We've raised $2M to build self-improving software

Software ate the world, but it sucks

In 2011, Marc Andreessen wrote the now-famous “Software is Eating the World” post, and oh boy was he right. Since he wrote that post, global software expenditure has risen 760% to $1.24 trillion, with the software-heavy Magnificent 7 now making up more than 30% of the S&P500.

AI has unleashed a new tsunami of personalization and problem solving, but most applications we use today are still rigid, static, and plainly… sucky.

And it’s not your tech team’s fault

At Uber, I saw how hard teams of Data Scientists, Product Managers, Designers and Engineers work to create new experiences.

Even experimenting on a piece of copy in Uber’s app could easily take a month, if not more. The typical workflow looks something like this…

• Brainstorm: Identify of a bunch of ideas to improve the Uber experience

• Stackrank: Dive into the data (or user research) to stackrank which ones to actually work on

• Define: PM writes a spec of the top option

• Design: Designer explores potential versions, working with the PM to align on the right experience

• Delay: Wait for the engineers to free up

• Develop: Implement the new experience

• Test: Walk through test steps to confirm no degredations

• Configure: Set up and launch the new experiment

• Delay: Wait weeks for usage data

• Forget: Probably forget it’s running. Come back months later

• Assess: Look into the results, identify the winner

• Finalize: Go back to engineers to ship the winner

• …then do it over again

And that's probably optimistic.

We’re getting agents for coding, but what about everything else?

First Devin, then Cascade, Bolt, Lovable, Replit, and a whole sea of other coding agents have arrived recently.

But what about understanding how your users use your products? Identifying new opportunities? Assessing them? Designing them? Launching them, and learning about them?

But what about the rest of the software lifecycle? Who's automating that?

Helium is building self improving software

Imagine the best growth team you’ve ever had. In just a few years, everything they did will be automated. Helium is bringing that future forward.

Helium's agents will work day and night to improve software applications, continuously identifying new opportunities, then building out new features to improve user experiences or lift business metrics.

We're starting by growing revenue for mobile apps

We looked for an opportunity where companies are already spending a substantial amount of time testing their apps to improve their business, and mobile apps quickly stood out.

Mobile was arguably the birthplace of product experimentation and growth, led by Zynga and mobile gaming. But mobile experimentation has been embarrassingly stagnant since then.

Apps still spend a huge portion of their time manually testing dozens of new onboarding flows, price points, and paywall experiences.

Starting a few months back, Helium launched the first automated mobile experiments.

We’ve built a series of models trained on hundreds of experiments that automatically generates and tests high-performing paywalls, built just for your app.

Here’s how it works…

1. Helium’s AI assess your existing paywall experience and conversion to identify a new opportunity

2. Then, it’ll generate an entirely new paywall experience for you.

3. You’ll get the proposal in your email

4. Use the Helium editor to make any small tweaks before approving to automatically start a new experiment

5. Helium will circle back with the results, then line up a new test for you

By automating paywall testing, Helium raises revenue (often >50%), while freeing up our customer’s tech teams to focus on building more powerful products for their users. We're also layering on new personalization features that automatically identify the best experience to serve each user to maximize revenue.

Driving Subscription Revenue >75% for Locket

We’ve been piloting with a handful of initial apps, including Locket, an app to share photos from your day with your closest friends.

We onboarded Locket late last year, and have been running experiments with Helium ever since — in that time we’ve been able to help grow Locket’s subscription revenue by more than 75%.

Read the full case study here.

But we aren’t stopping there

Over the next year, we're continuing to build out Helium to automatically test and improve your mobile apps, aiming to tackle tasks like…

• Re-vamping your onboarding flow, and personalizing it for each user to build the most excitement and understanding of your product.

• Identifying dead ends where users get confused or drop off, and automatically fixing them.

• Launch out-of-app messages to keep users engaged, increase retention, and boost conversion.

• Build out new growth features like referral programs, streaks, and more.

• Test new prices and discounts to maximize LTV for every user.

If you want to join some of the fastest-growing apps in the world to drive in-app revenue, join our waitlist. We're excited to work with you on our journey to build self-improving software.

Stop moving buttons around. Start making money in your sleep.