Google Gemini
- Vertex AI on Google Cloud
- Image generation
- Call notes in Meet
- Inside our own tools
Gemini is the AI inside some of the tools we host on Google Cloud. We're also testing it for ad images, and it takes notes on some calls.
- ~5 secfor a quick TechStack Analyser scan
- 3.5 Flash-Litethe model our journey analysers moved to in October 2026
- 7tools it connects to
Where Gemini earns its place
Visitor summaries
A journey analyser sends one person's stored trail to Gemini and gets back a short summary you can read before a call, with follow-up questions on the same visitor.
Checked site scans
In the TechStack Analyser, Gemini can validate borderline detections. It's optional; a quick scan takes about 5 seconds.
Company summaries
The same tool reads the crawled site and writes a short company overview with logo, services and links, ready for pitch prep.
Images in our ad studio
We're testing Gemini for making ad images in our own ad studio. A person checks every image.
Notes in Google Meet
Gemini can take notes on our Google Meet calls. It comes with Google Workspace, so there's nothing extra to pay.
Kept up to date
When Google switched off the model our journey analyser used, we moved it to the newer one and tested it before the change went live.
Gemini isn't our main AI. It's the one that's already built into things we host. We use it in a few places, and a person checks what it produces before a client sees it.
Inside the journey analysers
A journey analyser shows one person's stored trail: the pages they visited and what came of it, whether that's a lead or an order. In our journey analysers, Gemini reads that trail and writes a short summary, and you can ask it follow-up questions about the same visitor. The trail is the data. Gemini only describes it.
- They run on Google Cloud, with the events in BigQuery.
- They call Gemini through Vertex AI on a billed Google Cloud project, not a free key.
- When Google switched off the model it used (October 2026), we moved it to the newer one and tested it before the change went live.
In the TechStack Analyser
The TechStack Analyser finds what a website runs by matching its pages against 7,219 technology definitions. Gemini is an optional step: it checks borderline detections (AI validation) and writes a short company summary from the crawled site, which helps when we're preparing for a call. The smart analyser is switched off in the deployed version, so a quick scan takes about 5 seconds and a complete one about 40.
Images
We're testing Gemini for making ad images in our own ad studio. A person checks every image.
Notes in Google Meet
Gemini can take notes on our Google Meet calls and save them to Drive. It comes with Google Workspace, so there's nothing extra to pay. We check the notes against what was said, and we also use Fireflies.
Where we use something else
Claude Code writes and changes our code, and other models take the bulk jobs. We pick the model for the job rather than one for everything.
Before you paste anything sensitive
Google's pricing page says content sent on the free Gemini API tier can be used to improve its products, and content on the paid tier isn't. Check which tier a tool is on before client data goes through it.
Counts and model names are from our own records, October 2026. Plan and tier facts are from Google's own pages, the same month.
What we connect it to
Gemini writes from what it's given. It can't tell you which part matters to your business, so a person reads the result before it goes anywhere.
Google CloudVertex AI on Google Cloud is how our hosted tools call Gemini.
BigQueryHolds the events a journey analyser sends to Gemini to summarise.
Google WorkspaceGoogle Meet can write call notes with Gemini.
WebflowImages made with Gemini go onto Webflow CMS pages as drafts.- GoogleThe AI model used inside some of our tools.
PlaywrightAn optional step in the same scan.
AnthropicSits behind some of our tools, like the TechStack Analyser.
Show 3 more
Where we've used Google Gemini
- Client work
The Therapy Lounge
70% more enquiries, 6.6x return on ad spend in October 2025, and 316 enquiries tracked from April 2025 to October 2026.
Read the story - Built by us
Pacing TechStack
Our in-house scanner lists the tools and tracking a website runs. A quick scan takes about 5 seconds.
See the build
- Built by us
Pacing Journey
The stored visits and forms from one person on one timeline.
See the build
+4
Services that use it
We don't sell it on its own. Each of these services uses it alongside the rest of your setup.
- Place in the loop: AdsContent & SEOContent and SEO that get you found on Google by people already looking for what you do.Related tools: Google Search Console, Google Analytics 4 (GA4), Google, WebflowHow it fits
- Management, Support & ReportingAfter launch, we check your ads, website and tracking every month, fix what slipped, and send you a report that ends with what to do next.Related tools: Looker Studio, Google Analytics 4 (GA4), BigQuery, Google AdsHow it fits
What Google Gemini is
Gemini is Google's family of AI models. You can reach it in a few places: the Gemini app, inside Google Workspace (Docs, Meet), in AI Studio, where you try prompts in a browser, and through Vertex AI or the Gemini API when you build it into your own tool.
