--- title: "What will you build with infinite engineers?" description: "Turn customer behavior into your next product decision with Amplitude Wave. Investigate opportunities, ship improvements, and measure results." canonical: "https://amplitude.com/en-us/blog/wave-launch" --- # What will you build with infinite engineers? Turn customer behavior into your next product decision with Amplitude Wave. Investigate opportunities, ship improvements, and measure results. Source: https://amplitude.com/en-us/blog/wave-launch --- When writing code is easy, how do you decide what to build next? What do you do when you can ship almost anything on the roadmap, but your to-do list is getting short, and you’re not sure which new idea is worth your customers’ time? At the [Lenny and Friends Summit](https://www.lennysnewsletter.com/p/all-of-the-lenny-and-friends-summit), [Claire Vo](https://x.com/clairevo) described having the best coding agents and shipping more than ever, then admitted: “I am out of good ideas, y'all.” Agentic execution had outrun her conviction about what was worth building. Today, we’re bringing [Amplitude Wave](https://amplitude.com/self-improving-products) to more customers. Wave is a proactive product agent that investigates customer behavior and product code. From that analysis, it surfaces Opportunities tied to business goals, coordinates implementation plans, and measures the impact on users after changes ship. Your team sets the direction and decides what to pursue. Wave analyzes and executes. For a long time, most product leaders had more good ideas than their team could build. The roadmap forced choices and prioritization. The team had to debate effort, draw lines, and defer work that couldn’t fit. Now, coding agents remove the old capacity limits. In the time it takes to debate whether an idea is worth building, a team can get to a working implementation. Claire’s [critique of the roadmap](https://www.thepodcastsummary.com/episodes/VM5kuvWgwDY/claire-vo-roadmap-zero-ai-obsolete.html) followed from that logic. The plan used to be a necessary forcing function that allocated scarce engineering time. But as those resources become more abundant, the roadmap becomes less useful. She argued for stronger convictions, explicit evidence, and experiments with real customers to find out whether those convictions hold. ## **Turn your product loops into a software factory** The work underpinning product decisions follows a familiar pattern. You open a funnel to find where people drop out, then watch session replays to understand what they were trying to do. You search support conversations to see whether anyone has described the problem in their own words. A recent release could explain a broken user pattern. An experiment might already be testing it. After you finish researching and have an explanation you can defend, you still need to scope the fix and define what success would look like. Doing the job well takes time. While you are doing all that work, your team keeps shipping. New releases are adding new sets of questions to investigate. It never slows down. At the next product review, you are still responsible for explaining whether any of the changes actually made your product better. We want teams to be able to pursue ambitious product bets without needing to pause any of the other work streams. To focus on big decisions while a product agent investigates the experience that’s live for customers. In our ideal world, problems that would otherwise persist until someone noticed them should get flagged automatically. Better yet, they should come with an investigation and a proposed change attached. Your team can focus your rare attention resources on deciding whether the evidence is convincing and the solution is worth building. Every product team builds, ships, observes use, and learns from the result. AI has sped up the first two steps. Wave connects the rest of the loop by bringing agentic speed to investigation and measurement. Our vision for self-improving products is a software factory where customer evidence guides what gets built and each release informs the next change. Your team sets direction and decides what reaches customers. > I was genuinely shocked by how good the opportunities Wave identified were, even on its first ever run. Some of the opportunities and issues it identified were things we would almost certainly never have spotted without it. Daragh Kelly Chief Data Officer, The Economist In our [earlier look inside Wave](https://amplitude.com/blog/wave), we shared how we built this loop at Amplitude. Now, we’re bringing that approach to more of our customers, letting them set their own business goals that guide their product work from investigation through validation. ## **Start with the outcome your business needs** The first step to work with Wave is to focus a Product Area on a specific business outcome. For example, you could work on improving retention or helping more users finish signing up. Wave investigates behavioral data, session replays, and customer feedback to find parts of the experience that could change to achieve that goal. It turns the evidence into a documented Opportunity, along with a recommended change and a hypothesis your team can test. The longer-term ambition is to make this an ongoing assignment. As customers use the product and create new evidence, agents will continue looking for new ways to advance the goal. Your team can steer the work without having to manually kick off every investigation. > Wave is good at surfacing problems and showing their surface area. Each one comes with a potential solution that's easy to share with design and product teams. Once it's set up and defined well, it finds things worth looking at. Ryan Mach Staff Analyst, PetScreening ## **Give new users a reason to come back** Mewtant, the team behind the AI image creation platform [PixAI](https://amplitude.com/blog/pixai-wave-retention), wanted to improve retention among new users. Wave found a product action (publishing artwork) that correlated with retention, but usage data showed that many new users