The news
A new public dataset from SaaStr — the "SaaStr AI Agent API Report Card" — grades 152 B2B software APIs on six criteria that matter when an AI agent is doing the work: API design, events and streaming support, authentication, rate limits, SDK quality, and agent readiness. The findings, covered by MarTech, reveal that many of the platforms marketers depend on most were simply not built for machine-to-machine coordination.
Our take
This is the infrastructure conversation that's been missing from most AI-in-marketing discussions. Everyone's talking about agents and automation. Almost no one is talking about whether their martech stack can actually support them.
Here's the constant pattern: a team gets excited about automating a workflow — syncing lead data, triggering sequences, updating CRM records — and the idea is solid. Then they hit the API layer. Rate limits that choke under any real volume. Authentication flows designed for a human sitting at a keyboard, not a process running at 2am. Webhooks that don't exist, or exist but aren't reliable. Documentation written for developers who already know the system.
The SaaStr report card is putting a public score on something practitioners have been navigating quietly for years. And the timing matters. As more GTM teams start building agentic workflows — things that don't just trigger one action but chain decisions across systems — the API quality of each platform in your stack becomes load-bearing infrastructure. A weak link doesn't just slow things down. It breaks the whole chain.
The platforms that score well here are going to have a real advantage as AI adoption accelerates. The ones that score poorly are going to become the reason your automation doesn't work — and they'll be slow to explain why.
This isn't a reason to pause AI implementation. It's a reason to be deliberate about where you start. Build your first automations on the platforms with the strongest API foundations. Don't try to orchestrate across six systems when two of them can't reliably talk to anything.
So now what?
- Check your primary platforms against the report card before committing to a multi-system automation. Knowing where the weak links are saves you from building something that looks great in a demo and fails in production.
- Start automations inside strong API ecosystems first. HubSpot to HubSpot, Salesforce to Salesforce — native integrations exist for a reason. Earn your wins there before crossing stack boundaries.
- When a platform you depend on scores poorly, name it as a constraint in your planning. Don't build around a broken API and hope it holds. Either pressure the vendor or route around them.
The teams that win with AI automation aren't the ones with the most ambition — they're the ones who know exactly where their stack will hold and where it will crack.