SGS DIGICOMPLY: Powering a one-stop knowledge base to keep food compliant and consumers safe
About SGS DIGICOMPLY
A subsidiary of SGS SA, the world’s largest inspection, verification, testing, and certification company, SGS DIGICOMPLY transforms risk and compliance information into user-friendly actionable knowledge. Pushing the boundaries of AI and data analytics, SGS DIGICOMPLY helps companies provide safer, sustainable, and compliant products to consumers.
Tell us your challenge. We're here to help.
Contact usWith Google Cloud, SGS DIGICOMPLY has built an AI-driven, one-stop compliance intelligence platform that lets food producers act on safety information quickly and decisively.
Google Cloud results
- Helps food producers monitor compliance issues worldwide and respond effectively if needed, by supporting SGS DIGICOMPLY’s AI-driven platform
- Reduces costs and simplifies infrastructure management, with preemptible instances on Google Kubernetes Engine
- Supports a comprehensive COVID-19 knowledge base to help researchers keep abreast of 44,000+ scholarly articles and counting
Monitors 100K+ web pages, 150+ markets, 90+ languages
In today’s highly globalized world, there’s more to food production lines and supply chains than meets the eye. A food producer in Italy may source some of its ingredients from Africa, others from Asia, then export its goods to customers in North America. Yet, in each of those countries, the product standards, hygiene rules, labeling requirements and more, may all be vastly different. So how do you stay up-to-date with changing regulations across all the territories you operate in? And how do you make sure your food meets the standards of all your customers, wherever they are?
It’s a complicated business, but that’s where SGS DIGICOMPLY comes in, with a powerful compliance intelligence platform that enables food producers, as well as household goods manufacturers, to minimize any potential safety risks for their customers.
“There’s an avalanche of information that you need to deal with in this industry, especially when it comes to ensuring consumer protection and safety,” says Nicola Colombo, CEO at SGS DIGICOMPLY. “Our goal is to bring all of that together into a product that’s easy to digest and respond to. Something that enables industry professionals to keep delivering safe, sustainable, and compliant products that consumers can trust.”
“We wanted to transform unstructured information into actionable data using AI. We were impressed not only with the AI capabilities of Google Cloud, but with its flexible pricing and third-party integrations. We have very specific needs, and Google Cloud lets us work the way we want to work.”
—Nicola Colombo, CEO, SGS DIGICOMPLYWhat makes SGS DIGICOMPLY so compelling is its innovative use of machine learning and artificial intelligence to find new compliance information from non-traditional data sources, such as government regulations published as a PDF image or a social media post about a bad batch of yogurt. It harnesses the power of analytics to quickly and effectively structure this information and provide its clients with the most important data that they can act on quickly and effectively.
The platform currently monitors more than 100,000 web pages to identify thousands of documents and articles, commentaries, images, and labels for its constantly evolving compliance knowledge base. From here, users can select the exact kinds of information needed for a specific product and receive unique, AI-led, and customizable intelligence.
To deliver a platform with such far-reaching content, accessibility, searchability, and data management capabilities, SGS DIGICOMPLY runs its infrastructure on Google Cloud.
“We wanted to transform unstructured information into actionable data using AI,” says Nicola. “We were impressed not only with the AI capabilities of Google Cloud, but with its flexible pricing and third-party integrations. We have very specific needs, and Google Cloud lets us work the way we want to work.”
It was important for the SGS DIGICOMPLY team to find a provider that was evolving in the same direction, as Jakub Janostik, Head of Development at SGS DIGICOMPLY explains: “We feel that Google Cloud aligns with us, both from a DevOps perspective, as well as its evolving AI stack. We’ve been given access to AI features in beta, which is a way of supporting us to keep innovating,” he says. “The Google Cloud team also helps us identify areas where we can save costs, so we can put those savings back into our work.“
“Google Kubernetes Engine has a lot of features that makes handling Kubernetes instances very easy for us. Using preemptible nodes, for example, means that we can reduce our infrastructure spend dramatically when we’re setting up short-lived environments for testing.”
—Jakub Janostik, Head of Development, SGS DIGICOMPLYBuilding a unique infrastructure on a mature Kubernetes solution
SGS DIGICOMPLY needed a very particular kind of infrastructure built around Kubernetes containers. After evaluating its options, the company selected Google Cloud, initially for the maturity of its Kubernetes management and its ease of use.
SGS DIGICOMPLY’s technology platform comprises three main groups. The core infrastructure is made up of Compute Engine virtual machines, Container Registry to administer the Kubernetes instances, and Google Cloud operations tools for monitoring and logging.
For its proprietary machine learning solution, SGS DIGICOMPLY runs Compute Engine instances with Cloud GPUs, high-performance hardware to maximize power and minimize the time spent running intensive workloads. As well as its own tools, the company works with Document AI, which uses a number of different language-processing and image tools to extract insight from raw data. AutoML and its suite of custom machine learning models have been key. AutoML Vision has helped SGS DIGICOMPLY to incorporate data from images of physical objects such as labels on a jar, while AutoML Natural Language allows the company to find the most relevant parts of a text quickly and easily.