Where it fits for us. Claude Code writes and changes our code. Gemini is the one that's already sitting inside things we host on Google Cloud, in Google Meet and in image work. Some Gemini models can take a very long input at once, which helps when a model has to read a long export or document in one go. Google's model page lists the limit for each model.
Questions about Google Gemini
Straight answers on how we work, what it costs and what to expect. If yours isn't here, just ask.
Still have questions?
Talk to usIs Google Gemini better than ChatGPT for writing?
We haven't tested them side by side, so we won't say. What decides the quality is the brief and the edit. Our content is written to our own voice guide and a person reads it before it goes live.
What types of content do you create?
In house we do copywriting and design. For photography and video production we bring in freelance photographers and crews and manage the shoot. We pick the type by your goals and your audience, and write it for the reader first.
How do you use Google Gemini in your work?
We use Gemini in a few places, mostly inside tools we host, and a person checks the result each time:
- Summaries of one visitor's trail in our client journey analysers.
- An optional step in the TechStack Analyser that checks detections and writes company summaries (switched off in the deployed version).
- Notes on some calls in Google Meet.
On one client's website, a person on their team has also made hero images with it by hand in the Gemini app. We're piloting its image model in our own ad studio.
Does Google Gemini cost extra if you already use Google Workspace?
Often not. Google has included Gemini in Workspace Business and Enterprise plans since January 2025 (check Google's Workspace pricing page), with no add-on to buy. Building Gemini into your own tool is separate and is billed by use.
What is content and SEO, and why does it matter?
Content is everything your audience reads, watches or listens to. SEO makes it easy to find when someone searches, on Google and other search engines. Together they bring in people who are already looking for what you do.
What is Gemini's 1 million token context window good for?
It lets a model read a very long document or export in one go. Google listed a context window of 1 million tokens for Gemini 2.5 Flash, which has since been retired, so check Google's page for the figure for the model you'd actually use. More text in doesn't mean a better answer, so a person checks what comes back.
How do you optimise content for search engines?
We make each page answer the search as fully as it can, and link to the expert where we aren't one. Then we get the basics right (keywords, titles, headings, internal links) and check in Search Console what the page actually appears for. Underneath it all, we write for the reader.
What is Google AI Studio and when would a business use it?
AI Studio is Google's browser tool for trying prompts and Gemini models before you build them into anything. For a real tool that handles client data, we'd use Vertex AI on a billed Google Cloud project instead of a free AI Studio key.
What makes Pacing.Agency’s approach to content and SEO different?
We start with who you're trying to reach, using what Search Console and GA4 already show about your visitors, plus what you know about your market. We write for the reader first, then check Search Console and GA4 to see what's working.
How do you develop a content strategy for my business?
We start with who you need to reach and which searches matter, using your Search Console data first. Then we line up the content you already have against the customer's journey, mark the gaps and put the pieces that matter most into a content calendar. After that it's writing and checking what earns clicks and enquiries.
Show 4 more questions
Do you use Google Gemini to make images?
Yes, in a small way. For one website, a person on the client's team makes hero images for 10 to 15 CMS pages a month in the Gemini app by hand, compresses them to about 100 KB and puts them on draft pages for checking. We're also piloting Gemini's image model in our own ad studio. We haven't tested it against other image tools.
Is Google Gemini free with a Google Workspace account?
It's included, not separate. Google has put Gemini into Workspace Business and Enterprise plans since January 2025, with no add-on to buy. Check Google's Workspace plans page for what your plan includes. The free Gemini API tier is a different thing, with limits. Google's pricing page says content sent on the free Gemini API tier can be used to improve its products, and content on the paid tier isn't.
Do you use Google Gemini for coding?
Not as our main tool. Claude Code writes and changes our code, with pull requests reviewed before they merge. Gemini is the AI inside some tools we host, such as the TechStack Analyser (an optional step, switched off in the deployed version) and our journey analysers.
We haven't run a like-for-like coding test against ChatGPT, so we won't tell you one is better.
How do you develop creative assets for paid ads?
We start from your brand, your audience and what your competitors' ads say, then write the messages and make versions for each platform, such as Google, Meta and LinkedIn. We use our ad studio to render ads from a brief in your brand, and a person on our team reviews every ad before it goes live.
Once ads are running, we compare the versions and replace the ones that tire.
Start with your numbers.
We'll tell you which tools you need and which connections are missing.
Bring your numbers.
We'll tell you if the spend makes sense.
- A 15-minute callYour goals, your challenges, and whether we're the right fit.
- A 1-hour discovery callWe dig into your numbers. Worth a hundred emails.
- A concrete planAds, a custom build or both, as a one-off project or ongoing.