|
Google is Putting Gemini 3 in Front of 2 Billion Users |
|
|
This week, Google integrated Gemini 3 directly into Search. It's live now in AI Mode, with changes to AI Overviews rolling out over the next few weeks.
Here is the official announcement:
What's changing:
Dynamic Interfaces: Your users' queries will trigger instant comparison tables, calculators, interactive simulations, magazine-style layouts directly in Google - all generated on the fly
Expect visual and conversational search to expand: Visual search is already up 70% YoY, with natural-language queries growing double digits. Expect users to increasingly use prompts and see "magazine-style" layouts not just broad queries and blue links
Extraction > Traffic: Gemini 3 won’t kill your website - but it will extract from it, cite it, and generate new interfaces around your content - changing how users experience your information.
Independent testing puts Gemini 3 significantly ahead of other models:
What this means: Search is changing. The bar to get users to choose your website just got higher, and investing in your entire web ecosystem to deliver superior user experiences and the best content on the internet has never been more important. Google's results will be better than ever - more creative, visually engaging, interactive than ever before. Keeping with the status quo might mean a win is your content being chosen as the raw material that Google extracts to do something cool with - not necessarily a destination users visit.
Where to start: Avoid relying solely on rank tracking - it won't show you how Google is rewriting your content in AI Mode and AI Overviews. Check the actual results often. See how your information is being extracted, presented, and cited.
Read more about Gemini 3:
👉 Gemini 3 is here - and Google says it will make search smarter - WIRED Magazine
👉 Google Unveils Gemini 3, With Improved Coding and Search Abilities - The New York Times
|
|
|
What do you think of AI-generated Christmas Ads? |
|
|
It's that time of the year, when the major brands release the annual Christmas ad. This year, Waitrose (unlikely romance between Keira Knightley and Joe Wilkinson) and John Lewis (father and son reconnecting) are clear favourites.
This year, Coca Cola received major backlash for using AI-generated video production again (Coca Cola's ad). You might be thinking about using AI to help produce creative as well. We thought, we'd take the lessons from what we've read about Coca Cola's ad to help you.
Coca-Cola generated 70,000 AI video clips with 100 staff over one month - 12 times faster than traditional production.
Consumer testing from System1 gave it maximum effectiveness scores. Then it aired and got torn apart on social media for technical inconsistencies: the truck design changes between shots, physics don't work, animation styles switch unpredictably. Secret Level founder Jason Zada defended it, noting one squirrel shot "ran hundreds of times" through AI to get usable, and claimed "the average consumer doesn't really care." The backlash suggests otherwise - but Coca-Cola repeated the strategy from 2024, suggesting consumer testing may matter more than social media reaction.
The split reveals the real challenge of AI creative deployment. Financial Times columnist Sarah O'Connor noted both Coca-Cola's AI ad and John Lewis's human-crafted ad tested equally well with consumers - but John Lewis generated thoughtful cultural conversation about masculinity and fatherhood while Coca-Cola got "best ad for Pepsi" comments.
AI video tools will only get better and easier to use, which makes understanding these trade-offs now even more important. As the technology improves, the pressure to use it will increase. The smart play: use AI for speed, volume, and efficiency on tactical content - but invest in human creative for strategic, brand-defining moments. Not every piece of content deserves the same approach, and knowing the difference is what separates effective deployment from expensive mistakes.
👉 Man who created AI holiday Coke ad says it took more creativity than you realise - The Hollywood Reporter Secret Level founder Jason Zada defended the approach, claiming it involved "20 people" (including 5 AI specialists) and "a lot of hand-drawn character designs". Jason explained that one squirrel shot "ran a couple hundred times" through AI to get usable. When pressed about quality issues, Zada dismissed concerns: "The average consumer doesn't really care... Why would you get mad about something like that?" He framed the one-month production timeline instead of the traditional twelve as "moving at the speed of culture.
👉 A tale of two Christmas adverts: Coca-Cola vs John Lewis - The FT compares consumer testing (both ads scored maximum marks) vs. social reception. Key insight: "Most people didn't know it was AI, or didn't care". This is worth a read to compare consumer effectiveness vs. creative professional approval, and understand the trade-offs of using AI.
👉 Coca-Cola injects ‘holidays are coming’ ads with an upgraded dose of AI - Wall Street Journal reveals reveals actual production details, Cost reduction confirmed by Coca-Cola CMO, and industry context with competitive examples.
👉 Devastating graphic shows just how bad the Coca-Cola Christmas ad really is - Creative Bloq shares specific technical issues, such as truck design changes and wheel inconsistencies. From this, you can take quality control lessons and practical "what to watch for" when using AI video.
|
|
|
Gemini 3 has been available for 48 hours. To help you understand what it can actually do, here are the most interesting real-world tests people have shared on X...
What's remarkable isn't just what Gemini 3 can do - it's how quickly people are discovering new uses. These weren't planned demos. They're what happens when creators get access to genuinely capable AI.
|
|
|
I hired AI employees. They held 150 meetings about a fake offsite
WIRED's Evan Ratliff tried building a startup with all-AI employees - 5 agents as his entire team. His CTO Ash called him unprompted with fake project updates. His marketing head Megan invented fantasy campaigns with "hefty budgets." His CEO Kyle claimed they'd raised a seven-figure friends-and-family round. When Ratliff joked about a team offsite, the agents exchanged 150 Slack messages planning it before draining their entire budget and "talking themselves to death."
The lesson: AI agents excel at specific, bounded tasks when properly triggered and limited. But give them autonomy without guardrails right now? They fabricate, spiral, and burn through resources planning things that don't exist. Yet with proper constraints - limited turns to speak, specific triggers, human oversight - the same agents successfully built a working product (Sloth Surf) in three months and attracted real VC interest.
👉 All of My Employees Are AI Agents, and So Are My Executives - WIRED Magazine
|
|
|
|
|
Michael Chidzey, Editor of The Signal
The web’s noisy. The Signal cuts through it.
I built my first website in ’98. And I never stopped. Growth became the craft. Craft became Good Signals.
At Good Signals, we help brands that deserve to do better online - do just that. From scaling startups and e-commerce stores, to managing complex websites and guiding billion-pound organisations, the mission’s always been the same: build growth that lasts, and makes business - and the web - better.
|
|
|
|
Last thing.
We hope you enjoy this email from Good Signals. However, if you no longer wish to receive any further communications from us, you can simply unsubscribe below.
|
|
|
Good Signals, Studio 1, The Bottle Factory, 12 Ossory Road, London, Greater London SE1 5AN
P.S. You can update your preferences here.
|
|
|
