A data engineer evaluating workflow orchestration tools lands on "Automation for the context era" and has to read through open-source framework cards, a dual-platform section, and four testimonials spanning fintech trading desks and AI infrastructure before understanding what Prefect actually does for them specifically. The hero's tagline ("context era") is a category claim that requires the visitor to already know what Prefect is building toward — it doesn't name the problem the buyer came with. The move is to pick the primary buyer this page is built to convert (the data engineer or MLOps team evaluating Airflow or Dagster alternatives is the larger, more established segment) and build the hero around their specific situation, then introduce Horizon as a second capability further down the page, not as a co-equal product in the first scroll.

A visitor who arrived from a search like "workflow orchestration Python" or "Airflow alternative" reads a headline about "the context era," a subheadline listing two products, and then a button labeled with a brand name. None of those three elements tell them what happens when they click, what they'll see, or why now. The Endpoint testimonial — "73.78% reduction in invoice costs alone, switching from Astronomer" — is the single most conversion-relevant sentence on the page because it names a competitor, names a specific outcome, and names a real switching scenario; that sentence is buried in the testimonial strip below the fold. Pull the Astronomer-to-Prefect switching claim into the hero or directly beneath it, and change the CTA to name the action ("Start free" or "Try Prefect Cloud free") rather than the product.

A data engineer evaluating orchestration reads "10x faster integration at Nitorum Capital" and has to click through to discover that's an AI agent use case, not a pipeline use case. The testimonials have the same problem: the Cash App quote is about infrastructure architecture, the Nitorum Capital quote is about MCP and LLM chat UI, and the Endpoint quote is about cost reduction from Astronomer — three different buying situations presented as a single undifferentiated proof block. Segment the proof by buyer type or by product, so the data engineer reads the orchestration evidence and the AI engineer reads the Horizon evidence, rather than both reading everything and connecting nothing.

The implicit argument is: open source is popular, therefore the cloud platform is trustworthy and worth paying for. But the page never makes that argument explicitly — it presents the open-source frameworks as standalone products with their own feature lists and license badges, then pivots to "Two platforms. One mission." without a bridge. The visitor who came to evaluate the open-source tool has no reason to scroll into the platform section; the visitor who came to evaluate the paid platform has to read through framework documentation before reaching pricing-relevant content. Add a single transition sentence or CTA between the open-source section and the platforms section that names the specific upgrade moment: what breaks at scale with self-hosted that Cloud solves, in one sentence.