were leaving the product without ever discovering the publish button. Based on that information, Wave’s initial recommendation was to prompt users to publish an image after generating it. The PixAI team adapted the idea to point users to its Missions tab, where they can complete actions to claim free image generation credits. When a user runs out of credits, a tooltip now points to Missions, including rewards for publishing artwork. After making that change, PixAI reported a 76% increase in Missions opens, nearly 10% higher publishing, and a 3% increase in Day 1 retention. The change helped new users discover value in features PixAI had already built. Wave supplied a hypothesis from customer behavior, and the team used its product knowledge to choose how to act on it. > Wave connects session replay, web performance, and user behavior into a single view, uncovering opportunities we didn’t know existed and giving us the evidence to act on the ones we suspected were there. Brian McGinnis Director Platform Engineering, Siemens ## **Follow the friction past the first fix** We run Wave internally at Amplitude, too. In one replay, a user spent 76 seconds writing a question on our Agent homepage even though the same page showed three suggestion cards. We had offered three possible starting points, and none matched what that person needed to do. Wave’s analysis of 100 replays revealed a broader pattern. Users manually typed questions in 80 sessions and selected a card in only 12. The cards were rendering correctly. Their generic prompts were just missing the work people came to do. Wave recommended generating suggestions from each user’s recent chart and dashboard activity. Our team updated the experience to show personalized cards to existing users with in-product context. We kept the generic suggestions for people without it. Everyone could still type their own question. The initial experiment reported roughly 160% more suggestion sends with that treatment. As we continue testing and validating, our team can return to the same Opportunity to examine all the evidence and continue forming new hypotheses about what our users want. > What excites me most is how connected the Amplitude experience is becoming. We can identify opportunities through Wave, act on them with confidence, then use Guides and Surveys to create awareness. They’ve helped us uncover hidden bugs, friction in critical workflows, and spot improvements. Susana Fuentes Business Analyst, OneClickCode ## **Carry what you learn into the next release** Agents need customer context to form useful hypotheses and consistent metric definitions to evaluate them. They need to be able to turn that information into action: a path into implementation, a way to test changes, and boundaries on what can reach customers. Amplitude brings behavioral evidence and outcome measurement into that process. A failed hypothesis should change the next recommendation. A rejected idea can capture reasoning that wasn’t found in the data alone. Wave retains outcomes and team feedback, so future work can repurpose what the organization has already learned. > I have been impressed with Wave’s early capabilities. The way it structures evidence-based opportunities is easily my favorite feature. It continues to improve with feedback, and I’m excited for our teams to engage with it further in 2027. AJ Long Director Product Analytics, PrizePicks ## **Make room for your team’s next big bet** Delegating more investigation and implementation tasks gives product teams room for work that needs their judgment. Product managers set a direction and choose the consequential bets. Designers bring research and craft, while engineers shape architecture and release boundaries. Agents can carry more of the investigation and follow through between those decisions. > Bringing Wave together with Experiment and Amplitude Analytics has us well on our way to turning the mortgage app into a closed-loop software factory. It has the effect of adding an experienced growth PM, analyst, and engineering team—all wrapped into a single application. Cody Chin Director Mobile Product Management, Rocket ## **Close the loop on what you build** [DigitalOcean](https://amplitude.com/blog/digitalocean%2Dwave) is putting the product loop to work across five Product Areas, from developer onboarding to newer AI experiences. In each one, Wave surfaces ranked Opportunities with a hypothesis, wireframe, and experiment design. Product owners can decide within minutes which Opportunities to pursue, then create Jira tickets directly from Wave. Engineers build and launch the tests, and the team reviews results in Amplitude without additional tooling or analyst time. The results guide which experiences reach customers. One experiment made the payment requirement more visible on the Droplet creation page. More teams added a payment method, and DigitalOcean decided to roll out the banner. > “In only six weeks, Wave surfaced 55 opportunities across those product areas… My time spent writing separate product requirements documents (PRDs) went to almost zero.” Haley Eidem, PhD, Product Manager, DigitalOcean ## **Give your next business goal to Wave** Starting today, Amplitude Wave is available as a paid offering for customers on our Growth and Enterprise plans. To get started, our team will help you assess your product analytics and session replay data. Just bring a business goal and talk to our team about how Wave can help you investigate Opportunities, ship changes, and keep learning from the results. Try Wave on your product Take the next step toward self-improving products [Request Wave Access](https://amplitude.com/wave?siteLocation=blog-end-of-article-cta) ## Explore related content - [How DigitalOcean Improves Developer Onboarding With WaveCustomers](/blog/digitalocean-wave) - [How Wave makes me a better growth PMInsights](/blog/wave-product-friction) - [How I use Wave to find Amplitude's next product improvementInsights](/blog/wave-product-improvements) - [How PixAI lifted publishing nearly 10% with Amplitude WaveCustomers](/blog/pixai-wave-retention)