Finally, the data handling component of the infrastructure is orchestrated by Google Kubernetes Engine (GKE). The objective here is to extract, transform, and load the data into something that can be used by the end user. Using GKE gives SGS DIGICOMPLY the flexibility of incorporating proprietary and third-party components into its architecture, while keeping its management overheads low.
Due to one of the tasks in its information extraction workflow, for example, SGS DIGICOMPLY needs to process a lot of data in a short time. By using horizontal autoscaling on Google Kubernetes Engine and Google Cloud, it can complete this task in just a couple of days, instead of several months. This enables it to deliver insights to its clients faster, while also minimizing time needed for DevOps activities since the platform is almost fully automated and all deployments are scripted. Because Google Kubernetes Engine runs health checks and auto repairs nodes if necessary, SGS DIGICOMPLY can run the information extraction task on preemptible nodes, reducing costs by more than half.
“Google Kubernetes Engine has a lot of features that makes handling Kubernetes instances very easy for us,” says Jakub. “Using preemptible nodes, for example, means that we can reduce our infrastructure spend dramatically when we’re setting up short-lived environments for testing.”
“We collate somewhere between several thousand to tens of thousands of new documents weekly. What’s difficult isn’t the scale but the complexity of the data. Google Cloud helps us to quickly sift out the most important information, so that our clients have the information they need when they need it.”
—Nicola Colombo, CEO, SGS DIGICOMPLYTurning insight into action
With Google Cloud, SGS DIGICOMPLY has built a one-stop platform for compliance intelligence for the food and agriculture industries. Harnessing the power of AI, the platform is known for “horizon scanning,” where it keeps track of thousands of compliance events every day, from regulation changes to news of outbreaks to mislabeling in local markets.
“Our platform covers more than 230 markets using thousands of data sources in more than 100 languages every day. We collate somewhere between several thousand to tens of thousands of new documents weekly,” says Nicola. “What’s difficult isn’t the scale but the complexity of the data. Google Cloud helps us to quickly sift out the most important information, so that our clients have the information they need when they need it.”
More than a quarter of SGS DIGICOMPLY sources are monitored automatically on Google Cloud, along with more than half of its markets. Within that, the company records 252 million processed sentences from one million documents.
With a constantly updated feed of comprehensive compliance intelligence, SGS DIGICOMPLY helps food producers act quickly and decisively to ensure the safety of their customers wherever they might be in the world. Over time, the company has expanded the range of its sources and added cosmetic and household goods manufacturers to its list of clients.
Even as it expands, SGS DIGICOMPLY is already looking at the next evolutionary steps of its platform. The company has its own proprietary tools for data extraction and analytics and is looking at using Google Cloud AI to double down on its capabilities in handling text and translations. It’s currently taking advantage of Google Cloud AutoML Translation to help translate legal documents from Asia as part of a new initiative. It’s also delving deeper into Document AI to create a “Document Explorer” feature in its platform, which will perform tasks such as automatically identifying and flagging the most relevant paragraphs in a regulatory filing.
Launching an AI platform for teams on the COVID-19 research front line
In the early months of 2020, as much of the world went into lockdown due to the COVID-19 pandemic, SGS DIGICOMPLY saw a way to use its technology to support scientists on the research front line.
Its team of data scientists and developers responded quickly to this new challenge, and in just one month released COVID-19 Intel, a free platform for researchers working on the response to the virus. SGS DIGICOMPLY was able to build the new platform quickly thanks to its effective preexisting pipeline, says Nicola. First, the team imported data related to COVID-19, set up a pipeline to read the data from BigQuery datasets, then trained and set up a quick service using Cloud Run. A powerful content management platform, COVID-19 Intel uses natural language processing and other AI technologies to help the research community keep up with the new COVID-19 research developments. On its initial release, COVID-19 Intel saw peaks of up to 500 users who wanted to better understand the progression of the virus.
SGS DIGICOMPLY repurposed its horizon scanning techniques to look for information from the raft of new studies being commissioned about COVID-19 and its effects. At the time of launch, COVID-19 Intel was drawing from a dataset of more than 44,000 scholarly articles which are updated on a daily basis. “There’s a large and growing volume of research about COVID-19, and it can be difficult for scientists to stay on top of it,” says Nicola. “We want to use natural language processing and other AI capabilities that we’ve developed with Google Cloud to help the medical research community generate new insights. We’re proud to be doing our part to support researchers and their inspiring work in combating the virus."
Tell us your challenge. We're here to help.
Contact us
About SGS DIGICOMPLY
A subsidiary of SGS SA, the world’s largest inspection, verification, testing, and certification company, SGS DIGICOMPLY transforms risk and compliance information into user-friendly actionable knowledge. Pushing the boundaries of AI and data analytics, SGS DIGICOMPLY helps companies provide safer, sustainable, and compliant products to consumers.