Predict what users will do with your software in real-time, so you can change the outcome. WalkMe AI analyzes user data to predict the future behavior for your users so can use WalkMe to prevent churn and increase adoption of features.
The Short Version
WalkMe AI (Artificial Intelligence) is designed to empower you to take full advantage of WalkMe’s unique digital adoption platform. From data collection to engaging directly with end users, WalkMe AI replaces manual analysis, allowing you to focus on using WalkMe to engage with your users in meaningful ways that improve your business outcomes. With WalkMe AI you can predict user behavior allowing you to create WalkMe content that proactively drive users to meet your business goals.
WalkMe AI works on both traditional websites, via WalkMe’s robust rule engine, and on mobile applications, via Campaigns.
- Drive awareness of new or high-value features
- Improve adoption of processes and features
- Give users extra help if they’re likely to reach support
- Increase conversion rates
- Increase sign-ups for services or events
- Avoid churn and improve retention
How It Works
To predict future behavior, WalkMe AI collects data and generates a model based on behavioral trends from previous sessions. When a user visits your site, WalkMe loads the predictive model and based on a user’s current session can predict at a high level of accuracy if the user will end up meeting the AI rule criteria. If WalkMe AI evaluates that the user will likely meet your AI criteria, it triggers any WalkMe content that is set to play based on it.
The AI solution is made up of two distinct processes:
- Data collection: WalkMe constantly collects and analyzes behavioral data from users
- Prediction Model: Statistical analysis (derived from the data collected) predicts the likelihood of actions (visit a specific page, churn, etc.) to take place in a user session
The following list summarizes the common parameters used to generate each unique predictive model:
- User Journey: page views, usage of features, session parameters
- Technical Context: Application version, device, browser
- Location and Time: Country, date, time
- User Attributes: user role, user segment, onboarding progress
Predictive behavior models generated by WalkMe AI require statistically significant data sets to make accurate predictions. The number of end users on your underlying application and the frequency with which they use it, affect the time it takes for a model to be generated. Each model is created to be specific to your application and the defined goal. Over time, the model is consistently updated to maintain a high accuracy percentage.
To start generating your site-specific prediction model, publish a ShoutOut with your desired AI prediction. Once WalkMe gathers sufficient data so that the model can predict when a user is likely to meet the set criteria, it will start to trigger relevant WalkMe engagement features based on that criteria.
How to Use AI
Setting WalkMe content to be triggered based on WalkMe AI is incredibly easy, giving you the freedom to focus on creating a WalkMe solution that impacts your business goal, rather than spending your time setting it up. Initially, we support triggering ShoutOuts with WalkMe AI, which allow you create a customized call-to-action that can play other WalkMe content.
Create a ShoutOut that targets users based on AI
- Create a ShoutOut
- Click Next. Engagement Tab will appear
- From the Auto Play, select Play according to a rule
- Click Create an Auto Play Rule. The Rule Engine will appear
- From Type menu, select AI
- Choose your AI prediction
- If selecting likely/unlikely to visit page, add the desired URL
- Click Done, save your work and publish it to production
- The ShoutOut will play once WalkMe AI can accurately predict the desired behavior.
Likely to Visit a Page
WalkMe AI predicts if users are likely to visit specific pages in your website. Since pages (URLs) are usually related to features or functional parts of your application, you can easily correlate a visit to a page with a user’s intention or objective.
Likely to Need In-Application Support
If a user is likely to visit the support section of a website, it’s an indicator that they probably need help using your software. With WalkMe AI, you can identify these users in advance and show them WalkMe content offering on-screen, in-context assistance. This allows you to prevent users from having to search through a support site or create a ticket.
Likely to Miss a Relevant or High-Value Feature
Users that are unaware of specific features may not realize the full value of your software or may be confused in the middle of a process. If a user is not likely to visit the page of a specific feature, you can target this user with engaging WalkMe content, promoting features and driving the user towards using them.
Likely to Churn
WalkMe defines “churn” as the point at which a customer will not return to a website or application. Meaning, WalkMe predicts if this is a user’s last session with your application. This is a very important metric for customers that want to improve retention.
Once a user leaves your software there are limited options to re-engage them with your product. Many users at this point have decided that your software is not a good fit for their needs.
WalkMe AI can predicts which users may churn allowing you to create WalkMe content that re-engages them with your product improve retention. For example, you can create WalkMe content that promotes specific features, offers attractive discounts, drives better adoption of features, explains how things work, offering help and so on.
WalkMe AI for Mobile
AI for mobile apps works similarly to the web version, except that it can identify two additional user groups:
- Users likely to perform a positive action relative to a pre-set campaign
- Users likely to complete a pre-set business goal
For More information about WalkMe AI for Mobile contact you CSM