Google is bringing Gemini-powered video understanding to Ask YouTube in the coming months. It can draw on a video's frames, audio and transcripts to answer viewer questions. For UK service businesses using video to attract customers, that changes what is worth filming and how to film it.
Google's new Gemini video understanding system for Ask YouTube will draw on a video's frames, audio and transcripts, meaning service businesses should make video content that AI can follow, not just human viewers.
According to Search Engine Journal on 2 September 2026, Google is bringing new video understanding technology to Ask YouTube on video watch pages in the coming months. The system lets Gemini choose which frames, audio and transcript segments of a video to inspect, answering viewer questions about the content on demand.
For UK plumbers, electricians, dentists and other service businesses using YouTube to attract customers, this is a meaningful shift in how your videos get discovered and understood by potential clients.
Until now, YouTube search has leaned heavily on titles, descriptions, tags and transcripts. A potential customer searching for 'emergency boiler repair Manchester' would find videos based on those text signals.
The new Gemini system adds to that. It can look at what is happening on screen, what you are saying in the audio, and how the transcript ties it together. When someone asks YouTube a question about a topic your video covers, the AI has more to work from than the exact words in your title.
This matters because many service business owners film videos the wrong way round. They optimise titles and descriptions for search, then film whatever feels natural. The new system raises the stakes on the video itself. What you show and say matters as well as the metadata around it.
Start by speaking your expertise aloud. If you're an electrician replacing a consumer unit, say exactly what you're doing as you do it. 'I'm now isolating the main switch before removing the cover' gives the AI clear audio to work from.
Show the problem and the solution in the same video. Customers searching for answers want to see the before state, the work being done, and the after result. Because the system can inspect frames as well as words, what is visibly on screen, corroded pipes, blown fuses, cracked tiles, becomes part of what it can work from.
Use plain, consistent terminology, and say the everyday name as well as the trade name. If you say 'stopcock, sometimes called a stop valve', a customer using either phrase has a better chance of landing on your video.
Film in well-lit conditions with stable framing. The system analyses video frames, so blurry, dark or shaky footage gives it less to understand. A £30 phone tripod and a work light solve this.
YouTube is widely described as the world's second-largest search engine. When a boiler breaks at 11pm, plenty of people search YouTube for 'boiler not firing up'. They want to see if it's something they can fix themselves or if they need to call someone in the morning.
If your video appears in that search and clearly shows the problem they're experiencing, you've just become the expert they trust. When they decide they need a professional, you're well placed to be the first call.
Better video understanding makes this more useful because it can help match intent, not just keywords. Someone asking 'why is my tap dripping' might find your video titled 'How to replace a worn washer', because the problem and the solution are the same thing.
Start with your ten most common callouts. For a plumber, that might be dripping taps, running toilets, low pressure, no hot water, radiator cold spots, boiler error codes, frozen pipes, blocked drains, leaking joints and noisy pipes.
Film a short video for each one showing what the problem looks like, what causes it, and how you fix it. Speak clearly throughout. Show close-ups of the faulty component and the replacement. These videos become your searchable knowledge base.
Then film your emergency scenarios. 'Burst pipe what to do first' and 'boiler leaking water emergency' are high-intent searches from people who need help right now. A three-minute video showing them how to isolate the water and minimise damage, ending with how to reach you, turns viewers into enquiries.
YouTube videos appear in Google search results, especially for how-to queries. When someone in your area searches for a service you provide, Google often shows video results alongside traditional web pages.
Better video understanding could help here too, because context counts for more. A video about 'replacing a bathroom tap' is relevant to someone asking about a 'leaking basin mixer' or a 'dripping bathroom sink', and a system that understands the footage has a better chance of making that connection.
This is where Rosie, our Rankings and GEO Agent, fits in. Rosie's job is making sure a business is visible on Google and in AI search, publishing content built around the local search terms that matter in your area. Video you film yourself sits alongside that work and gives both people and AI systems more evidence of your expertise.
The biggest error is filming once and forgetting about it. A channel with one video from three years ago tells a customer very little. A steady run of recent, useful videos shows you are active and gives you more chances to be found.
Second is poor audio. You can film on a phone, but if the audio is muffled by wind, traffic or machinery noise, there is little useful information to extract from what you say. A £20 lapel microphone solves this.
Third is assuming video is just for brand awareness. Video can be a direct source of enquiries. Every video should end with a clear next step: call this number, visit this website, book through this link. People who find your video in the middle of a problem need to know how to reach you.
Google says the technology will reach Ask YouTube on video watch pages in the coming months, according to Search Engine Journal, with no exact date, regions or languages confirmed yet. That gives service businesses time to build a library of clear, well-filmed videos before it arrives.
Service businesses that start filming structured, well-spoken, clearly-shown content today will have a body of useful videos in place by the time this becomes standard. Those who wait will be starting from nothing against channels that already have it.
The barrier to entry is low. You need a smartphone, basic lighting, a microphone, and the willingness to explain your work on camera. The aim is simple: appearing when people with a real problem go looking for help.
Video is becoming hard to ignore. It is an increasingly common way for potential customers to discover and evaluate service businesses online, and better video understanding pushes further in that direction by making video content more searchable.
For UK plumbers, electricians, dentists, fitness trainers, solicitors and estate agents, this is an opportunity to demonstrate expertise at scale. One well-filmed video explaining a common problem can keep being found for years, especially as AI systems get better at understanding and surfacing relevant content.
The businesses that do well will be those that treat video as a proper part of their visibility, not an afterthought. That means regular filming, clear explanations, and content that AI can follow as easily as a person can.
Rosie helps service businesses get found in local search and AI search by publishing the right content in the right places. If you are not showing up when potential customers search for your services, you are losing work to competitors who are. Get your free AI visibility report at The AIgency free report.
The AIgency home · Q&A · Blog