Every quarter we review our automation and AI stack and it keeps growing. Used to be we had one or two tools. Now we’ve got a workflow platform, an AI API suite, some integration tool, a few specialized services. Each one does something slightly better than the others at that specific thing. But the operational overhead of maintaining all of it is real.
So the obvious question: why hasn’t the market settled on one unified platform? Or why don’t we just standardize on one if they’re claiming to do everything?
I think part of it is lock-in anxiety. You commit to one platform and they own you. Pricing changes, features get deprecated, something breaks and you’re on their support queue. With separate tools, you have leverage and optionality. But that optionality costs money—integrations, data mapping, operational knowledge spread across teams.
Another part is probably that “everything” platforms are still kind of bad at the specialized stuff. They’re good enough at most things, but if you really need performance or specific features in one area, you’ll find a dedicated tool is better. So you end up with hybrid stacks anyway.
But I’m wondering: is the era of the hybrid stack ending? Or is it just shifting? Like, instead of having separate platforms for workflows and AI models, you have one platform that does both, but you still need something else for whatever niche thing comes next?
We’re at the point where consolidating would probably save us money, but it requires a bet that the unified platform won’t hit a dead end in two years. How are other teams thinking about this? Are you consolidating or doubling down on optionality?
We went through this exact deliberation. The funny thing is we started with three platforms, tried to consolidate to one because of cost and complexity, and gradually added back a second platform because the first one hit a wall in one area.
Here’s what we learned: consolidation works if you’re willing to accept 80% solutions. That’s usually fine. But 80% compounds. When the platform is 80% good at your core use case and 80% good at your backup use case, and they break in different ways, you end up frustrated.
We settled on a hybrid: one unified platform for 85% of our workflows and AI needs, and one specialized tool for the 15% where we really need depth. Operational overhead is maybe 30% higher than if we’d stuck with one tool, but our outcomes are better and we’re not underwater if the main platform changes direction.
Lock-in anxiety is real, but we decided the bigger risk was vendor complacency—betting everything on a platform that loses innovation momentum. That risk felt bigger than the operational overhead.
We chose to consolidate about two years ago and haven’t regretted it. We’re not doing anything too exotic—standard workflows, common AI models, standard integrations. For that, one platform handles everything fine, and the operational simplification is worth way more than the 5% performance we’d gain from a specialized tool. The team thinks in one mental model, there’s one contract to manage, one vendor relationship.
The unified platform thing only works if you’re willing to adapt your processes to fit the platform. If you’ve got workflows that are specifically optimized for your old setup, migration is painful regardless of cost. But if you’re building new workflows or you’re willing to reshape existing ones, being on one platform is cleaner. We found that about 60% of our workflows could move easily, 30% needed rethinking, 10% were too specialized to fit anywhere.