The first question an enterprise architect asks about an AI agent isn't about its reasoning capabilities. It's about where it writes, what it can read, and who gets an alert when it goes off script. Runflow, a Brazilian startup that emerged from the AI leadership school IFTL, is building its entire platform around that second set of questions.
Launched in January 2025, Runflow is not another playground for prompt engineering. It is an enterprise platform designed to create, orchestrate, and govern AI agents that are directly connected to a company's existing systems, from CRM and ERP to legal and finance databases. The company's early traction, including an Annual Recurring Revenue (ARR) of R$2.6 million (approximately $500,000) with 33% monthly growth, suggests a specific appetite for this approach in the Brazilian market [Startups]. The bet is that the path to production for AI agents is paved with compliance guardrails and pre-built connectors, not just more powerful models.
From classroom to corporate spin-off
Runflow's origin is its most distinctive feature. The company is a spin-off from IFTL, an edtech focused on training corporate leaders in AI, strategy, and technology [PERPLEXITY SONAR PRO BRIEF]. The founders, Edmee Moreira, Danrley Morais, and Mariana Pastorelli, developed the initial Runflow platform as an internal tool to automate commercial and operational flows within IFTL itself [runflow.io]. When the tool proved effective at scale, they packaged it as a standalone product. This genesis provides a natural wedge: IFTL's network of AI-oriented executives across Brazilian corporations became the initial beachhead for customer conversations and product feedback.
The founding team blends educational and technical backgrounds. Edmee Moreira comes from the AI leadership education side of IFTL. Danrley Morais is a co-founder and serves as CTO, bringing the engineering depth required to build a platform that promises system integration [LinkedIn]. The company has also enlisted Joaquim Neto, former Technology Director at iFood and co-founder of DeliveryCenter, as an advisor, lending credibility in scaling complex technology operations in Latin America [LinkedIn].
The governance-first platform wedge
Runflow's technical differentiation rests on three pillars: integration, governance, and time-to-value. The platform provides an SDK and a library of tools that allow developers to connect agents to internal APIs and data sources, such as pulling customer data from a CRM or submitting a legal document for review [runflow.ai]. This moves the agent from a conversational interface into the workflow itself.
The governance layer is treated as a first-class product feature. The platform offers visibility into agent costs, performance metrics, and operational impact, along with controls to manage permissions and audit trails [runflow.ai]. For regulated industries like finance or healthcare in Brazil, this addresses a primary barrier to adoption. The promise is to move from weeks or months of custom integration and compliance work to a platform where agents are "ready to operate in weeks" [PERPLEXITY SONAR PRO BRIEF].
| Feature | Runflow's Approach | Typical AI Tool Approach |
|---|---|---|
| Primary Focus | Production deployment & system integration | Model capabilities & prompt crafting |
| Integration | SDK & pre-built connectors for corporate systems (CRM, ERP, etc.) | API-centric, often stateless |
| Governance | Built-in cost tracking, performance monitoring, audit trails | Often an afterthought or custom-built |
| Target User | Enterprise IT/development teams | Developers & business users experimenting |
Traction in a nascent market
While the global market for AI agent platforms is crowded with well-funded incumbents, Runflow's early focus is distinctly local and vertical. The company reports automating processes in sales, legal, customer service, and finance for its initial clients [Startups]. Its disclosed ARR of R$2.6 million, with a goal of reaching R$3.6 million by year-end, indicates it has moved beyond the conceptual stage and is closing business deals [Startups]. The 33% monthly growth rate, while from a small base, points to strong initial product-market fit within its niche.
The platform's architecture suggests a focus on mid-to-large enterprises that have existing processes ripe for automation but lack the in-house infrastructure to safely connect LLMs to their core systems. By offering a complete platform,from agent building tools to deployment orchestration and monitoring,Runflow aims to become the "operating system" for corporate AI agents, an ambitious claim that positions it as an ERP-like layer for automation [runflow.io].
The technical breakdown and scale risks
From an infrastructure perspective, Runflow's value is in abstracting away the plumbing. The platform likely handles the messy work of managing API rate limits, maintaining state across long-running agent workflows, securing credentials for system access, and logging every action for compliance reviews. This is non-trivial engineering work that most enterprises would struggle to build robustly in-house.
The primary technical risk is integration depth. The promise of "connected to your systems" is only as good as the breadth and robustness of its connectors. A platform that offers shallow integrations or fails to keep pace with updates to Salesforce, SAP, or local Brazilian ERPs like Totvs will quickly hit a ceiling. Furthermore, the performance and cost profile of running hundreds of stateful, long-running agents concurrently is a different scaling challenge than serving stateless chat completions.
The competitive landscape and execution hurdles
The most direct competitor named in the available data is Runloop, though details on that company are sparse [PERPLEXITY SONAR PRO BRIEF]. The broader competitive set includes every major cloud provider's AI orchestration tools (Azure AI Studio, Google Vertex AI) and a host of venture-backed startups globally. Runflow's answer is its focus on the Brazilian enterprise context, its governance-first design, and its embedded distribution through the IFTL network.
However, several hurdles stand between early traction and lasting category leadership.
- The platform dependency trap. Success requires deep, ongoing investment in connector maintenance and expansion. If engineering resources are diverted to chase new AI model features, the core integration value erodes.
- The consultant versus product scale. The company's origin in solving IFTL's own problems is a strength, but it risks building a services-heavy model tailored to a few large clients rather than a scalable product. The reported ARR suggests product revenue, but the line can blur.
- The global incumbent factor. While hyperscalers may move slowly on region-specific compliance, a well-funded international player with a similar governance focus could quickly replicate Runflow's approach in the market.
The company's near-term milestone is clear: hit its R$3.6 million ARR target and prove it can sustain growth while moving beyond its IFTL-affiliated launch network. The next twelve months will test whether its governance and integration wedge is defensible enough to build a standalone platform company, or if it becomes a valuable acquisition target for a larger entity looking for an enterprise foothold in Latin America. For now, Runflow has identified a real pain point,getting AI agents out of the sandbox and into the corporate workflow,and is executing against it with a focused, ground-up build.
Sources
- [runflow.ai] Runflow website | https://runflow.ai/
- [Startups, 2026] Com Runflow, fundadores da LinkAPI lançam ERP para agentes de IA | https://startups.com.br/negocios/inteligencia-artificial/com-runflow-fundadores-da-linkapi-lancam-erp-para-agentes-de-ia/
- [LinkedIn] Joaquim Neto LinkedIn profile | https://br.linkedin.com/in/jocaneto/
- [runflow.io] About Runflow: AI Infrastructure Team & Company Story | https://www.runflow.io/about-us