Explore how Sage upholds the highest standards in data and AI ethics, ensuring your data's integrity and security. Our commitment is rooted in transparency and trust, safeguarding you and your business with ethical diligence.
Sage data and AI ethics principles
In a constantly changing world, we are committed to integrity, transparency, and trust. Our data principles may adapt to stay aligned with our customers' needs, the evolving business landscape, and regulatory requirements.
Your product data - This refers to the data about your business and employees that we process for you in our Accounting, HR and Payroll products and other services. These principles are mainly concerned with how we use your product data.
Other data - As well as your product data, we also use other data, sometimes alongside your product data, to get more powerful insights and build more powerful products, features and experiences. This includes data that we collect about how our products are used and data that we obtain from third-party sources.
Anonymisation and anonymised data - Anonymisation means aggregating data, removing whole data records, or masking specific fields, so that it is not possible to tell which person or business the data relates to.
Identifiable data - This is data that has not been anonymised, so that it is possible to tell which person or business the data relates to.
Sage’s trusted network - This refers to a collection of products and features that help our customers digitise their business processes by connecting with their own customers, suppliers, accountants, banks, tax authorities and employees.
Artificial Intelligence or AI - This is the simulation of human intelligence by computer systems to enable them to solve problems in sophisticated ways. It is often “trained” to solve problems by using data.
Some of our products, features and experiences are built using your product data or other data or involve sharing your data across our trusted network. When we do this we are transparent about how we use your data and the choices you have.
We do not share your identifiable data with third parties unless you give us permission first, either explicitly or by choosing to use a product or feature that requires sharing. We make it clear when sharing is required and share only the minimum data necessary.
When using your anonymised product data to build products, features or experiences, or as part of our Using Data for Good work, we do not offer a choice on whether your anonymised data is used.
Our Employment Verification service enables sharing of employees’ pay information with lenders as part of a credit application. This is identifiable data about employees, so Sage Payroll customers can opt out of this service for their whole company or for individual employees. Also, lenders must get permission from the employee before requesting the data from us.
We use your product data and other data to develop innovative and powerful new products, features and experiences to help your business and your employees thrive.
We believe data has enormous potential to help you and your employees thrive.
Where it is allowed under applicable laws, we analyse your product data and/or other data, together with data from other Sage customers, to enhance your experience with Sage, simplify your business processes and solve common problems and to develop and train AI models to create new products, features and experiences.
When doing this, we may give carefully selected Sage employees temporary access to your product data as part of a sample, with data from other customers. If this happens it is done with strict controls to ensure security and protect your privacy.
We developed advanced machine learning models, trained using invoice data from thousands of customers, to automate accounts payable processes by extracting invoice data from emails and inserting it into accounting systems without the need for any manual data entry. Find out more about Accounts Payable Automation.
Case study 2
We use data that we collect about how our products are used to understand patterns of behaviour and to help identify areas for improvement that make our products easier to use.
Charging for access to data
We may charge for access to products and features built using your product data.
We build products and features that provide insights or predictions derived from your product data, together with data from other customers and other data. We also build products and features that allow you to share your identifiable product data with third parties.
We may charge for use of these products and features. If we do charge, we are transparent and open about the charges.
With our Employment Verification service, employees of our payroll product customers can give permission to potential lenders to access their recent pay data as part of the affordability check in online credit applications. This service is free to payroll customers and employees, but lenders are charged for accessing the service.
Using data for good
We use your product data to help find and fix common problems facing small and medium sized businesses and wider society.
Our purpose is to knock down barriers so everyone can thrive. We analyse your product data together with data from other customers and other data, to help us better understand the barriers and problems SMBs are facing and to identify solutions. We publish these insights or use them to help us be a more effective public voice, speaking up for all SMBs’ interests.
We sometimes work together with governments or trusted partners to do this analysis and may share your anonymised product data with them. We assess the security approach of partners before agreeing to share data.
We make sure data quality is as good as possible so you can trust the insights and recommendations that flow from it.
We understand that the value of insights and recommendations that we make from data depends on that data being accurate, relevant, complete and up to date. This includes data that we control in our websites and internal systems, and also data in our products and services that you control as a customer.
We have put in place teams and processes internally to look after data quality.
We build features into our products, services, websites and internal systems that help detect and remove data errors and improve the quality of your data.
When we build products and features that involve data sharing, we include features that help data recipients have more trust in the accuracy of the data they receive and where it came from.
On our website, when customers are entering company details, we use technology that predicts which company they are referring to and then automatically completes the relevant business information, based on data that we have collected. This reduces data entry for customers and also ensures that the data is complete and of good quality.
Diversity and bias management
We work to manage bias in our AI models to help make sure that everyone affected by our AI models is treated fairly and without discrimination.
We look carefully at the datasets we use to manage the risk of bias in our AI models.
We design our AI models with the management and prevention of discrimination and stereotyping in mind.
We build AI models using diverse teams from a wide range of cultural, academic and professional backgrounds.
Our AI and Data Ethics Council, which will help us make ethical decisions on how we use data, will include members who are specifically focused on diversity and bias management.
We are partnering with HBCU Morehouse College in Atlanta, Georgia to foster a more diverse pipeline of talent for technology jobs, focusing especially on AI. Internationally, we have a multi-year partnership with Teens in AI, an initiative launched at the AI for Global Good Summit at the UN in 2018, which aims to inspire the next generation of ethical AI researchers, entrepreneurs and leaders, with a particular emphasis on diversity.
We will always empower you to make decisions on how AI in our products affects your business.
We ensure that you can understand and control how AI in our products affects your business.
We build features that allow you to decide how much automation you want to allow or how much you want to review recommendations before they are applied.
Where possible, we make it simple for you to understand the recommendations made by AI.
You can override recommendations made by AI. When you do this, we use your feedback for improvement, so your trust in the recommendations increases over time.
Our General Ledger Outlier Detection feature uses machine learning to detect and flag potential errors in transactions. Customers are able to see why each transaction was flagged and either correct the transaction, or accept it as is.
Need guidance? We are here to help
If you have questions or need clarity, get in touch with our team of experts who can help you find the answers you need.