The case-study format that converts readers and AI tools

In short

• A case study that brings enquiries reads like evidence, not like a brochure: who the client was, what problem they had, what you did, what changed, and how you know.
• The same qualities help with AI search. Specific numbers, quotations and self-contained facts are what AI assistants find easy to pick up and repeat.
• Use a fixed structure for every project, put the key facts in plain text near the top, and only publish results the client has agreed to share.

Why most case studies do not convert

Many company websites have a portfolio page full of logos and photos with one line of text each: "Brand campaign for a leading retailer." It looks busy, but it answers none of the questions a buyer has. Was the project like mine? How big was it? How long did it take? Did it work?

A buyer reading a case study is trying to reduce risk before they contact you. An AI assistant answering "which agency in Hong Kong has done event projection work?" is doing something similar: it looks for clear, specific statements it can trust. A gallery of images with no text gives neither of them anything to work with.

Google's guidance points the same way. Its page on creating helpful content asks whether content provides original information or analysis, and whether it clearly shows first-hand expertise. A case study is one of the few pages where a business can show that directly, because nobody else has the details of your project.

What Google and the research reward

Google's guidance on writing high-quality reviews is written for product reviews, but most of its advice fits a case study well. It asks writers to:

• show evidence of their own experience, such as visuals or links;
• include quantitative measurements where they matter;
• explain what sets something apart from the alternatives;
• focus on the factors that matter most for the decision;
• care about quality and originality rather than length.

The research behind GEO adds a second reason. The 2023 paper GEO: Generative Engine Optimization found that its best methods, adding statistics, quotations and citations of sources, improved a page's visibility in AI-written answers by 30–40% on its main measure, while keyword stuffing did not help. A good case study already contains those three things: the numbers from the project, a quote from the client, and links to the live work.

Microsoft's advertising team gives practical advice in the same direction. It suggests writing content that AI systems can split into clear pieces: descriptive headings, sentences that make sense on their own, specific facts in place of vague claims, and important details written in HTML rather than only inside images.

A structure you can reuse

Use the same layout for every project so readers, and AI tools, know where to look:

1. A one-line summary. Client type, what you did and the main outcome, in one sentence. For example: "A two-week product launch event for a Hong Kong skincare brand, with projection mapping and a live-streamed opening."
2. A facts block. A short list near the top: industry, location, year, duration, services, deliverables and team size. These are the details buyers scan for and AI assistants quote.
3. The challenge. Two or three sentences on the client's situation in their own terms: the deadline, the budget pressure, the audience they could not reach.
4. What you did. The key decisions and why you made them, not a list of every task. This is where first-hand expertise shows.
5. The results. Specific, agreed numbers where you have them (attendance, enquiries, time saved, pages delivered), and plain-language outcomes where you do not.
6. The evidence. Photos or video of the real work, a short client quote with a name and role if permitted, and links to the live site or coverage.
7. What we would do again. One or two lessons. This makes the page more useful and more believable than a list of successes.
8. A next step. One clear call to action, such as asking for a similar project review.

You can see this kind of layout across our portfolio, with a facts block and evidence for each project.

Writing rules that make it work

• Lead with facts, not adjectives. "Delivered in 14 days for 3,000 guests" says more than "a highly successful event". If you do not have a verified number, leave it out.
• Write each key fact as a full sentence. A sentence like "The campaign ran in English and Traditional Chinese across Hong Kong and Macau" can be quoted on its own; a heading that only says "Scope" cannot.
• Keep facts in text, not only in images. Google advises keeping important content in textual form and making sure any structured data matches the visible text. Results shown only on a slide image are invisible to many tools.
• Publish both languages properly. If your customers ask in Chinese, the Chinese version needs the same facts, not a shorter summary.
• Get written approval. Confirm with the client what may be named and which figures may be published. Accurate, permitted claims build trust; inflated ones damage it.

The same principles apply to FAQ pages, which we covered in how to write FAQ pages that AI cites.

Wrap-up

A case study that converts is a short, well-organised piece of evidence. Give every project the same structure, lead with agreed facts and numbers, show the real work, and write the important details in plain text in both languages. That helps a buyer decide to call you, and it gives AI assistants something accurate to say about your business. If you would like a second opinion on your own case studies, contact us.