Ask YouTube starts your video at the answer. Does your next step start there too?
Ask YouTube plays long videos from the answering timestamp, and those plays count as views. Move the funnel step into the chapter that gets clipped.
- Ask YouTube starts long videos mid-way and counts the preview as a view, so views can rise while clicks per view fall.
- End screens run only in the last 5 to 20 seconds, which most clip viewers never reach.
- Put a one-line spoken next step and a timed card inside each chapter that answers a question.
- Tag every card and description link by chapter and judge results per 1,000 engaged views, not per view.
On 12 October 2026 Ask YouTube reached phones for every signed-in viewer in the US, which completed its spread across web, mobile and TV. Three days later YouTube added a test that puts an Ask YouTube answer above the ordinary results for some desktop searches, mostly how-to and skill questions. The feature has been public since July. What changed in October is reach.
Most of the coverage asks how to get a video into the answer. This piece asks a narrower question that matters more if your videos sit inside a funnel. When Ask YouTube plays your 20-minute walkthrough from minute nine, where does that viewer go next, and can you see it?
What does Ask YouTube actually show?
YouTube's help page describes an answer made of text, long-form videos and Shorts. Clips carry the video title and channel name. On a computer, a video plays on hover, starting at the timestamp YouTube judges most relevant. Clips are ranked on relevance, engagement and quality, in the same way as standard search. Playlists and Premium-only videos cannot appear. The feature is closed to supervised accounts, YouTube Kids and Apple TV devices, and YouTube says the conversations are not used to show ads.
Early testers saw the same shape. In April, Search Engine Journal described a text summary with one main video linked to a timestamped section, followed by galleries of long videos and Shorts. Digital Trends reported long-form videos linked to the relevant section rather than the start.
The unit YouTube retrieves is a passage, not a video. That is the whole funnel problem in one line.
Who can see it, and since when?
The dates are worth getting right, because they tell you which weeks of your data to compare.
| Date (2026) | Change | Source |
|---|---|---|
| April | Opt-in test for US Premium members aged 18 and over | Search Engine Journal, Digital Trends |
| 19 May | Announced at Google I/O, still Premium only | YouTube Blog |
| 6 July | All signed-in US viewers aged 13 and over, English, desktop | PPC Land, Search Engine Land |
| 12 October | Mobile added, completing web, mobile and TV | Digital Trends, The Keyword |
| 15 October | Answers tested at the top of ordinary desktop search for some users | The Keyword, Android Authority, Social Media Today |
One report disagrees. MediaPost wrote on 13 August that the tool was already open to everyone aged 13 and over on desktop, mobile and TV. Every October report, including those that cite YouTube's 12 October post, treats mobile as new that week. We read the August item as describing the announced scope, not the rollout, and use 12 October.
Why can views rise while your clicks fall?
Two rules combine here. In July, YouTube said views of Shorts, videos and previews inside Ask YouTube answers count toward total views and Partner Program eligibility. Since 24 August, a public view also counts from the first frame, as we covered in our piece on engaged views. So a hover preview that plays a few seconds of chapter three can add a view.
The next step in most funnel videos lives somewhere else. YouTube's end screens run only in the last 5 to 20 seconds of a video. The external link element needs Partner Program membership, end screens do not show on mobile web apart from iPad, and YouTube says they may not always appear. A viewer who starts at minute nine of a 22-minute video and leaves once the question is answered never gets there.
The count goes up. The share of viewers who see your offer goes down. Neither number is wrong. They measure different things.
Is this the AI Overviews click problem again?
Partly, and the evidence comes from web search, not YouTube. Pew tracked 900 US adults through March 2025 across 68,879 Google searches. When an AI summary appeared, people clicked a traditional result on 8% of visits, against 15% without one. They clicked a link inside the summary on 1% of visits. Seer's 2026 update, built on 53 brands and 5.47 million queries, found informational click-through of 2.07% when the brand was cited in the AI Overview, 0.94% when it was not, and 3.35% with no overview at all. Seer calls this correlation, since stronger brands are cited more often, and notes that one large account pulls the not-cited figure down.
