• Written by Muhammed Abdulla NC, CEO

A hiring manager checks the employee database to answer a question from a line manager about staffing levels, and wonders how the new enterprise AI system for human resources will affect this routine task. The system, built by Harns Technologies, is designed to answer questions about the organisation's own data in natural language and provide answers in real time. Before deployment, several questions need to be settled in procurement, in writing, to ensure the security and privacy of human resources data.

Processing and storage locations

The first question to settle is where the data will be processed and stored, and whether the system will train any models on this data. This is crucial because human resources data is sensitive and subject to various regulations. The organisation needs to ensure that the data is handled in compliance with these regulations, and that the system does not introduce any new risks. For example, if the system will be processing data on employee salaries, the organisation needs to ensure that this data is stored securely and accessed only by authorised personnel.

The location of the processing and storage also matters, as different countries have different regulations regarding data protection. The organisation needs to ensure that the system complies with all relevant regulations, including the General Data Protection Regulation (GDPR) in the European Union. This may involve working with the vendor to ensure that the system is configured to meet these requirements.

Prompt injection and data boundaries

Prompt injection is a potential security risk that arises when an attacker is able to inject malicious prompts into the system. This can be used to extract sensitive data or to manipulate the system into performing unintended actions. In the context of human resources, prompt injection could be used to extract sensitive employee data or to manipulate the system into providing unauthorised access to this data. To mitigate this risk, the organisation needs to ensure that the system has robust data boundaries in place, and that all prompts are thoroughly validated before they are processed.

This requires careful consideration of the system's architecture and configuration, as well as regular testing and monitoring to ensure that the system is secure. The organisation should also work with the vendor to ensure that the system is designed with security in mind, and that any vulnerabilities are promptly addressed.

Human resources data dependencies

The enterprise AI system for human resources depends on several types of data, including the policy library, headcount and skills, attrition history, and hiring plan. The policy library provides the system with the rules and regulations that govern human resources practices, while the headcount and skills data provides information on the organisation's current workforce. The attrition history and hiring plan data provide insights into the organisation's staffing needs and trends.

For example, a category manager might use the system to compare the skills of different employees, or to identify trends in attrition rates. The system can provide valuable insights and answers to questions about human resources, but it requires high-quality data to do so. The organisation needs to ensure that this data is accurate, up-to-date, and properly configured for the system.

More information on human resources data and how it is used in the system can be found on our human resource page.

Impact on policy questions and planning

The enterprise AI system for human resources can have a significant impact on the policy questions that arise every week, as well as on planning that runs on impressions. By providing accurate and timely answers to questions about human resources, the system can help to inform decision-making and reduce the risk of errors or misunderstandings. For example, a shift supervisor might use the system to answer a question about employee scheduling, or to identify trends in employee productivity.

The system can also help to improve planning by providing insights into staffing needs and trends. By analysing data on attrition rates, hiring plans, and employee skills, the system can help the organisation to identify areas where it needs to improve its human resources practices. This can lead to better decision-making and more effective use of resources.

Measurement and evaluation

To evaluate the effectiveness of the enterprise AI system for human resources, the organisation can use several metrics, including repeat question volume, time to answer, and attrition by tenure. Repeat question volume can help to identify areas where the system is providing value, while time to answer can help to evaluate the system's performance. Attrition by tenure can provide insights into the organisation's staffing needs and trends.

By tracking these metrics, the organisation can gain a better understanding of how the system is being used, and where it can be improved. This can help to inform decision-making and ensure that the system is providing the best possible support for human resources practices.

Limitations and wrong choices

The enterprise AI system for human resources is not a solution for every problem, and there are certain situations where it may not be the best choice. For example, the system may not be suitable for organisations with very small or very large workforces, or for organisations with highly complex human resources practices. In these cases, the organisation may need to consider alternative solutions, such as custom-built systems or manual processes.

If you have any questions about the enterprise AI system for human resources, or would like to learn more about how it can support your organisation's human resources practices, please contact us for more information.