AEO helps ensure your learning content is discoverable, understood, and surfaced by AI-powered tools such as ChatGPT, Google AI Overviews, Gemini, and Perplexity. Instead of requiring users to browse multiple links, AEO allows your content to appear directly within AI-generated answers or as a cited source.

Not set up yet? You're still welcome to join and watch it in action.

Seats are limited, so register early.

These sessions are not for troubleshooting configuration. For help getting set up, visit our Support Article.

AEO helps ensure your learning content is discoverable, understood, and surfaced by AI-powered tools such as ChatGPT, Google AI Overviews, Gemini, and Perplexity. Instead of requiring users to browse multiple links, AEO allows your content to appear directly within AI-generated answers or as a cited source.

Not set up yet? You're still welcome to join and watch it in action.

Seats are limited, so register early.

These sessions are not for troubleshooting configuration. For help getting set up, visit our Support Article.

Key things to know:

• Conversational AI indexes a range of topic types, transcripts (including video transcripts), recipes, text, PDFs, presentations, and more. Content needs to exist as an asset in the asset library (whether attached directly to a course or added via the global resource library) to be indexed and surfaced as a source.

• Turning it on is self-serve: enable it from Apps in your dashboard and accept the terms of service. Initial indexing can take up to a couple of hours the first time, with a notification once processing completes.

• Conversational AI Learning is now in general availability and is a paid, usage-based service. Billing and payment details live under AI usage settings, separate from the learning experience settings themselves.

• On the learner side, it appears as a floating action button (FAB) and an inline search bar (if enabled) — both configurable, including FAB color, position, and where the search bar displays.

Key Takeaways

• Don't rely on the FAB alone. Not every learner knows to look for it, so deciding where the search bar lives — inline in the header, floating, or hidden — is worth prioritizing for learner engagement, not just an afterthought.

• Small branding tweaks go a long way. FAB color and position, and the default support link (which points to Thought Industries support unless changed), can all be customized in a few clicks from the settings.

• Content exclusion works at a global level, not per panorama. You can fully turn Conversational AI on or off within a specific panorama, but tag-based content exclusion rules apply site-wide today. Excluding a subset of content from just one panorama while still surfacing it elsewhere isn't natively supported yet.

• Answer quality is an active, ongoing investment. The team is improving results using real usage data and beta customer feedback, so flagging inconsistent results or empty-answer queries directly helps prioritize fixes.

• No bring-your-own-model support yet. Running Conversational AI through a custom AI provider, API key, or local model isn't available today, though it's expected to come up in future discussions.

• There's a free trial, but no self-serve spending cap. The first 14 days are free once enabled, but if you want a spending cap or a prepaid budget arrangement, that's set up through your Account Manager rather than in the settings.

Q&A Recap

• Does Conversational AI scan the "resources" area (tabs on detail pages)?

  • It indexes the asset library — if those resources exist as assets in the library, whether attached to a course directly or added via the global resource library, they should be indexed and surfaced as sources. Content types like PDFs uploaded to a course page are also indexed. If a specific resource type isn't showing up as expected, sharing the exact example allows the team to test and confirm directly.

• What counts as a "query"?

  • Each learner question and the AI's answer together count as one query — a single question-and-answer exchange, regardless of answer length.

• What if I am only seeing AI usage settings, not Conversational AI Learning settings?

  • Conversational AI Learning settings are currently gated behind a feature flag. If you only see AI usage and not the learning settings, reach out to Support to have it enabled for your account.

• Can text be added to the FAB icon, not just a color change, to make it more obviously an AI search icon?

  • Not currently supported — today you can change color and position, but not add a text label. The team is taking the idea back for consideration and suggested submitting it as a community idea so it can be tracked. The search bar option is currently the best way to accomplish this type of engagement.

• Can content exclusion rules be created separately for each panorama, rather than globally?

  • Not today. You can turn Conversational AI fully on or off at the panorama level, but content preference rules (like tag-based exclusions) apply globally. Tagging content consistently across panoramas can approximate some use cases, but it doesn't solve for excluding a single tag from just one panorama when that same tag is also used for provisioning. This was flagged as a good candidate for a community idea so real-world use cases can be gathered.

• If no answer is found for a question, does it still count as a billable query?

  • Yes, currently a query with no answer still counts. The team is actively working to improve the answer success rate (currently around 98%) and is looking at ways to address empty-answer queries, including from a billing fairness standpoint.

• Can Conversational AI be run through our own preferred AI provider, API key, or locally?

  • Not today. This is expected to come up in further discussions in the coming weeks, but there's no bring-your-own-model or custom API key option currently.

