Where do people search today?

The search experience used to start on traditional engines like Google, Bing, and DuckDuckGo. Today, it can start in an AI engine like ChatGPT or Claude, on social media Explore tabs, and directly in many of the apps, websites, and platforms we use every day.

How do people search today?

People used to write succinct queries (“heart attack signs”), open a result — or a few — and read the content to get their answer. Today, searchers use longer, more contextually rich prompts (like “does fast heartbeat and sweating mean I’m having a heart attack?”), and expect AI to surface highly personalized answers.

In other words, we aren’t reading on-page content as often as we used to. Machines are reading it for us, synthesizing the information, and providing an answer that closely matches the intent of our prompt.

Do people really trust AI for health information, Ahava?

Yep, they really do. Nearly 3 out of 4 people, to be exact. Check out:

• Our original research about AI trust and usage in healthcare search
• 6 ways to humanize the algorithmic experience and build trust today

How have search engine appearances changed?

Organic search results used to only compete with paid results and maybe a SERP feature or two. Today, AI-generated results (like Google’s AI Overviews) and several SERP features push organic results farther down the page.

Organic results often land beneath the fold now — where users are less likely to see and click on them.

Why is my organic traffic down?

Because it’s harder for users to see your organic results on SERPs now, and they often don’t even need to look for them. AI provides many of the answers — both in traditional search engines and in AI answer engines.

If you’re seeing fewer clicks from search, you aren’t alone. Read this: What’s Happening in Search? What Healthcare Marketers Need to Know [Webinar]

The shift: Treat content as a data layer

For content teams and website owners, we can’t keep creating content with the expectation that people will see our source material. We need to operate with the understanding that people will frequently interact with our content via AI.

Our content is already being trimmed, flipped, reworded, mashed up, and delivered to searchers in a way that’s meant to best serve their search intent. They often don’t see our full atrial fibrillation condition page — they see a personalized remix of the section or two that best answers their question.

Your most important task right now is to get your content ready to perform as a data layer, feeding information to both people and machines.

CRISP: How to prepare your content for agentic AI

I developed the CRISP (Conversational, Retrievable, Interoperable, Structured, Personalized) framework to help healthcare marketers prepare their brands’ content to survive in an AI-first world. Think human-consumable, machine-readable, AI-scalable, and personalized brand experiences in every corner of the web.

In other words, humans create content for other humans and structure it for machines. Then, machines disperse it to the appropriate channels — ready to answer human questions using rules and code to govern those interactions.

What does this acronym look like in practice?

Conversational

Content needs to answer real patient, customer, and stakeholder questions using current, approved, reader-friendly content. To do this, write content in Q&A or intent-based formats that mirror how people ask questions, like “Do I need a referral?” or “How does this integrate with Salesforce?”

How to create conversational content:

• Break content into answer blocks: Treat condition, treatment, and service line pages like FAQ pages with questions and answers related to the H1 title.
• Front-load answers: Put the clear, direct answer in the first 1-2 sentences under a heading. Don’t make AI (or users) dig.
• Use consistent language: Standardize how you describe services, conditions, and policies. (Because AI compares content across your site and penalizes inconsistencies.)
• Govern your answers: Regularly review content to ensure accuracy, tone, and compliance.

Want to learn more? Read about AI search optimization strategies.

Retrievable

Content needs to be chunked, tagged, versioned, and indexable as a single source of truth — quickly and easily. If AI can’t extract your answer in seconds, you won’t be the answer.

How to make your content retrievable:

• Break content into components: Turn long pages into modular components, such as definitions, steps, FAQs, and disclosures.
• Tag content: Give each component its own metadata, version history, and unique identifier, so AI systems can retrieve only the relevant answer rather than an entire document.
• Structure content so it can be pulled into AI answers: Use clear headings, tight answers, and supporting details — not fluffy intros.
• Use 1 question per section: Avoid mixing multiple ideas in long paragraphs.

Interoperable

Content needs to work with apps, APIs, CRMs, EHRs, and portals via open, documented schemas. Integrate APIs or structured feeds that allow website content to appear in all other platforms without manual copying and pasting.

How to make your content interoperable:

• Standardize key terms and definitions: Define services, conditions, and programs once and reuse consistently. Your content needs to mean the same thing across platforms, so it can work everywhere.
• Align messaging across platforms: Ensure website, social, and other channels describe things in the same way.
• Create a shared content model or taxonomy: Establish a single source of truth for core entities — the people, places, services, and conditions your content is fundamentally about.
• Implement governance and ownership: Assign responsibility for maintaining consistency over time.

