For instance, a sales page filled with hype but no substance makes me skeptical.

So for my sales page, I decided to scatter supporting facts and figures throughout to increase conversion rates.

Obviously using AI research capabilities to find exactly what I needed!

Let's get started

Evidence: The Difference Between Saying and Proving

In Part 3, we refined our basic content structure and flow. Now we're going to add an additional layer to step it up.

This step is optional depending on your content type, but it's particularly valuable for

• Sales pages (to back up claims and reduce skepticism)

• White papers and case studies (to establish authority)

• Blog posts and articles (to increase trustworthiness)

• Presentations and pitch decks (to convince stakeholders)

• Video scripts (to add credibility to verbal claims)

Not every piece of content needs this level of evidence, however. For instance, personal stories, creative writing, routine updates, or internal communications might not require extensive research. So consider this step optional.

Identifying Claims That Need Support

The first step is finding every assertion in your content that would benefit from supporting evidence.

Basically what in your sales page or document is going to raise eyebrows??

Here's the prompt I use

Please identify all claims, assertions, or statements in this draft that would be strengthened by supporting data, statistics, case studies, or expert quotes. 

This creates a clear shopping list of places where you need to add credibility through evidence. Good first step.

And yes…we are 100% going to look for evidence that supports our conclusions. Cheeky? Yes. But we’re not doing balanced academic research here. If you DO need more balanced arguments specify this in the research step.

Leveraging AI Research Tools

So we could go off and find our own supporting evidence.

This is where having access to the right AI research capabilities becomes crucial. At the time of writing, I recommend:

• ChatGPT Deep Research

• DeepSeek (R1) - particularly strong for in-depth reasoning and research, and it's free to use (horray!)

The key is to use models with strong research skills, ideally more than one since they'll often find different information from various sources. You can combine all their findings.

Whatever the best research tools are when you're reading this, use those. This space is evolving rapidly, but the approach remains consistent.

The Research Process

For each claim that needs support, craft a focused research prompt. Here's a generic template you can adapt to any topic:

Find credible information about [your specific topic], including data points, case studies, expert opinions, and industry examples. Focus particularly on [specific aspect you need]. Include sources for all information provided.

The key to effective research prompts is specificity

• Particular industries or contexts

• The type of information you need (statistics, case studies, examples, quotes)

• Credible source requirements (academic, industry reports, news articles)

Note: you can also ask the models to go and find support data for ALL the points you need to make but be aware that it’ll find less info per point. Experiment to your own needs here.

Information Synthesis and Integration

When using specialised research tools like ChatGPT's deep research capability or DeepSeek's R1 model, you'll often get extensive results. Here's how to bring it all together:

1. Copy all the research from your research tool back to your primary chat (this is one step where you will need to use your keyboard)

2. Use this prompt with the pasted research

Using the research copied below, please distill the 5-7 most compelling and relevant pieces of evidence that match my shopping list of claims needing support. For each data point, provide:The specific claim from my shopping list this evidence supports
*The statistic or finding itself
*The source
*A suggestion for how to incorporate it naturally into my content

3. Finally, integrate the evidence

Please integrate these statistics and evidence points into my draft. Ensure they flow naturally within the existing text and include appropriate attributions.

For my sales page, this process allowed me to transform general statements like "AI automation helps businesses operate more efficiently" into specific, credible claims like "Businesses implementing AI automation solutions report an average 35% reduction in process time and 28% decrease in operational costs (Harvard Business Review, 2024)." and even provide a link to the research for the very analytic!

That's the difference between making a claim and proving it.

And

Identifying Supporting Materials

Content is rarely just data and statistics. Visual elements like charts, graphs, infographics, and images can dramatically enhance credibility and engagement.

While AI can't create most of these elements for you (yet), it can identify exactly what supporting materials would strengthen your content.

Try this prompt

What supporting materials (images, graphics, charts, videos) would enhance this content? For each suggestion, please specify:

*What the element should show/contain
*Where it should be placed in the content
*Why it would be valuable to include

For my sales page, this analysis suggested

• A comparison chart visualising efficiency gains before and after automation

• Specific places where client testimonials would address potential objections

• Areas where process diagrams would clarify complex concepts

• Suggestions for before/after scenarios that would demonstrate transformation

While you'll need to create or gather these materials separately, having a list of targets to create makes the process much more efficient.

Strengthening Social Proof

Perhaps the most powerful form of evidence is social proof—testimonials, case studies, and examples from others who have experienced what you're describing. I’ve written whole Playbooks on this elsewhere - it’s super important. But let’s cover some additional help here.

If you already have testimonials or case studies, you can ask the AI to identify the optimal places to include them:

Here are testimonials from previous clients. Please suggest the best places to integrate them into my content to address potential objections and strengthen key claims

Use this alongside a copy/paste or page full of your prior testimonials and reviews and the AI will pick those that’ll back your arguments up.

If you don't have testimonials yet, the AI can help you identify what kinds would be most valuable:

Based on this content, what specific types of testimonials or case studies would be most effective in building credibility? For each suggestion, explain what kind of client/result would be ideal and what specific concern it would address.

This analysis can guide your customer feedback process, helping you gather testimonials that directly address the specific objections or concerns potential customers might have.

Validation: The Final Credibility Check

OK one final piece! Once you've integrated all your evidence and supporting materials, it's time for a final credibility check:

Please evaluate this content from a skeptical reader's perspective. Are there any remaining claims that lack sufficient evidence? Are any statistics or quotes used in misleading ways? Is the evidence balanced and fairly presented? What else could be added to make this content even more credible and convincing?

This helps identify any remaining weaknesses before you finalise your content.

Remember, this is also an excellent time for a human review. Show your content to another person (shock horror!); a colleague, friend, or even potential customers. Consider posting parts of it on social media to gather feedback as you "build in public."

What's Next?

In Part 5, we'll explore how to prepare your content for deployment and discover how this approach applies to various content types beyond sales pages.

Keep Prompting,

Kyle