When using AI bots turns customers away

New to Science Says? This is a 3-min practical summary of a scientific study 🎓 Join 31,651 marketers who make science-based decisions 📈 Subscribe here.

🤖 Working with AI? I’d love to have a quick chat

Are you working on AI implementation or AI products at a large or mid-size organization? (100+ employees)

We’re working on a follow-up to our Harvard Business Review article, around the topic of making the most effective use of AI.

There is plenty we could cover, but I want to focus on what would be most useful to you.

I’d love to have a quick chat so I can ask you a few questions to understand what are the biggest pain points you are facing now in AI (could be anything from AI agents to improving retention or trust for your AI product).

Is this you? 👉 Reply to this email so we can set up a chat! 😊

P.S.: Of course, the ‘picking your brain’ goes both ways. During our chat I’ll be more than happy to give you recommendations and ideas based on all the latest evidence in AI that we have been analyzing.

Note: want to promote your product in this space? Here’s all you need to know.

📝 Intro

You want to improve the speed and effectiveness of check-ins at your hotel. AI-powered automated check-in desks are all the craze in China, so you are strongly considering implementing them too.

Still, you have a nagging feeling. Your customers are used to real people, friendly voices, and warmth. Now you’d be asking them to talk to a machine. It might speed things up and cut costs, but it also feels… risky.

What if it makes the experience colder? What if it drives people away? Science Says your instincts might be right.

P.S. To understand the psychology behind AI interactions, check out the Blueprint we created with Wharton. Download for free here.

When warmth matters, using AI bots turns customers away

Topics: AI | Customer Experience
For: B2C. Can be tested for B2B
Research date: March 2025
Universities:  Southwestern University of Finance and Economics, Nankai University, Macau University of Science and Technology

📈 Recommendation

Avoid fully replacing humans with AI if warmth is important to the service you’re providing (e.g. scheduling appointments at a spa, ordering room service in a hotel). 

If you do, people will be more likely to stop using your services.

You can use machines and AI to provide services if customers mainly look for efficiency and practicality (e.g. grocery check-out).

🎓 Findings

• People are more likely to stop using a service provided by AI-powered bots (vs humans) when warmth is important (e.g. using a receptionist to check in at a hotel vs. a self-service kiosk).

• Scientists ran 5 experiments and an analysis of online reviews, and found that when service was provided by AI (vs humans), people: 

  • Reported 23.7% lower intention to keep using the service on average

  • Said they were 29.8% less likely to keep using the AI service (vs human service) when they were looking for fun and pleasure, but only 17.5% less likely when they were looking for functional and effective service (e.g. a quick check-in during a business trip)

  • Saw AI service as 19.5% less warm than human staff when compared directly

• The negative effect:

  • Is stronger when people are looking for fun and pleasure

  • Is weaker when people are looking for 

  • Disappears when the AI uses informal language and punctuation (e.g. “thanks!” vs “thank you”)

  • Is the same across different cultures and geographies (China and US)

🧠 Why it works

• Our impression of a service (e.g. business services, hospitality, travel, healthcare) is largely based on our interaction with the service provider. 

• We think humans feel empathy and have genuine intentions to help. We don’t think AI or machines do.

• This makes us perceive their service as less warm and friendly. 

📊 The Data Behind Rhode’s $1B Influencer Strategy

Most celebrity brands fade fast. Rhode didn’t. 

It hit $200M+/year with just 3 products and a Gen Z–powered influencer engine.

Click Analytic broke it all down in this data-packed playbook:

✅ Audience-first collabs that drive sales
✅ Viral waitlists with 400K+ signups
✅ A repeatable UGC engine that fuels TikTok and IG

📘 Grab the playbook + get 1 month of Click Analytic free to launch your own Rhode-style campaign.

This announcement was sponsored. Want your brand here? Click here.

✋ Limitations

• The research only looked at hotels. The effect might differ in other industries (e.g. travel, healthcare) and other hospitality contexts like restaurants.

• The study compared services fully provided by AI or humans. Further research found that providing services with both AI and humans can improve customers' attitudes and experiences.

• The research only looked at people’s intention to continue the service at that moment, it’s unclear whether people would be demotivated to use the service in the long term too.

🏢 Companies using this

• Fully AI provided services seem to be unpopular among customers

  • Hotel chain Henn na hotel was the first to be fully AI robot-operated, but low customer satisfaction led the chain to stop the AI service, and employ human staff

  • Mcdonald’s stopped its drive thru AI chatbot trials due to negative customer sentiment

• AI-human collaboration does seem to be more successful

  • Starbucks uses AI chatbots integrated into the app to provide ordering services and drink recommendations 

  • AI is employed frequently for security services where efficiency is prized (e.g. home safety, airport security)

  • Retailer Zara employs AI in self-checkouts and smart mirrors to provide recommendations of product pairings

Pepper Parlour in Tokyo welcomes restaurant guests with human-like robots. While right now AI-powered robots are a novelty and can be entertaining, this research finds that they are unlikely to be what people are looking for when they go to a restaurant.

⚡ Steps to implement

• Consider the context in which you use AI or machines to provide your services, and the customer need that is being addressed. 

  • If your service requires efficiency and speed (e.g. booking transport for business hotel transfer), go ahead and use AI

  • If warmth, empathy and fun are important (e.g. a restaurant for families) stick to fully human service, or employ AI as a way to support the human staff (e.g. use robots to clean the tables)

• Opt for a blend of human and AI provided services. Employ AI to provide services where interactions are at a minimum (e.g. cleaning rooms), to enable staff to focus on higher value tasks (e.g. assist and recommend customers).

• Program AI to use informal language and a casual, friendly tone. Research shows that using interjections (e.g. wow) can also improve satisfaction.

🔍 Study type

Online experiment, field experiment (survey of 94 customers outside an AI-service hotel and a comparable human-service hotel in Hangzhou, China), and market observation (analysis of 71,934 hotel reviews in China)

📖 Research

AI service may backfire: Reduced service warmth due to service provider transformation. Journal of Retailing and Consumer Services (March 2025)

🏫 Researchers

• Xingyang Lv, Southwestern University of Finance and Economics

• Yufan Yang, Nankai University & Southwestern University of Finance and Economics

• Dazhi Qin, Macau University of Science and Technology

• Xiaoyan Liu, Southwestern University of Finance and Economics

Remember: This is a new scientific discovery. In the future it will probably be better understood and could even be proven wrong (that’s how science works). It may also not be generalizable to your situation. If it’s a risky change, always test it on a small scale before rolling it out widely.

Here is how else Science Says can help your marketing:

• 📈 Join the Science Says Platform to unlock all 250+ insights, real-world case studies, and exclusive playbooks

• 📘 Boost your sales and profits with topic-specific Science-based Playbooks (e.g. Pricing, Ecommerce, SaaS, AI Best Practices)

• 🔬 Get on-demand evidence to make better decisions. My team of PhDs and I regularly help leading brands in FMCG (e.g. Mars), retail, and tech. Reach out here.

🎓 It took 3 of us 12.5 hours to accurately turn this 10-page research paper into this 3min insight.

If you enjoyed it please share it with a friend, or share it on LinkedIn and tag me (Thomas McKinlay), I’d love to engage and amplify! 

If this was forwarded by a friend you can subscribe below for $0 👇