Build Better Products: AI Edition

Envisioning the AI-powered future of product development with Amplitude

Inside Amplitude
August 8, 2023
Image of Jeffrey Wang
Jeffrey Wang
Co-founder & Chief Architect
A geometric 3D shapes in Amplitude colors on a blue-background with white text indicating Amplitude AI

When we started Amplitude a decade ago, we had a strong conviction: product and user behavior data is the key to building better digital products and experiences. By deeply understanding and drawing insights from what your users are actually doing in your product, you learn so much about what to build and improve on. From personalized homepage experiences at NBC to correlating spikes in traffic with new products at Walmart to uncovering the value of meditation reminders at Calm, we have seen so many companies find success by integrating data into their decision-making. Today, our conviction is stronger than ever.

In parallel, the field of artificial intelligence (AI) has witnessed incredible advancements. The past year has been particularly exciting with the emergence of ChatGPT, Stable Diffusion, and other generative models that demonstrate human-like intelligence. At Amplitude, we have long been at the forefront of helping product teams leverage AI and machine learning (ML) techniques with features like our recommendation engine and product monitoring. Deep down, we’ve always known that the day would come when AI technology would transform the way everyone builds products.

That future is now on the horizon, and we’re eagerly anticipating the arrival of a new era of AI-powered digital products and experiences. Given our front-row seat into how product teams of the future are building, we feel compelled to share our vision of where this all leads. Spoiler alert—it’s going to get wild!

The product improvement loop

All good product and growth teams understand the iterative process of product development: build, ship, use, and learn. The speed and effectiveness with which you navigate this loop determines your success as a product team. It's safe to say that AI will impact each step of this process in a significant way.

Diagram of AI at the center of Product Development with Build, Ship, Use, and Learn as the stages of product development

Generative AI has already accelerated the “build” and "ship" phases and will continue to do so. Tools like GitHub Copilot and Figma AI are making it easier than ever to create new products, enabling you to go from idea to prototype to production in record time. The best teams ship multiple times a day, and as tooling improves, they’ll only get faster.

At Amplitude, we are working hard to make the “use” and “learn” phases as effective as possible. These stages are becoming the biggest bottleneck in the product development process, but fortunately AI presents abundant opportunities to unlock the potential of products. We’re seeing the early implications today with our launch of Amplitude AI. Data Assistant ensures consistent measurement and data quality, while Ask Amplitude makes it even simpler to go from business questions to insights. We’re making it effortless to find your product’s “Aha” moment by having AI sift through your data to uncover behavioral patterns.

AI simplifies every step of the loop, enabling products to improve more quickly than ever before.

Experimentation and AI

There are also more fundamental ways that the product improvement loop will change.

Today, teams look to A/B testing to inform the “learn” phase. This helps them ensure the product is getting better with each iteration. Typically, this process requires human involvement at every stage, ideally facilitated by a tool like Amplitude Experiment. Identifying the hypothesis, designing the experiment, building the variants, deploying the changes, and assessing the results—all demand time and attention, making the process expensive. Regardless, teams forge ahead with experimentation because it is the best way to get real insights into how different features impact user behavior.

An illustration of the A/B testing process from hypothesis to design, launch, analyze and results.

Importantly, experimentation provides a rigorous framework for evaluating product changes, which is exactly what AI needs to excel at the task. Take the example of optimizing marketing copy on a website. Many companies run A/B tests to ascertain which copy drives the most engagement, signups, purchases, etc. Thanks to AI and large language models (LLMs), teams can already generate diverse variations of marketing content with ease. When you combine this with an experimentation framework, you end up with an AI-powered system that can continuously improve your website with minimal human supervision. Magic!

Toward self-improving products

The impact of AI extends far beyond marketing copy. Products of the future will be designed in ways that harness the power of AI for autonomous improvement versus manual iterative development. Think of the TikTok feed or the Netflix homepage as very early examples of products that adapt on their own.

We will inject AI into more and more of our digital products: navigations will be reordered, layouts will be shuffled around, content will be dynamically selected (or generated), and entire flows will be created and broken down—and this will all happen automatically, in real time.

Iterations of the product improvement loop that take weeks or months to develop today will happen seamlessly in the background, as a side effect of the product being in the hands of users.

Experiment ideas previously ignored due to low expected ROI will become negligible in cost, enabling a plethora of small product tweaks and UX delighters to get shipped. This frees up product builders like ourselves to focus on what products do, rather than the details of how they work.

Chart showing improved product quality due to AI-powered functionality

Together, let’s build better products

Join us on this journey. Are you an existing Amplitude customer that wants early access to Amplitude AI solutions and features? Sign up for our AI design partner program.

Be sure to check out our AI principles and learn more about how we’re implementing AI at Amplitude.

About the Author
Image of Jeffrey Wang
Jeffrey Wang
Co-founder & Chief Architect
Jeffrey owns the infrastructure that enables us to scan billions of events every second. He studied Computer Science at Stanford and brings experience building infrastructure from Palantir and Sumo Logic.