It’s automatic: digitising food safety compliance

Australia maintains an enviable reputation as a world-class food producer. We’re renowned for our safe, high quality, sustainable and ethical produce. There is a stringent regulatory system in place to verify that Australian produce meets this high standard.

Regulatory compliance in food supply chains is clearly important. But, like in many other industries, it’s onerous on business and regulators. Compliance is an administrative challenge, involving many systems, reporting and players across the supply chain.

There’s a huge opportunity to better coordinate, automate and streamline compliance across the supply chain through new digital technologies. Doing so would result in hundreds of millions of dollars in estimated savings for Australia.

There’s an opportunity to streamline Australia’s regulatory compliance system that ensures produce is safe and high-quality.

Automating food safety compliance

Safe Food Production Queensland (Safe Food) regulates primary production and food processing in the second largest state in the country: from meat and dairy to eggs, seafood and horticulture (seed sprouts).

Recently, Safe Food has shifted towards digitising its food safety compliance monitoring. This business need became pressing during the COVID-19 pandemic when there was a need to reduce in-person visits.

Since 2020, Safe Food has transitioned an iOS application into a web-based portal. Poultry processing businesses now add their compliance data into this portal. But it’s critically important the data that companies provide is objective and trustworthy. Safe Food requires a solution for this verification step.

Mark Chan is the Safe Food manager for risk and response. Safe Food has staff who are trained to go into the physical site and conduct checks onsite to verify the data provided.

“If we could add a layer of technology that allows us to trust the data, we could reduce reliance on manual audits.”

That’s where a new CSIRO-Safe Food project comes in.

Through the Trusted Agrifood Exports Mission, CSIRO is working with Safe Food to complete the missing piece and to ensure regulators get the trusted data they want.

Industry trials of ‘automated compliance’ kick off in 2022

CSIRO’s research team is developing a cloud-based ‘data trust engine’ in consultation with industry. In 2022, they will begin trials of a prototype, starting with poultry processors.

Poultry processing plants are ideal first candidates for these trials due to the vast amount of data involved in the compliance process. Processing plants typically require highly mechanised, fast running, high throughput and largely automated operations. Given this, many plants are early adopters of new sensors and methods.

Sensors on, or around, a poultry processing facility continuously, autonomously, and objectively monitor critical events. Trust in that data means it can be used to its full potential.

As an example, poultry food processors may use sensors to measure the temperature to ensure the product is kept at a safe temperature during certain processes. The next step in digitising and automating compliance checks is to add a level of surveillance. This would verify that the temperature reading supplied is accurate.

This data would then go into CSIRO’s data trust engine. In the engine, it will be integrated with data from different sources to attribute a trustworthy metric. This metric will indicate that the data is valid, has been correctly collected, can’t be tampered with, and importantly, it can be relied upon by industry, auditors and regulators without the need to physically inspect (audit) it themselves.

The data can then be used to automatically assess compliance against regulatory requirements and guidelines.

Dr Philip Valencia is the tech lead from CSIRO’s Data61.

“Not only can this data be used to auto-populate compliance forms, but receivers will be able to know why the data can be trusted,” Dr Valencia said.

“It will help coordinate compliance processes for industry who have, in the past, experienced a lot of duplication across audits.”

The poultry industry will be the first to trial our ‘data trust engine’ to automate regulatory compliance.

Deep analytics to improve operations

Businesses are already collecting and providing this kind of data to auditors manually for compliance purposes. But by going digital, businesses can gain additional deeper data analytics and insights on their operations.

“An audit only provides a snapshot of the day. So the idea with automated compliance is that we can get data over time and look at trends, such as temperature dips and their frequency of occurrence,” Mr Chan said.

Loss of product and time due to a problem, such as a contaminated or ‘out of specification’ product, costs a lot of money. But looking at things holistically through a continuous data loop would allow issues to be addressed in real time.

Rick Jacobson is the Safe Food director of compliance and verification

According to Mr Jacobson, this speed of insights would be the biggest benefit for operators.

“When you’re capturing data, you are joining up the end-to-end process, to show how the system of food safety controls is working,” Mr Jacobson said.

“Operators would be in a better position to respond and correct the problem before there’s any product wastage, or impingement to product shelf life or public health.

“The end goal would be to have sensors that talk to each other and data that is uploaded in real time.”

How can we enable industry adoption?

The project is on an ambitious timeline. The goal is to roll-out a product to industry in 2023.

“Safe Food is a critical partner for rolling out our data trust engine technology to industry,” said Dr Valencia.

“We’re leveraging their strong relationships with industry to coordinate trials with more agrifood sectors, including red meat, later this year.”

So what’s needed to enable widespread industry adoption?

“The compliance system needs to be easy to use and cost-effective to be used across the board,” Mr Chan said.

Mr Jacobson added that Safe Food has had success with data sharing to date because they don’t have a ‘one-size-fits-all’ approach.

“There are different levels of maturity across the industry around how the data is shared, some requires a manual input,” Mr Jacobson said.

“So we have different solutions for different businesses and can receive data securely via multiple methods through our web-based portal.”

This will continue to be critical, as it will be up to companies to have the sensors and technology to be able to supply data to Safe Food. This means their system will need to integrate with a range of different sensors.

But regulators are only one part of the system. Customer certification programs and exporters also need access to similar data.

The big picture for data and exports

Mr Jacobson said that the ultimate solution could be in creating a ‘data lake’ to streamline data collection.

A data lake is a repository where data is stored.

“We talk about having a ‘data lake’ where all the data collected for export purposes, large retail supermarkets, food safety and quality goes into the one platform,” Mr Jacobson said.

That doesn’t mean everyone can access all the data. Rather, in order to protect company privacy, data would be carefully managed and shared according to user specifications.

It aims to underpin the Trusted Agrifood Exports Mission’s ‘tell us once’ approach to regulatory compliance. This approach will reduce paperwork and build on Australia’s reputation as a trusted source of high-quality foods.

Through better coordination and enhanced trust, ultimately Australia will be able to grow the value of its agrifood exports.