Version 1.0, 2026-05-06

Webtrends Optimize is a rapidly evolving platform, and as such we often give users early access to our features. Further, we provide AI-driven experiences, some of which are heavily content-led.

We encourage users to be aware of what they are doing and the potential risks associated with these, and endeavour to make transparent as much detail as possible around these, which is the purpose that this document serves.

Alpha features are early-stage ideas we have. Past the stage of R&D and ready for early feedback - we may wish to get these features out to customers to help shape the direction of the feature.

Often these won't come with a credible UI, either delivered as a crude feature such as our Labs interface or operating out of other means like a Slackbot.

Availability - Alpha features are unlikely to be accessible 24x7. The operating hours may be very slim and unreliable.

Reliability - Alpha features come with no warranty to work consistently. They are under active development, and may break as part of this.

Stability - The features may evolve multiple times a day based on feedback collected and this will happen without warning. Everything from the layout to features can change or disappear without warning.

Security - We aim to keep your use safe and reasonable, but consider these to be early-stage prototypes and so not yet designed. They are typically intended to only use publicly-available data, nothing sensitive that requires additional guardrails.

We make a lot of features and move pretty quickly between them. It makes managing a public list very difficult.

At the time of giving you access to these features, we'll make clear what stage they're in.

Beta features have gone through early rounds of feedback, and are ready to put in the hands of users.

They still likely have bugs and missing "ideal" features, but are in a usable stage and ready for exposure to a wider audience.

We introduce features to customers at this stage still hoping for feedback and stories to help us promote the feature to customers when it goes live to General Availability.

Availability - Beta features will have better availability than Alpha features, but this still won't be at the level of our Service Level Agreements (SLAs). The operating hours may be unreliable at times.

Reliability - Beta features still come with no warranty to work consistently. They are under active development, and may break as part of this.

Stability - These features will evolve less frequently, but may still happen during working hours and without warning.

Security - While we design with security in mind, these features do not come with the level of security we commit to for General Availability features. For the same reason, we often limit the scope of these features to non-destructive actions and with limited/no access to sensitive data.

We make a lot of features and move pretty quickly between them. It makes managing a public list very difficult.

At the time of giving you access to these features, we'll make clear what stage they're in.

AI can be applied in many different contexts, and you will find instances of its use across Ideation, Build, Analysis, etc.

While we adopt layers of checks and reasoning into the work we do, AIs are not 100% accurate and do hallucinate replies.

The scope of this may be slim in cases like reviewing your regex, where the potential responses are small and problems are easily reverted.

The scope of this may be very wide in cases like Synthetic User Testing, where the entire interface is AI-generated content and the checks we can apply are very limited.

While we endeavour for this to not be the case, these features have the potential to return content that is wrong, misguiding, offensive, etc. For example, suggesting test ideas for your website that would reduce conversion rate, or grouping your surveys by themes you find offensive or challenging to interpret.

We recommend caution and a healthy sense of skepticism when using such features.

Wherever possible, we are designing for inline feedback with our AI features. Whilst using them, you'll typically be able to feed back (for good or bad responses) in the moment.

This helps us focus on things we're doing right and help shield everyone from wrong or dangerous replies. At the same time, where we choose to experiment with the AI strategy for our replies, it will help guide the best approach for users. This is in everyone's benefit to contribute to, and so we encourage feedback wherever possible.

Where there are no options for doing so, feel free to send feedback through your normal support channels such as JIRA Service Desk, Email, Slack/Teams, Weekly calls and QBRs, etc.