InKeep raises $13M seed to build no-code/code AI agent builder for technical customer support

Summary

InKeep has closed a $13 million seed round and is targeting a structural gap in how technical companies deploy AI agents for customer support. Founded by Nick (last name not provided in transcript), the company counts Anthropic, Midjourney, Pinecone, and Solana among its early customers.

The core product thesis addresses a friction point that emerged from those early deployments. Support and documentation teams want out-of-the-box tooling that connects to ticketing platforms like Zendesk and messaging tools like Slack, while engineering teams need API-level control, data governance, and custom integration. Those two requirements have historically forced a hard choice between no-code SaaS tools that lock out engineers, or developer-built systems that support teams cannot modify without filing internal requests.

InKeep's answer, launched alongside the fundraise, is a hybrid no-code visual builder paired with a TypeScript SDK. Non-technical teams can construct and iterate on agents independently, while engineers retain the ability to inspect, extend, and control the same workflows in code. The platform connects to shared backend systems including Stripe, CRMs, and support ticketing infrastructure, so customer-facing agents and internal copilots draw from a unified knowledge and data layer.

Architecture is fully agentic, not sequential. Unlike workflow automation platforms such as n8n or Zapier, which follow structured predetermined paths, InKeep's agents use LLMs to determine each step dynamically, decomposing tasks into sub-agents and synthesizing outputs, similar to the approach behind ChatGPT's deep research mode. According to Nick, that design makes the platform better suited to high-variability inputs like support tickets than rule-based alternatives.

The Anthropic relationship surfaces an apparent tension: a company whose flagship product is AI-assisted software engineering still procures a third-party tool for support automation. The explanation is straightforwardly practical. Anthropic's support and documentation teams are not going to build their own chat UI, Zendesk integration, and Slack bot, even if their engineers could. Claude Code, Anthropic's developer tooling, operates at a low level of abstraction intended for software engineers. InKeep positions itself one abstraction layer higher, making agent-building accessible to go-to-market and support functions. The frontier labs also serve as a product signal: Anthropic and OpenAI's own guidance to minimize scaffolding and let LLMs drive decision-making directly shaped how InKeep designed its no-code builder.

The fundraise gives the company runway to expand that customer base beyond the AI-native companies that defined its early traction and into broader technical enterprise accounts where the same engineering-versus-support divide exists.