Evaluating AgentKit: avoiding YAPPy solutions in wealth firms
OpenAI’s new AgentKit is a fascinating new entrant in the integrated workflow automation space — joining Microsoft Power Automate, Azure Logic Apps, Workato, n8n and Zapier.
The number-one question on my mind as I evaluate any tool like AgentKit for wealth firms: is it YAPPy — Yet Another Proprietary Platform?
When I help firms design their technical architecture, one guiding principle is always how YAPPy a solution is. The more proprietary a tool, the more tightly you’re bound to its ecosystem, pricing model and pace of innovation.
In an advice firm, your pillar systems are already well-defined: CRM, document storage, email/calendar/team collaboration, portfolio management, custody. These are systems of record — critical stores of data that justify deep, strategic vendor relationships. You invest in them because they’re foundational to compliance, client service and scalability.
But workflow automation is different. It’s not a store. It’s a pipe between stores — an orchestration layer. And because it doesn’t hold core data, it’s an opportunity to stay open, portable and minimise long-term lock-in.
That means looking carefully at how you automate:
👉 favouring open protocols over proprietary connectors
👉 choosing platforms that are exportable, self-hostable, even open source
👉 avoiding builders that capture logic in unreadable formats
👉 favouring transparent pricing over opaque “call-for-quote” models
👉 leaning into tools that can evolve as your architecture does
👉 choosing ecosystems with large, active communities for support — human and AI alike.
AgentKit is still emerging — and it may become a powerful new option for wealth firms. But my evaluation lens remains the same: Does it make your architecture more flexible and resilient, or more YAPPy?
Because pipes should be open, well supported and replaceable. That’s how you manage risk and protect optionality in a fast-moving tech world. 🙂
Originally shared on LinkedIn