Website operations
When treatment information exists only in images, where should you start?
A practical sequence for moving essential patient information into page text without removing every long information graphic at once.
An attractive screen and readable information
Consider this hypothetical operating scenario: first-visit guidance is contained in one long image, so every wording change requires rebuilding the design file. The first task is not to remove all graphics, but to identify information that people frequently check and staff frequently update.
Google recommends making important information available as text among its basic practices for AI features in Search.[1] This should not be interpreted to mean that every AI system is unable to read images.
Choose what to move first
Extract the text from the image and have the person responsible for the content check it against current information. Copying outdated opening hours or missing conditions simply gives old information a new format. The following is also a hypothetical example.
| Information in the image | How to present it in page text |
|---|---|
| Opening hours by day | A table separating days and times |
| Preparing for a first visit | A readable list of preparation items |
| An explanation of treatment stages | Paragraphs for each stage, with conditions for applicability |
Describe images as images
Diagrams and photographs of the premises can remain beside the text. Put the sentences patients need to understand in the page content, and add descriptive alternative text and captions where needed for the images.
Google explains that it uses alternative text and surrounding context to understand images, and advises against filling alternative text with repeated keywords.[2] We do not recommend placing the entire treatment explanation in a single alternative-text field.
Check the content before a button is pressed
Being able to select words on screen is a simple initial check, but it does not establish whether search engines can collect the content. Ask the developer to check whether the essential explanation appears in the HTML rendered by the search engine.
Google can execute JavaScript to process content. However, not every bot runs it in the same way, so serving important content from the server or prerendering it can also be considered.[3]
Complete one page first
Choose one first-visit guidance page and compare the original image with the new text side by side. Dividing sentences into shorter sections must not remove conditions or points to check before making an inquiry.
- Does the title alone make the purpose of the guide clear?
- Can the main information be read without enlarging an image?
- Do links to related guidance open the intended pages?
Use the same source copy for later updates
If the same operating information remains in both text and images, assign an owner and a source document for updates. Editing the page text while leaving the banner unchanged creates conflicting information on a single page.
Record the wording changed and the date checked, and read the page on mobile too. Converting information to text is a content management and communication task. It cannot, by itself, guarantee search rankings or citations in AI answers.