Building an AI-Powered Website Workflow for the Academy of Natural Sciences

The Challenge

The Academy of Natural Sciences of Drexel University was preparing to launch a redesigned website built with a flexible library of content components. Think of them as building blocks: depending on the page, the team can combine different modules to structure and present the content.

That flexibility gave the team more options for creating pages, but it also added a new step to an already busy website workflow. For every new page, Web Coordinator Ramon Torres needed to determine which components made sense, then account for additional page requirements like title tags and meta descriptions.

Capacity saw an opportunity to make that process faster. Through our AI Strategy partnership, we worked with the Academy to build a custom AI tool that could recommend the right website components and generate supporting page requirements, with Capacity’s SEO expertise built directly into the workflow.

Finding the Right Job for AI

We started by learning how the Academy’s marketing team actually works.

Ramon Torres coordinates the website, managing page builds and updates that come from both the marketing team and departments across the Academy. During our discovery process, the team described just how central this work is to his role.

The redesigned website introduced another consideration. Instead of relying primarily on fixed page templates, the site uses components as building blocks that can be combined to create individual pages.

More flexibility meant more choices each time Ramon built a page.

We identified this as a strong AI use case because the task was frequent and governed by a defined set of rules. We gave AI the Academy’s website documentation and taught it how to apply those rules to a specific page.

Turning Website Documentation Into a Practical Tool

The website’s technical documentation contains detailed guidance for its component library, including when individual modules should be used, which content types they correspond to, and how they function within the CMS.

Capacity translated that documentation into instructions an AI tool could use as part of Ramon’s regular workflow.

When Ramon provides the content for a new page, the tool can use that information to recommend appropriate components based on the Academy’s actual website system. The goal is to give him a useful starting point without requiring him to reference extensive documentation for every page.

Building Capacity’s Expertise Into the Workflow

A useful AI tool needs clear instructions for how to apply the information it receives.

We designed the tool around the Academy’s specific website architecture and layered in Capacity’s SEO expertise and knowledge of the Academy from years of partnership. The output is grounded in the organization’s real component library while also accounting for the requirements that help each page perform in search.

This was especially important because the Academy’s component system includes different rules for different page and content types. Some headers, for example, are intended only for specific types of pages, while other modules can be used more broadly throughout the site.

By codifying those distinctions, we made the tool specific to the Academy rather than relying on generic AI recommendations.

Testing With the Person Who Will Actually Use It

We built the tool to fit Ramon’s workflow, so his feedback was essential.

Instead of treating the first output as finished, Capacity tested the tool with Ramon and refined its behavior based on how its recommendations worked in practice. That process gave us a chance to identify where the AI understood the component system correctly, where its recommendations needed adjustment, and how the output could become more useful for day-to-day page building.

That human review remains part of the process. The tool provides a structured starting point, while Ramon brings the context and judgment required to build the final page.

Making More Room for Strategic Website Work

The potential impact of a few minutes saved on a single page becomes much more meaningful at scale.

The Academy builds approximately 30 pages each month and estimates that choosing components takes about five minutes per page. That adds up to roughly 30 hours of work each year. With the custom tool, the team estimates the potential to cut that time in half.

Those hours have somewhere valuable to go. During discovery, Ramon told us that more available time would allow him to think more strategically about website decisions, including what belongs on the site, how content is worded, and how visitors use it.

That’s the opportunity we look for in our AI Strategy work: a specific application that solves a real problem inside an existing workflow. For the Academy of Natural Sciences, a repetitive part of building website pages became a chance to put institutional knowledge and Capacity’s expertise directly into the team’s hands.

Capacity helps arts and culture organizations find practical ways to put AI to work. Our AI Strategy partnerships can help you identify the right opportunities, build tools around your team’s real workflows, and create more capacity for the work that needs your expertise.