Short teaching sections set up practical decisions the group can use immediately.

Policy, platform, governance, and ownership questions get separated from quick content edits.

Can a real visitor understand the page quickly enough to keep moving?

Which fixes belong to content owners, and which need platform, policy, or brand decisions?

Make pages easier for people to scan, understand, trust, and act on.

CWRU leaves with rules it can reuse beyond the pages discussed in the room.

A shared page-quality framework you can use immediately.

A short list of priority changes CWRU can standardize after the training.

Focus first on decisions content owners can change.

Distinguish quick content fixes from policy, platform, and governance questions.

Which kind of web reader are you: skimmer, swimmer, or diver?

Use this lens when reviewing CWRU pages.

A useful page answers a real reader question.

It removes institutional clutter that does not serve the page job.

A reader should understand the page from the headings alone.

Use headings to expose structure, not add decoration.

List the current H1/H2 sequence.

Fix headings before rewriting paragraphs.

Automated extracts show some pages include navigation headings before content headings.

For AI/search, meaningful content headings still matter because they create retrievable sections.

Use precise institutional language where it matters.

Do not make expert content vague in the name of simplicity.

Put the next action near the decision point.

Do not make readers infer the destination from surrounding copy.

Focus on content decisions the room can control.

Accessibility, usability, and AI-readable clarity often improve together.

Useful alt text names the meaningful content of the image.

It does not stuff keywords or turn into a marketing paragraph.

Decorative image that adds no information.

Avoid forcing screen-reader users through irrelevant image details.

Image conveys unique information.

Image is also a link or action trigger.

The destination should be clear if the link text is read by itself.

Specific links help accessibility and page comprehension.

Use lists for grouped items that do not need row/column relationships.

Do not use tables for layout.

Broad and exploratory.

'Engineering at CWRU combines research, design, and hands-on problem solving from year one.'

Add a short section under the hero: what students can study, how they learn, and where to go next.

Link to makerspace, co-op, and first-year engineering experience with descriptive anchors.

Nurses comparing specialties and application paths.

'Choose the MSN path that matches your nursing goals.'

Lead with who the program is for and how specialty choice works.

Use Frances Payne Bolton School of Nursing, Master of Science in Nursing, and nurse practitioner specialties consistently.

Add format, location, audience, and professional outcome.

Use proof sparingly and connect it to applicant decisions.

Multiple giving methods are visible.

AI/search is secondary; this is primarily UX writing and accessibility.

Add a 'Choose the giving method that works for you' intro.

Add a 'Why give to CWRU' path that connects philanthropy to outcomes and institutional momentum.

Useful pages begin with purpose.

The same clarity carries into Day 2, but AI systems interpret it through different mechanisms.

Yesterday was about human comprehension.

It raises the cost of vague, inconsistent, disconnected content.

Retrieve and rank documents.

Make the facts, entities, relationships, and evidence clear enough to be retrieved, understood, and corroborated.

Index, retrieve, rank, snippet, link.

A hybrid of retrieval and generation.

The model answers using learned patterns and may provide no citation.

AMA deck slides on training data vs. grounding.

Pulls from Google's index and search-quality systems.

'And Google AI Overviews?' from AMA AI Search deck.

Keywords still matter, but alone they are not enough.

External corroboration matters because AI systems recognize patterns across sources.

Ubiquity increases probability.

UPCEA 2026 Winning AI Search, slide 16.

Prospects rely on search engines and university websites.

CWRU pages have to serve search, site visitors, and AI-assisted discovery.

50% of prospects use AI-powered tools at least weekly.

Use these as orientation stats, not as a reason to abandon web fundamentals.

You have to be found to be considered.

AMA 2025 AI Search deck, slide 14.

Weak SEO vs. strong SEO.

Not every AI answer is search-index grounded, but strong public SEO signals still matter.

Subject -> Predicate -> Object.

Use triples as an editing tool, not as visible jargon for every page.

Newsroom stories connect people, topics, discoveries, gifts, rankings, research, and institutional momentum.

They give AI systems repeated patterns around what CWRU is known for.

Name the institution consistently.

Use boilerplate intentionally so external pickup carries the right facts.

A CWRU page says what CWRU says about itself.

This is why PR, newsroom, and SEO should not be separated in the AI-search conversation.

Consistent verbiage used in each release or talking points about the university, program, or priority topic.

UPCEA webinar slide on PR and media best practices for AI search.

LinkedIn posts and articles can reinforce program/faculty/newsroom language.

The point is not volume; the point is consistent, useful, corroborated signals.

Ask AI for its sources.

Fix the owned source first, then reinforce with stronger external signals.

AI visibility tools are directional.

Use tools to find page priorities and citation gaps, then apply editorial judgment.

Question-style queries show what users ask before they reach the page.

AMA 2025 slide 48.

Prompt around priority audiences and real decision moments.

Turn findings into page edits, citations, or newsroom/PR priorities.

The tool should support the content decision, not become the focus.

Use examples that keep the discussion focused on findings and next actions.

Page/content issue.

Pull from live examples discussed across both days.

What changed in how you read your own pages?

What question should be answered in the written recommendations?