PRE-ORDER: The Science-based Playbook of AI Best Practices

⏳ PRE-ORDER OFFER: Buy before launch (January 29th) and get an early bird gift - an exclusive case study where I tackle a real-life AI marketing question using the latest scientific evidence ⏳

🎓 The science of using AI in your marketing 

Everyone has an opinion about AI. But opinions don’t drive revenue - facts do. 

Scientists from the world’s best universities have spent the past years researching when AI works best and how people respond to it.

We’ve turned this scientific evidence into this practical playbook, so you can make fact-based decisions. It answers questions such as:

• Does an AI chatbot fit your brand, or will it backfire?
• How do you increase trust in your AI’s recommendations?
• When does AI improve your marketing (vs is just a distraction)?

And so much more. 

📘 What’s in the Playbook

1) Science-based guide for using AI in chatbots, product recommendations, and optimizing your marketing

2) Insights on how to get AI to deliver best results

3) Practical examples and field guides

4) The future directions of AI and upcoming research

For example, you’ll learn from studies such as:

• Use AI to deliver bad news (e.g. price increase) (Garvey et al., 2023)
• Never use AI for complex customer service issues (Lou, et al., 2022)
• Tell people that your AI is still learning, to boost trust (Chacon et al., 2025)
  

I created this Playbook in collaboration with Stefano Puntoni, Professor of Marketing and Co-Director of AI at Wharton Business School.

✅ Is this Playbook for you?

This playbook will be useful if you’re using or planning to use AI in your marketing or your products. It’s meant for: 

• Marketers looking to improve their operations and results (e.g. streamline customer support, give higher-converting ecommerce product recommendations)
• Product managers and designers looking to improve UX / UI (e.g. building more effective AI chatbots)
• Business owners looking to optimize parts of the business (e.g. determine when to turn to AI vs hiring new customer support agents)

❌ What you should not expect

• A beginner's guide to understanding or using AI. 
• Basics or a guide to generative AI (e.g. ChatGPT) prompting. These are insights on when and how to best leverage AI to grow your business. 
• Answers to all AI-related questions. The playbook focuses on the three currently most researched and most popular AI use cases - chatbots, product recommendations, and marketing processes.
• Opinions or generic, vague “advice from experts”. This is science.
  

📈 How to use it

Use these recommendations to make better, evidence-based decisions about how and when to use AI in your marketing or products.

When possible, test any changes on a small scale before implementing them widely. Differences in the context of your business can make these effects unexpectedly weaker or stronger.

A (non-exhaustive) list of the areas the playbook helps you with:

1. Building AI chatbots that customers will like
2. The best chatbot personas for your business
3. Avoiding uncanny valley and AI aversion
4. Deciding when to use AI vs. stick to humans
5. Giving AI recommendations that drive sales
6. Improving user trust and adoption rates
7. Handling sensitive and private information
8. Running more effective customer research
9. Creating effective content, faster and cheaper
10. Understanding the future of AI impact in business

❓ FAQs

Q: Can I find this research elsewhere?

Yes, you can analyze the research papers yourself (and should, if there’s one you’re deep diving into) in scientific journals such as the Journal of Marketing or Marketing Science. All studies are referenced and linked.

But unless you are someone who reviews the 10,000+ marketing studies published each year, buys each research paper for $30+, and knows how to interpret academic research - you will have never heard of most of these techniques. 

Q: Is this for beginner or advanced AI users?

This playbook will not teach you what AI is or how to use or prompt LLM’s. The insights are meant to give high-level, strategic guidance on how to approach applying AI to your business and your growth goals. 

If you are just starting out using AI or you’re an advanced AI user, as long as your goal is to improve and grow your business or optimize user experience, you will find useful insights that will help to make AI decisions with less guesswork and more confidence.

As always, no one strategy fits all, no insight will perform in all cases. Be critical about implementing any changes and heavily A/B test variations of the recommendations (e.g. different messaging and designs).

Q: Is this just a collection of insights covered in the Science Says newsletter?

No. Some of the insights in the playbook have been covered in the newsletter, others are brought in from new research papers. Science Says insights in the playbook are enriched with more examples, focused on AI best practices, and explained so that you can rapidly apply them to your business.

You can subscribe to the Science Says Platform to access the library of hundreds of insights.

Q: What kind of research is included in the playbook?

All the studies covered in the playbook have survived rigorous statistical analyses, peer-review (i.e. picked apart and challenged by other scientists), and been published in a reputable scientific journal (e.g. Journal of Marketing Research).

Q: What if I don’t see any results?

You get your money back. If you tested techniques and didn’t see a statistically significant lift in conversions you will get a refund (within 90 days of purchase).

I’d appreciate it if you can share how you applied what didn’t work, so academic researchers can use the learning to advance our scientific understanding, but this is optional.

Q: Why is this playbook more expensive than others?

The type of research we’ve covered in this playbook took more specific and complex resources than any previous one to analyze and put together into a playbook.

Additionally, we were working with the newest publications which just hit the market recently. We did not have the chance to collect the research along the way of writing insights and doing our work as usual - this was a very focused and intense effort. The turnaround times and deadlines have been extra tight.

Finally, this resource is absolutely one of a kind. Nothing like that exists in the market and making that happen took some extra resources on our end.

Even given the price increase, I am sure you will find it more than worth the investment. 

Note: If you become a Platform Member member, you get an additional 10% off!

🎓 About Science Says

• Science Says turns the latest science in marketing into 3min practical recommendations. You can join thousands of research-based marketers getting new insights every week here - at no cost.
• And I am Thomas McKinlay, the founder of Science Says. Before starting Science Says in 2020, I was a product marketer at Google. I found it crazy how little we use the latest discoveries from the brightest scientists in the world, so I decided to do something about it.

Playbook bundle offers: Looking to buy one of the bundles of Playbooks? You can find info about each Playbook on their individual pages: Pricing & Promotions ⏺Ecommerce Optimization ⏺ Ad Creatives ⏺ SaaS Optimization