What messaging techniques are effective to build greater trust in AI/ML features of products and solutions?

AI/ML means a lot of things to a lot of businesses. Some companies have high data maturity where AI/ML tools are a vehicle for data science teams to create proprietary value for the business. Others are fairly low on their data maturity (which is fine!) and they want out-of-the-box AI/ML for fairly simple use cases, usually around personalization.

The best messaging technique here is clarity. Do the work for the prospect. Is this product for advanced or basic users? Technical users? Does it require a data lake? By being really clear about where the product and solution fit, you give the right people the opportunity to raise their hands.

While I often bristle when people call out that technical audiences don't like being marketed to (the implication being other personas like fluff marketing?!), I absolutely believe that core differences in how personas evaluate solutions matter. To build trust in the value of AI/ML features, here are a few ideas that come to mind:

1. Given the amount of hype in the AI/ML space, customer proof becomes even more powerful. Showcasing concretely how users have used your solution, their goals, and the impact they achieved is among the most impactful way to build confidence. 
2. When targeting data scientists/developers, in addition to anchoring to the value and impact a given capability will have, being explicit about how the capability works technically and what the experience is like to use it are just as important. This doesn't need to mean your content is overly long or dense; you can rely on visuals and workflows heavily. 
3. When crafting messaging for less technical business stakeholders or economic buyers, we anchor to the big-picture business impact they'll achieve, things like accelerating time to value. To build trust in that promise, we speak directly to the pain the currently experience and how we resolve it. We also empower our technical champions to connect the dots betwen the technology and the value as their endorsement is the most powerful evidence!

To build greater trust in AI/ML features of a product, I would focus on a couple of different aspects. The first is the quality and quantity of data being used to train your AI/ML models. People like to know that the data is ethically sourced and accurate. Usually quantity of data also helps to validate how well your AI/ML models work and how well it scales. Second, I would focus on expertise - this could be in terms of years developing and training the AI/ML or in terms of data scientists who are working on your AI/ML features. Messaging on expertise gives users more confidence in the data output. Finally, any customer case studies or testimonials that support the feature also reinforces trust in the solution working.

One of the biggest mistakes I see marketers make when talking about Artificial Intelligence (AI) or Machine Learning (ML) is they just throw the terms around. Like their employer can just sprinkle some AI salt on something to give it extra marketing zest.

Think like a journalist for a small-town news site, instead, and try to explain why AI/ML is important and how it works for the use case at hand. Interviewing a software developer working on the AI/ML aspect of your product is one of my favorite shortcuts. They're used to explaining complex stuff to all kinds of people!

Less helpful example: Our customer data platform uses machine learning to target customers who are "likely to convert."

Perhaps a more helpful example: Say you're a marketer for a health and wellness website and you want to send a "paid subscription" email promotion to a set of free subscribers. You could use our customer data platform to find a small group of free subscribers that the software thinks are likely to engage with your promotion and renew their subscriptions. The way we do that is by using an approach called "machine learning" that combs through thousands of interactions of readers who have already subscribed and looks for common (they've visited your website's subscription-renewal page) behaviors as well as seemingly odd ones (they use a Chrome browser). We then find the free subscribers who have done similar things (visited the renewal page plus use a Chrome browser) — yet not subscribed — and allow you to target them. Our customers say that using this approach to target "likely to convert" customers doubles the effectiveness of their paid-subscription promotions.

If you look at my writing for Calendly, I'm always using either real customer examples or making up hypothetical ones to explain stuff. I recommend trying it.

With any emerging technology, I've found the best approach is to compare to the previous way of doing things. People digest a new way of doing something by comparing it to how they do it today. Think about a lot of the technologies we use today: Email = electronic mail, Software as a Service = Software that you don't run locally on your comuter, TV (tele-vision) = vision at a distance, etc.

There's some real fear and trust issues with AI/ML. But there's also a lot of just fear of the unknown. Think about what AI/ML is doing for your customers and how they do that today. Make that comparison. Use metaphors/analogies to work harder for you. Make sure people know they have options to do as much or as little as they want with AI/ML.

It’s true that customers are both skeptical of AI outputs and excited about AI’s potential. At Zendesk, we’ve definitely seen this in the customer service space, since this is one of the most obvious areas that AI is poised to disrupt. Here are some tactics we’ve found useful:

• Show value: Focus on demonstrating tangible value through real-world examples and clear use cases that customers can latch on to. Showcase specific customer success stories that highlight ROI and impact, making the benefits of AI adoption more relatable.

• Streamline adoption: Simplify the adoption process by providing curated prompts, best practices, and guided implementation paths. This approach demystifies AI and helps customers quickly leverage it in their workflows.

• Be transparent: Get specific about your AI's capabilities and limitations. Use plain language to explain how the technology works and how its outputs are designed to be clearly understandable by customers.

• Mitigate fears: Position AI as a tool that augments human expertise rather than replacing it. Illustrate how human-AI collaboration leads to better outcomes, addressing fears about job loss.

• Highlight responsible AI: Proactively address data privacy and ethical considerations. Clearly communicate your commitment to responsible AI development and usage, showing the measures in place to protect user data and ensure fair, unbiased results.

• Share your unique perspective: Articulate a differentiated perspective on AI's role in your industry, which can come through in your marketing narrative and sales assets. Emphasize your commitment to guiding customers toward success, positioning your company as a trusted partner in their AI journey.

Building trust in AI products starts with the product and company itself—is there a dedicated AI safety team? Are the right AI safety checkpoints in place to safeguard from abuse? Is the company taking a principled approach to rollout?

If the foundations of AI safety are there, then the messaging becomes much easier. I approach messaging with a few things in mind:

• Being transparent about the intended use of the technology and the risks, including how the company is mitigating those risks through safety precautions

• Knowing your audience and using the right language that will resonate with them (which generally means: no unnecessary jargon!)

• Illustrating the use cases of AI to make the technology feel more tangible and less mysterious