Session Replay Summaries help you quickly decide which replays to watch, by using Generative AI to give you a summarized breakdown of what happened during a replay, including:
- A quick summary of user behavior in a session replay
- Summarized key friction points encountered in the session
- Links to specific examples of friction points
About the replay summary
Each replay summary includes the following sections:
- Key insights: Bulleted summaries of user behavior found in the replay.
- Potential issues: Information about potential issues or friction encountered by the user with time-coded links allowing you to jump to specific moments of interest in the replay.
- Detailed journey: Use the arrow icon(s) to expand user journeys details detected in a single replay. Review user behavior and click the time-coded links to watch back the specific moment of interest.
Additionally, use the ‘thumbs up/down’ icons to leave us feedback about the replay summary.
How to use session replay summaries
Note: Can’t access replay summaries? It likely means the AI Service Schedule hasn’t been accepted on your Contentsquare account. See the ‘Accessing session replay summaries’ section below for more details.
You can access two types of replay summaries:
- Individual summaries
- Group summaries (requires Pro or Enterprise plans)
Individual summaries
Get a summary of a single replay to review key insights, journeys, and potential issues.
Access these by:
- Clicking the ‘sparkle’ icon on a replay from the list
- Clicking the ‘Summarize’ button when watching a replay
Group summaries (Pro / Enterprise plans)
Group summaries allow you to summarize up to 100 replays, so you can analyze behavioral patterns and highlight key friction points across the grouped data.
Access these by:
- Group 10 replays: Click the 'Summarize replays' button
- Group up to 100 replays: Click your preferred number of replays to summarize (25, 50, 100) next to 'Results per page' at the bottom of the list, followed by 'Summarize replays'.
Tip: Refine your group summaries
Summarize a different group of replays by narrowing down which are displayed in the list. Update the analysis context (device and date range), apply a filter, show your favorites, or use the column sorting to change the scope of replays being summarized.
Grouped summaries with over 25 replays
Any summary created with 26 sessions up to 100, will not include Potential Issues and is limited to 5 Key Insights only.
Leverage Sense Chat with summaries
After summarizing a single, or a group of replays, use the Sense Chat side panel to ask any follow-up questions about the summary to gain further insights for your analysis.
Note: to ask follow-up questions in Sense about a single replay summary, you will first need to trigger the summary from the replay list view (by clicking the 'sparkle' AI icon next to the replay).
Alternatively, you can use Sense Chat from another capability, for example Journey Analysis, and ask it to show you/analyze one or more related replays. This allows you to leverage Sense Chat to deepen your analyses across other capabilities alongside Session Replay.
Accessing session replay summaries
In order to access replay summaries you must accept the AI Service Schedule. This can only be done by an Admin user on your account, following the steps below:
1. Navigate to Session Replay and from the right-hand panel click 'Get started'.
2. Review the linked AI Service Schedule, check the box and click ‘Accept'.
FAQs
I don't know who the Admin users are on my account, how can I find out?
You can contact your Customer Success Manager, or ask your internal team to identify the Admin users on your Contentsquare account.
You can learn more about user management and permissions here.
How can I share the AI Service Schedule with my legal team for review?
You can share the direct link to the AI Service Schedule here: https://contentsquare.com/legal/service-schedule-cs-ai
Do Session Replay Summaries work on projects with personal data masking applied?
Yes, as the input used for session replay summaries does not rely on masked data. Learn more about how personal data masking works with Session Replay.
Is it possible to disable Session Replay Summaries from displaying on my Contentsquare project?
Yes, account Admins are able to enable/disable Session Replay Summaries from the Console, on a per project basis. Go to 'Console' > 'Modules & Features', then next to the 'AI' row click 'Access'. Then simply toggle on or off 'Enable Session Replay Summaries' for the project to control visibility of the feature in that project.
Which AI models is Contentsquare using?
Please refer to our AI Model List for information on the models used for each generative AI feature in the Contentsquare platform, as well as Contentsquare’s Sub-processors List for more information about what data is processed by Contentsquare AI feature.
What data is sent to AI models?
For session replay summaries specifically, we are currently sending the following data to the AI model as part of a prompt:
- Project-related data, such as website name, website URL.
- Field-related data: industry, subindustry, subindustry category.
- Data related to the user environment: such as the os or the device type.
- Data related to the user session itself such as the session time, its duration, the URLs visited during the session. For example:
- For each pageview, we send events that the user performed on the page including user clicks, frustration events (e.g. rage clicks). We also send technical data such as Core Web Vitals metrics and technical errors (e.g. Javascript errors, API errors).
- Aggregated data about the URLs visited during the session. This allows the comparison between the user experience during the session and the "average" experience. This includes: traffic volume (visits, pageviews), behavioral metrics (bounce rate, exit rate, etc), and core web vitals metrics.
Where can I learn more about the privacy, security and compliance?
Please visit our Trust Portal for further details.
Do Session Replay Summaries process any personal data?
While individual Summaries do not intend to process any personal data, group Summaries collect the following visitor personal data:
- the description of the issue (`text`)
- URL where the issue happened (`pageURL`)
- the user that encountered the issue (`userIdHashed`)(generated random userID)
- the session where the issue happened (`sessionId`)
- the event that caused the issue (`eventId`), the event can be a page visit, error, or a user action.
Because our cloud providers (AWS & Azure) host their own copies of those models, no customer data is ever sent to the original model developer organizations (i.e., Anthropic, OpenAI).