For the last year I have been integrating systems and creating automated workflows using a wide variety of software platforms. My goal has been to learn how to leverage APIs & automated workflows to streamline operations within SaaS companies, and ended up with my own platform in the process. There's a lot of discussion in tech circles about the current and future state of system integrations and workflows with emerging technologies, and many of these conversations are led by business leaders that can't write code. Here's what I've learned so far:
SaaS companies must have secure and reliable integrations to stay relevant. SMBs manage a large number of software platforms, the ones that provide data visibility without requiring logging into their dashboard has the edge. The expectation is seamless communication and data sharing via APIs or enabling secured access without sign-in through authentication methods like OAuth.
Workflow automations are difficult to get right. When done right, they can streamline portions of an SMB's operations, but are widely under-utilized and often misused. Many business leaders don't understand how to design workflows that actually improve efficiency while maintaining visibility of the process. Startups and small businesses looking for automations often turn to modern platforms such as Zapier and Make.
Automation platforms have restrictions for growing SMBs. They enable users to build and implement their own automated workflows, but require time investment to learn and have limitations with complex workflows and high volumes. They are typically sufficient for startups and SMBs with low volume, non-critical workflows, but become economically inefficient for high volume businesses like eCommerce or SaaS that need them for revenue driving tasks. Additionally, most platforms "fail silently" and kill workflow execution without notification, which can result in loss of potential clients.
AI implementation in backend workflows is difficult. The barrier to entry for AI implementation into backend workflows is beyond the capabilities of most SMBs. Implementation through retrieval augmented generation (RAG) requires more advanced database architecture and dedicated workflows to create & manage tokenized context.
"Vibe coding" is fascinating as a concept, but not much else beyond that. This new craze has taken the scene by storm, but evidence of completely revolutionizing software automations is not apparent (yet). Most coding models excel at building landing pages, simple apps, etc., but struggle with context when creating larger projects. For the most part AI is nearly capable of replacing frontend developers, but the demand for backend engineers and architects (who can leverage AI) is still extremely high.
Growing SMBs need more control than automation workflow platforms can provide, but they can't afford IT teams to create & manage their own. I'm excited to see how I can help tackle this problem by providing custom automation solutions.
