Launches platform to entire existing customer base overnight with scalable Google Cloud infrastructure
Automatically populates form data in just 300 milliseconds with Gemini
Runs 100,000s jobs a day with just 15 developers thanks to Google Cloud managed services
With BigQuery and Cloud Run, Precis built a scalable data platform and added a co-pilot using Gemini models, giving customers a faster, smarter way to hone their marketing campaigns.
With BigQuery and Cloud Run, Precis built a scalable data platform and added a co-pilot using Gemini models, giving customers a faster, smarter way to hone their marketing campaigns.
For the tech-powered marketing agency Precis, data is key to driving marketing performance. A core part of the company's offering is helping customers invest their marketing budgets where they'll drive the greatest impact. This involves analysis of vast amounts of data. Previously, these insights were gained using Precis' internal data platform, built on third-party software, and were presented to customers by one of the company's consultants. However, the cost of this consultancy meant that only larger businesses could afford the service. Additionally, hiring and training new consultants took time, slowing Precis' ability to take on new customers when existing consultants were at full capacity.
Precis realized that giving customers direct access to data as a service would lower the cost for the customer, give them more control over how they use their data, and enable Precis to scale up its offering to reach more customers more quickly. Precis set about trying to transform its existing internal data platform into a customer-facing SaaS platform, "but we hit so many scaling limitations with our existing provider that we couldn't keep going," recalls Jonas Bæk, Partner and Head of Tech Product at Precis.
We looked around at the other solutions out there. Only Google Cloud offered all the services we needed.
Erica Petersson
Engineering Manager, Precis
The company realized the only way to scale the platform using its existing provider would have been to hire a large team of DevOps engineers to keep it running effectively. "We didn't want to hire a fully fledged DevOps team, so we looked around at the other solutions out there," explains Erica Petersson, Data Engineering Lead at Precis. "Only Google Cloud offered all the services we needed."
Building a new data platform to make a forecast of millions SEK ARR
Precis set about building its new customer-facing data platform on Google Cloud in 2021, using a microservices architecture hosted on Cloud Run, with BigQuery as its analytics database. By using these fully managed services, Precis was able to keep its team lean. "We're taking advantage of the fact that Google Cloud manages all our infrastructure for us, leaving us free to develop the business instead of worrying about the infrastructure," explains Bæk.
When Precis first launched its platform — named Alvie — to its existing customers, the scalability of Google Cloud meant it could immediately go from zero to serving its entire customer base overnight without any performance issues. The platform now serves twice as many customers as when it launched.
It came as a surprise that we could use BigQuery as a real-time database for customers to interact with through the UI, but it works really well. And the latency is so low that for the majority of our queries, the user barely notices they're loading.
Jonas Bæk
Partner and Head of Tech Product, Precis
"We see the platform grow every month and the system just scales automatically," says Bæk. "Our annual recurring revenue is high and forecast to grow significantly next year, so we're already operating at a very large scale. That wouldn't have been possible without the serverless capabilities of Google Cloud."
Now, Precis customers can use Alvie to access all the data they need from various marketing platforms in one place, along with pre-defined reports from Precis that help them understand what the data means for their campaigns. With BigQuery powering the underlying data insights, users can receive granular information about the success of specific parts of their marketing campaigns, along with automated data-driven recommendations about how they could tailor their campaigns for better returns. "All that data in the user interface comes straight from BigQuery," Bæk explains. "It came as a surprise that we could use BigQuery as a real-time database for customers to interact with through the UI, but it works really well. And the latency is so low that for the majority of our queries, the user barely notices they're loading."
Forging a faster path to smarter insights with generative AI
With Vertex AI Agent Builder, users can ask the specialist that knows the most about a given subject. They not only get their insights faster, but the agents have so much internal documentation that they can offer new perspectives the user might not have thought of.
Erica Petersson
Engineering Manager, Precis
Having put all this data into customers' hands, Precis soon realized this presented a challenge of its own. "Suddenly there was no expert consultant to tell customers all about their data," Bæk explains. "We needed a way to hold the customer's hand as they explored their data. When gen AI arrived, it was the obvious solution."
Precis joined the PaLM Trusted Tester Program as part of its launch, before implementing Codey, and later Gemini 1.5 Flash, as it built an assistant to enable customers to interact with the Alvie platform using natural language. Now customers can use this assistant to ask specific questions of their data and receive easy-to-understand answers, which the assistant pulls from Alvie's extensive how-to guides.
This assistant is helping Precis to win new business. "The appeal of being able to use gen AI in the user interface is helping us to open more doors in the sales process and ultimately close more deals," adds Bæk.
For its internal users, Precis has gone even further, organizing its Alvie assistant into specific marketing specializations to support its various specialist teams. By tagging all its best-practice and consultancy documentation by content, and connecting it to Vertex AI Agent Builder, Precis has effectively built a range of digital marketing specialists that its internal users can turn to for specific insights. For example, if the SEO specialist has a question about creative best practices on social media, it could turn to the dedicated agent for that field. The assistant uses Vertex AI Agent Builder to direct customer questions to the relevant documentation and combines this with the data stored in BigQuery to return highly relevant, data-driven recommendations in easy-to-understand, natural language.
"With Vertex AI Agent Builder, users can ask the specialist that knows the most about a given subject," explains Petersson. "They not only get their insights faster, but the agents have so much internal documentation that they can offer new perspectives the user might not have thought of."
Precis is also using Gemini behind the scenes to help customers get insights more quickly. When customers first use Alvie to analyze campaigns, there are a number of rules they need to define using dropdown boxes. Previously, this was a tedious process. Now, Precis has integrated Gemini into the workflow to auto-populate the necessary fields for the customer.
"Time-to-insights is what we're really focused on," explains Bæk. "How quickly can we get someone through the whole flow of setting up a model and into the dashboard? With Google Cloud it was really easy to install the software development kit (SDK), and it's super quick. We average around 300 milliseconds to add the information, so the customer barely even realizes we're doing it."
A firm foundation for future success
Precis is now continuing to expand its use of Gemini and explore more ways to use gen AI to improve the Alvie platform. The company recently tested Gemini's multimodal capabilities to automatically identify and tag the various visual elements of advertisements, enabling users to understand which elements performed better and how the creative could be improved. Precis is now considering how this could be incorporated into Alvie to provide even richer insights to customers. "The idea is to use Gemini to tag the elements of advertising creative and then run our existing models on top of that information," Bæk explains. "That would allow us to tell the client what made ad A better than ad B and make recommendations about the kind of ad to make next time."
Today we have just 15 developers on our team, yet we run hundreds of thousands of jobs a day. With Google Cloud, we can keep our team lean and there's no limit to how far we can grow the business.
Erica Petersson
Engineering Manager, Precis
For Petersson, the flexibility and scalability of Google Cloud have been key to Precis' success in building the Alvie platform, and will remain key as the company builds on those successes. "Today we have just 15 developers on our team, yet we run hundreds of thousands of jobs a day. With Google Cloud, we can keep our team lean and there's no limit to how far we can grow the business."
Founded in Stockholm in 2012, Precis is a modern, data-driven digital marketing agency specializing in media buying, analytics, creative, and proprietary technology. It delivers clarity, actionable insights, and measurable impact for ambitious brands in a complex marketing world.
Industry: Media & Entertainment
Location: Sweden
Products: BigQuery, Cloud Run, Gemini, Vertex AI Agent Builder
