Chameleon is pivoting to a more complete product improvement system

We're pivoting from in-app guidance to a self-improving product system, with AI agents that track user behavior, surface product friction, and personalize UX, to do the full job of improving your product automatically.

For eleven years, Chameleon has delivered the best in-app experiences (tours, walkthroughs, checklists etc.) to customers that prioritized great user experience.

Now the market has changed, and we're empowered to do more.  

Accordingly, Chameleon is pivoting from a Digital Adoption Solution to a Self-Improving Product System. 

This means we're adding significant new capabilities to our existing in-app functionality, and committing to being your system that automatically learns what's working and not in your product, and helps you fix, communicate, monitor, and learn.

Previously, we had relied on partners to complement in-app experiences for user behavior tracking, whether via a CDP or traditional product analytics. However, we believe the traditional event-based analytics architecture is at risk and will be replaced by an AI-native, inference-based model. 

No one needs to create charts and run manual queries. Instead, we can set up a monitoring and alerting system that proactively identifies issues and helps you take action. Accordingly, we saw an opportunity to rebuild the product analytics category at a much lower price and with a much stronger signal. 

Chameleon can now track all in-product user behavior, generate insights, and identify issues.

You can ask Chameleon about a specific user’s use cases, review recent sessions related to a Feature, compile the overall friction points in your product, understand the depth of adoption of a feature by a particular company, and more. 

For those who choose to enable Compass to track user behavior, Chameleon will identify “friction candidates”. These are points in the journey and product where users struggle, and are worth solving to improve your product. 

You can review a report of these, download them, or send them directly to our integrations, including Linear. You can turn this on autopilot, meaning you can have a feed of issues to triage in Linear generated from real usage. 

And of course if you use a coding agent (Linear Agent, Claude Code, Codex etc.) you can have that triage and kick-off a build directly.

We’ll monitor any issues created and their status, so once the fixes or improvements are deployed, we’ll assess whether they truly solved the underlying user friction or if more work is required. 

As we track, summarize, and infer from each user session, we also build a profile per user (and per company, from all their users). This becomes a context graph for how each user uses your product, which features they adopt, what their regular workflows are, where they struggle etc. We supplement this with any attributes you pass to us to generate a rich understanding of each user, and their needs. 

You can then leverage this within Chameleon to create personalized in-app Experiences: 

• Better follow-up questions in Microsurveys that respect what a user knows and go deeper in extracting the right feedback

• Copy within in-app guidance that speaks to a user, their industry, their level of expertise, their usage etc. 

• Checklists that show items the user will much more likely to care about, at the right place and time

This truly enables 1-1 user experience for your product, bringing us back to our mission of personalized software. 

A major reason the old Chameleon was hard to drive successfully was the requirement that a human lead it. If no one on your team had the job of managing in-apps, customers would lapse in their usage, experiences would not be updated or optimized, and, over time, usage of our tool would decay. 

We’re making big strides towards Chameleon taking on more responsibility for success and making our customers orchestrators, managers, strategizers, and overseers, rather than builders and doers. 

Chameleon’s Copilot can quickly plan and build any in-app Experience, just with a simple text prompt. It knows your account data, so it can choose the correct URL rules or apply the right Theme. It can analyze an Experience, create a variant to test, or optimize the copy for your brand’s tone of voice (which is known and stored in the Context settings). It can summarize Microsurvey feedback to themes, generate Linear issues, and create follow-up campaigns. It’s also all available as an MCP and Slack app. 

And now, we have Scheduled Jobs in Chameleon's Copilot, which can take actions on a regular and recurring basis, for example, checking your release log/notes or what’s new page and generating relevant in-app announcements, or running a quarterly NPS report to be able to share with stakeholders (or take to Claude Design to create a presentation). 

Our Ranger Agent keeps your account in good standing: monitoring any elements in your app that might have changed and need to be re-attached; identifying teammates that haven’t accessed the product for a while and can be removed to free up seats; alerting when a theme doesn’t adhere to accessibility standards and a lot more. We continue to develop more “jobs” for Ranger so that it becomes a simple inbox feed of things that need your attention, and soon you’ll be able to put adjustments on autopilot, like turning off Experiences with low engagement, or renaming Segments to match your taxonomy. 

As we develop more AI systems, we will be moving deeper into usage-based pricing. Already, Chameleon prices are based on the number of end-users (MTUs) that are in your platform, engaging with in-apps. We will layer on top credit-based billing for systems and features that consume tokens, and offer granular control over what you want to spend credits on. We believe this will give you the most control over where you invest your resources, and we hope it will deliver outsized ROI for your product and company. 

Parts of the system are live today, and the full picture will land over the coming weeks. If you want to see what a self-improving product system can do in your own product, book a demo. I'd like to hear what you make of it.

We’ve built a lot in the last 6 months, and a big part of that foundation is our new internal coding agent, a harness that knows our systems and processes and is automating planning, building, testing, deployment, and more. It’s enabled our whole team (including support engineers, CSMs, marketers, and others outside the core EPD group) to ship product and systems at 10x speed, and that are self-learning and improving. 

We expect to continue at this pace and believe that a nimble, high-performing team at the cutting edge of AI is the most powerful force to create impact through software. 

We are also exploring new products and ideas to attack the wider problem space we are so familiar with, and we’re excited to share more developments with you in the coming months!