Two readings follow. Android Authority warned on 16 October that viewers may stop clicking through to videos at all. YouTube frames the feature as one more way to be found, and its counting rule protects the view. Both are right about their own metric. Being the clip beats being left out. But on the web evidence, an answer box costs clicks even for the cited source, and nobody has published the YouTube version of that number. YouTube has shared no usage or click data for Ask YouTube.
What you can control is where the next step sits relative to the clip.
Do chapters decide which part gets played?
They help, but they do not decide. In July YouTube told creators to publish unique, high-quality videos with "clear chapters and descriptive titles", so its systems can match segments to questions. A creator-education blog, KDCC, said after live tests in May that almost every answer it saw came from a chapter. That was a handful of searches on a stream, with no count. YouTube's help page lists only relevance, engagement and quality.
The fair reading is that chapters are a named aid, not a switch. They matter for a funnel reason too. A chapter gives the clip a clean start, and it gives you a defined place to put the next step. YouTube's rules are simple: the first timestamp is 00:00, you need at least three in ascending order, and each chapter runs at least 10 seconds. Automatic chapters are on by default for new uploads, though not every video gets them.
A worked example (hypothetical)
These numbers are made up and rounded to show the arithmetic. They are not data.
A software company runs a 22-minute product walkthrough. Chapter three answers a common question. Before Ask YouTube, the video gets 10,000 views a month. A quarter of viewers reach the end screen, and 2% of those click it.
| Views | Clicks | Clicks per 1,000 views | |
|---|---|---|---|
| Before | 10,000 | 50 | 5.0 |
| Plus 4,000 clip views, no change to the video | 14,000 | 56 | 4.0 |
| Plus a card and a spoken line in chapter three | 14,000 | 96 | 6.9 |
The middle row assumes 30% of clip viewers keep watching, a quarter of those reach the end, and 2% of those click. That adds 6 clicks. The rate falls by a fifth even though the video did nothing wrong. The last row assumes 1% of the 4,000 clip viewers act on a card timed to the end of chapter three's answer. That adds 40.
The point is not the 1%, which is an assumption. The point is that the clicks you can add sit inside the clipped chapter, and the end screen cannot reach them.
Where should the next step go, by video type?
| Video type | What a clip usually covers | Where to put the next step | End screen |
|---|---|---|---|
| Short how-to with one answer | Most of the value | One spoken line right after the answer, plus a card at that moment | Keep, but expect few to see it |
| Long walkthrough or demo | One step of several | A card at the end of each chapter, pointing to the full video or a trial | Keep for full watchers |
| Webinar replay | One audience question | Chapter per question, with a card to the next live session in each | Keep |
| Comparison or review | The verdict | The next step straight after the verdict chapter | Keep |
| Case study | Rarely a clean answer, in our view | One chapter that states the result in plain words, then the next step | Main call to action stays here |
Cards have limits. You get five per video, link cards need Partner Program membership, and neither cards nor end screens work on videos set as made for kids. Without a link card, point a video card at a short video whose description carries the link. If you run webinars, the replay row ties into what a replay viewer is worth.
Can you measure Ask YouTube traffic?
Not directly. YouTube's list of traffic sources names fifteen internal sources, from Browse features to YouTube Search, and none of them mentions Ask YouTube, AI answers or previews. We could find no YouTube document saying where these views are filed. Retention key moments help, but only on videos of at least 60 seconds with 100 views, and the segments split viewers by organic or paid, new or returning, and subscriber status, not by search surface.
So measure what you place. Tag every card and description link with the chapter it belongs to, for example a campaign content value of ch3. Then read clicks per 1,000 engaged views, not per public view. A video whose views rise while average view duration falls, with a new bump in retention at a chapter start, is the pattern to watch for.
Ask YouTube is also one surface among several. If you want to know where video earns citations outside YouTube, see which AI engines cite YouTube.
A seven-step check
- List the funnel videos. Any long-form video with a link or offer that leads to your site.
- Write the questions each one answers. Three to five per video, in the words a viewer would type or say.
- Fix the chapters. Start at 00:00, at least three, each at least 10 seconds, titled as the answer, not as a label.
