-market customers enterprise AI vendors ignore.

The thesis is structural: Big Tech chases enterprise and micro-businesses, leaving consultative mid-market sellers with high ticket sizes and no automation.

Per-conversation pricing and client-trainable AI workers share performance risk, making the sale easier than enterprise AI contracts.

300+ clients in 20 countries with $2M+ ARR by August 2025 (then $3M+ by March 2026) is concrete evidence the middle market will pay.

Price the outcome, not the model: per-conversation pricing and client-trainable AI workers make 'AI agent' a controllable employee — killing the mid-market's 'we can't build AI' objection.

As of March 20, 2026, Darwin AI is scaling from Sao Paulo with a 60-person team, more than $3M in annual sales, operations across 20+ countries and about US$7M raised (Base10-led $4.5M follow-on seed in August 2025). The Darwin 2.0 platform sells per-processed-conversation pricing and role-specific AI workers (prospect qualification, reactivation, satisfaction, collections, post-sales), with plans to expand to Spain and Portugal. Open risks: competition from larger AI platforms absorbing the SMB-automation layer, and the need to keep cutting hallucination in synchronous voice calls.