I don’t usually talk about what is around the corner. I am more about understanding history, and telling stories in the present. However, after several cycles now I am feeling like can develop a “vibe” for how this moment will shake out around the corner. Meaning, as the current artificial intelligence bubble reaches peak hype and begins shrinking back down to reasonable levels of investment and building, what can we expend in coming months and years when it comes to the integration and automation conversaitn.
AI isn’t going away. I saw someone compare it to asbestos — we’ll be digging it out for many years to come. There are also good use cases of artificial intelligence, although I am not in the business of championing those — I’ll let you find and champion them. I am in the business of equipping SMB, SME, and enterprises with what they need to deal with the past, present, and future, so I feel it is important to provide guidance in all three areas.
So, here are how things will shake out with the AI bubble. It is going to come back down to a reasonable size. There will be plenty of folks doing the work to patch over the pop, so I don’t know if I’d call it a “soft landing”, but there is a ton of money on the table and a ton of eating crow needed in a market of people who aren’t very good at doing it—so they’ll do whatever it takes to save face. From a business standpoint things are going to hurt. Lots of middle management who made decisions in last couple of years will lose their jobs and move on to green pastures before the accountability takes hold, with investors getting a haircut.
From a technological standpoint, you’ll see AI shrink back down to a realistic size in the wider application, system, and cloud integration and automation toolbox alongside other machine learning models. Think GraphQL, came on big, didn’t live up to hype, but still very useful in some situations—AI will do similar, but at a different scale. We’ll see lots of small language models and other proven machine learning models living alongside HTTP APIs, GraphQL, event-driven, and RPC interfaces in our diverse API toolbox. This is a world where discovery, rating, routing, and observability across all tools in your integration and automation toolbox will matter a lot.
A2A will outpace MCP, and find it’s place alongside OpenAPI, Overlays, Arazzo, AsyncAPI, JSON Schema, GraphQL, and Protocol Buffers. Knowing your spec will continue to be important in how you apply your integration and automation toolbox. We will continue to perpetually live in an XKCD cartoon and every cycle new youngsters who are unaware of what already is will come along and add a new spec because they were told to. Moving forward we will see more trends come along, but none will be as big as AI, but it won’t stop folks from trying. All while SMB, SME, and enterprise leadership continue to struggle with the sprawl adopted across each moment like we find ourselves in today.