- Run the questions in Ask YouTube. This needs a signed-in US account searching in English. Log whether your video appears and which timestamp it starts at.
- Place the next step inside each clipped chapter. One spoken line after the answer, and a timed card at the same point.
- Tag links by chapter. Use a separate campaign content value for each chapter's card and description link.
- Review after four to six weeks. Compare clicks per 1,000 engaged views with the weeks before 12 October, chapter by chapter, and move the step where the clicks are.
Where this stops
YouTube has published no figures on how often Ask YouTube answers appear, how often viewers open the full video, or where those views land in Analytics. The search-results test covers only some desktop users in the US. Everything here about lost clicks is inferred from web search studies and from where YouTube allows funnel elements to sit. The fix costs little either way. A clear next step inside the chapter that answers the question helps full-length viewers too.
If you want help working out which of your pages and videos AI answers pick up, join the waitlist.
Questions
Who can use Ask YouTube?
Signed-in viewers in the US aged 13 and over, searching in English, on computers, phones and TVs. Mobile was the last surface, added on 12 October 2026. It is not available to supervised accounts, in YouTube Kids or on Apple TV devices.
Do views from Ask YouTube count?
Yes. YouTube said in July 2026 that views of Shorts, videos and previews shown in Ask YouTube answers count toward total views and Partner Program eligibility. Since 24 August 2026 a view also counts from the first frame, so a short hover preview can add a view.
Can I see Ask YouTube traffic in YouTube Analytics?
Not as its own line. YouTube's list of traffic sources does not mention Ask YouTube, AI answers or previews, and YouTube has published no reporting for the feature. Chapter-tagged links and changes in watch time per view are the best proxies for now.
Do chapters decide whether a video is used in an answer?
Not on their own. YouTube ranks clips on relevance, engagement and quality, as in standard search, and has named clear chapters and descriptive titles as aids to matching. Chapters also give the clip a clean start point and a place for your next step.
Sources
- YouTube Help, Ask YouTube support.google.com
- YouTube Blog, All the YouTube news from Google I/O 2026 blog.youtube
- YouTube Help, Video chapters support.google.com
- YouTube Help, Add end screens to videos support.google.com
- YouTube Help, Add info cards to videos support.google.com
- YouTube Help, Traffic sources for views support.google.com
- YouTube Help, Measure key moments for audience retention support.google.com
- Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results pewresearch.org
- Seer Interactive, AIO impact on Google CTR. 2026 update seerinteractive.com
- PPC Land, YouTube blocks Ask search feature for viewers under 13 and signed out users ppc.land
- PPC Land, YouTube counts views from the first frame across all formats on August 24 ppc.land
- Search Engine Land, YouTube expands Ask YouTube AI search to signed-in US desktop users searchengineland.com
- Search Engine Journal, Google tests Ask YouTube conversational search experiment searchenginejournal.com
- Digital Trends, Ask YouTube expands to mobile, web and TVs across the US digitaltrends.com
- The Keyword, YouTube tests Ask YouTube answers at the top of search thekeyword.co
- Android Authority, YouTube's new search feature takes a page from Google Search androidauthority.com
- Social Media Today, YouTube tests AI replies in searches socialmediatoday.com
- MediaPost, YouTube deploys chatbot for mobile users mediapost.com
- KDCC, Ask YouTube explained blog.kdcc.social
- Social Lady, Ask YouTube. What conversational search changes for your videos social-lady.com
See where your brand stands.
Your Digital Authority Score, who AI recommends for your target questions, and the placements that close each gap. Your first report is free.
Get your free report →TikTok search ads now pick keywords from your video. What should the video say?
TikTok's Smart+ search update builds keywords from the ad creative. What TikTok documents, what the video likely feeds it, and how to brief for search.
→2026.11.01 · 8 MINGoogle now tells you how many readers chose you as a preferred source. What is that number worth?
Search Console's first preferred source count is not traffic. What a selection changes, how it differs from a Search profile follow, and how to read it.
→2026.11.01 · 8 MINAI engines cite pages that answer first. Does the answer have to come first?
A new study links direct openings to AI citations. Within-site data says the answer's match matters everywhere, and its position mainly on Gemini.
→