Structured

Content needs to use shared metadata, schema, and taxonomy to power search, AI, and personalization. In other words, structure your content to live in fields, not just paragraphs, using schema markup. Schema is a specific vocabulary used to tell search engines exactly what is on a webpage.

Structure turns content into something AI can understand and trust. It tells machines what the content is, who it’s for, where it applies, and how it can be safely used.

How to structure your content for AI retrieval:

• Break content into labeled sections: Use clear headings like “What is …” “Who treats …” “When should …” instead of long narratives.
• Define core entities consistently: Standardize services, conditions, providers, and locations across all content.
• Use schema markup and metadata: Use structured data so machines understand what each element represents.
• Align content with templates: Build repeatable structures for each page type (service, provider, condition pages, and more) with consistent fields for schema.

Personalized

Content needs to match your audience, journey stage, and context across channels. To do this, incorporate audience, lifecycle, and contextual signals into your content. Systems use these fields to dynamically select the appropriate version or explanation without you needing to rewrite the underlying content.

How to create content that supports personalization:

• Tag content by audience, intent, and journey stage: For example, tags could include new patient vs. returning, prospect vs. customer, or symptom vs. treatment decision.
• Create modular content components: Break content into reusable pieces that can be assembled dynamically. This helps AI pull from relevant parts of your content to create personalized answers.
• Define variations for key scenarios: For example, variations could be location-specific, pediatric vs. adult, or urgent vs. non-urgent.
• Integrate with systems that use context: Ensure content can be delivered through AI tools, portals, and digital assistants.

Healthcare content before CRISP

As we move toward agentic AI for patient navigation and access, we need to prepare our content to work in that landscape.

Most healthcare content today is:

• Written for static webpages
• Owned and duplicated across teams — marketing, clinical, patient education
• Created fresh for every channel, rather than reused
• Outdated and not consistently governed
• Risky for AI to interpret without context

In practice, this looks like a webpage titled “Knee Replacement Surgery,” with a long narrative detailing eligibility criteria, pre-op instructions, recovery timelines, surgeon bios, insurance disclaimers, and scheduling links. That information then gets rewritten, reapproved, and republished for other channels — how tedious!

Healthcare content after CRISP

Future-ready website content that follows the CRISP framework looks and works like this:

• Eligibility criteria are stored as structured clinical guidance.
• Pre-op instructions are modular care steps.
• Insurance disclaimers are version-controlled compliance objects.
• Scheduling details are tagged by location.

The H1 defines what the page is about (“Knee Replacement Surgery”). The H2s answer people’s questions: “Do I need knee replacement surgery?” “How long does it take to recover from knee replacement surgery?” “What are the risks of knee replacement surgery?”

Under that structure, when a patient prompts, “Do I need a referral for knee surgery in Chicago?” the AI agent retrieves the relevant approved components, rather than crawling the page to generate a response from scratch.

Creating future-ready content takes a team

As much as we’d like to own every layer of our content as marketers, we simply can’t prepare our content for agentic AI alone. CRISP is a team sport.

In most organizations, marketing owns the conversational, retrievable, and personalized layers. Digital, IT, or product teams own the interoperable and structured layers. To execute CRISP, these teams have to collaborate — with clear roles and responsibilities documented and followed.

This brings us to your next steps: what you can do right away.

Next steps to prepare your content for the future

First, don’t set out to build Rome in a day. You don’t need to rewrite and reformat all of your content immediately. CRISP is a helpful framework to follow for future content — not a mandate to tear down and rebuild your website.

Here’s what you should focus on right now:

1. Bring your teams together: Align marketing, digital, IT, and clinical stakeholders around a shared content strategy. Talk about how the agentic web will impact patients and how you’ll adapt to prepare for it as an organization — and how your organization will adapt to prepare for it.
2. Audit a high-impact section of your website: Choose one service line and audit the content for AI readiness. Is the content modular? Dynamic? Tagged with appropriate schema? Contextually rich? Version-controlled with governance? Accessible via API?
3. Start building a content model: Define key entities (such as services, conditions, providers) and how they connect across platforms. For example, a physician entity might include name, specialty, location, and accepted insurance. Once implemented, this information can be pulled into a provider directory, service line page, and patient portal without being rewritten each time.

If you don’t feel ready, don’t worry. We’ve got your back with the on-page content and the content operations side — including systems and workflows to support cross-functional collaboration. Our sister agency can help with any technical elements that we don’t cover.

In an AI-first world, your content isn’t just what you publish. It’s what powers everything. And our team is here to power you.

Questions? Let’s talk about it.