• Can search results be made more consistent? (e.g., searching "color" vs. "colour" returns different content)

  • There's an ongoing initiative to improve answer quality using real usage data, including feedback and testing from customers. Spelling and regional variations are a known, fixable pattern — sending specific example queries to the team directly allows them to investigate and improve consistency.

• Is there a trial period to test it before committing?

  • Yes, the first 14 days are free once enabled. It's self-serve, though activating it requires providing payment or ACH information upfront.

• Since this is a chargeable, usage-based service, can we set a cap or prepay a fixed amount to avoid unexpected bills?

  • There's no self-serve spending cap today. The recommended path is to connect with your Account Manager, who can help arrange a prepay or budgeted arrangement.

• Why don't the AI Usage dashboard and the downloaded query history show the same number of queries?

  • The AI Usage dashboard displays queries for the selected date range, while the downloaded query history includes all queries recorded over time. If you've used AI features outside the selected date range, the download total will be higher.

Check out our support article or microcourse on this topic to learn more!

Join us monthly for our What’s New? Webinar to understand the impact of our product releases and latest innovations. Each session delivers a dynamic, visual walk-through of our most impactful enhancements including what’s new and what you should do next.

Need to catch up on what happened? Find Release Notes here.

We’ll take a tour of the new capabilities and how they work, where they shine, and how to put them to work quickly. This is your shortcut to staying current in the platform, so you can turn Thought Industries updates into your business momentum!

🚀 Join us each month at 11am EST for an overview of our most impactful releases from the past month.

Note: While we won’t be providing live support or troubleshooting during this time, we’re always happy to point you in the right direction or connect you with resources that can help.

1. Learner launches the External Activity from the course.

2. Learner completes the activity in the external app.

3. The external app POSTs a completion JWT back to TI and redirects the learner back into TI.

Issue:
The completion is registering correctly — the External Activity shows as complete in TI, and I've confirmed the return JWT is well-formed and valid (correct claims, correctly signed with the shared secret). However, instead of landing the learner back on the course/activity page, TI redirects them to /not-found.

What I've checked so far:

• JWT signature and payload are correct (verified against the External Activity spec).

• The activity itself does register as completed in the backend/reporting — so the JWT is being accepted.

• The redirect URL being used is the one issued by TI for the return trip (not a hardcoded URL on my end).

  Any pointers — especially from anyone who's implemented External Activity with a custom app and hit this — would be much appreciated. Happy to share the JWT payload structure or my redirect handling if it's useful for debugging.

Join us monthly for our What’s New? Webinar to understand the impact of our product releases and latest innovations. Each session delivers a dynamic, visual walk-through of our most impactful enhancements including what’s new and what you should do next.

Need to catch up on what happened? Find Release Notes here.

We’ll take a tour of the new capabilities and how they work, where they shine, and how to put them to work quickly. This is your shortcut to staying current in the platform, so you can turn TI updates into your business momentum!

🚀 Join us live on the first Wednesday of each month at 11am EST for an overview of our most impactful releases from the past month.

Note: While we won’t be providing live support or troubleshooting during this time, we’re always happy to point you in the right direction or connect you with resources that can help.

Here's what that means in practice:

Conversational AI Learning means your learners don't have to go dig through courses to find an answer, they just ask, in plain language, and get an instant response pulled straight from your own content, with a citation so they know exactly where it came from. And if the answer lives inside a course they haven't taken yet, the Enroll Me Agent lets them enroll right from that same conversation.
→ Learn more about Conversational AI Learning

AEO works a little differently. It's less about answering questions inside your academy, and more about making sure your content shows up when people ask tools like ChatGPT or Claude a question your academy already has the answer to. That matters in two ways: your customers get a trustworthy answer no matter where they went looking, and prospects doing their research run into your expertise before a sales conversation ever starts.
→ Learn more about AEO

And we're not slowing down - here's what's cooking in Beta

Both of the above are fully GA now, but this is just one chapter of where Omnichannel Learning is headed. Here's what's next:

• Thought Industries for Claude has been in Open Beta since July 10, bringing governed, agentic authoring and content workflows directly into Claude.

• Thought Industries for Slack opens up for Open Beta starting today, ask a question, get a cited answer from your academy, all without leaving Slack.

To be clear: these two are still Beta, not GA like Conversational AI Learning & AEO, but getting in is easy. Head over to the Apps area to enable these on your site. If you need assistance, support is always here for you.

And if you've got developers building on top of the platform, now's a good time to send them to our developer resources. This is a great resource to bookmark!

We are genuinely curious to hear from this community! If you were part of the beta, what's been most useful so far? And if today's your first time seeing this, ask away below. We'll be in the comments.

Not connected yet? You're still welcome to join and watch it in action.

Seats are limited, so register early.

These sessions are not for troubleshooting installation. For help getting connected, visit our Support Article.