In our AI series so far, we’ve covered ChatGPT, automation, and the future of AI for employers. In our third and final article, we’ll look at how to use AI in HR data management. We'll explore challenges HR teams face when it comes to managing data, as well as how they can save valuable time here with the help of AI tools.

The future of HR and why AI is in the picture

So firstly, what’s the general picture of AI in the UK to date? The future is looking highly promising for businesses looking to invest in AI technology. The UK government hosted the first global AI summit back in June, and has developed a 10-year National AI Strategy in order to turn Britain into an 'AI superpower'.

Forbes found that the AI market in the UK is already worth more than £16.9 billion, and the number of AI companies based in the UK has grown by 688% in the last 10 years. 

The UK government has also committed £1bn in their 2023 budget for AI research. This means that there’s plenty of support and backing for companies not only looking to invest in AI, but also for creating their own AI tools that could solve many issues within their own industry – and of course – developing further their human resources data management solutions. 

So why should human resources teams be investing in AI? As found in Indeed & Glassdoor's Hiring and Workplace Trends Report 2023, the UK workforce is shrinking, leaving behind a 'post-pandemic participation gap'. This means that HR teams need to look at finding new ways of hiring and retaining staff in what is a competitive jobs market for recruiters. With the help of AI, they might be able to develop more compelling interview questions, or save time that would otherwise be spent on manual tasks like managing data.

In order to understand how to develop AI-based solutions to human resources data management, let’s look at what AI is capable of when it comes to data analysis, management, and streamlining of workflows.

Automation and AI can save HR teams valuable time

Automation and AI can help make it easier for your HR team to save time on repetitive manual tasks. However, could lead to being asked to do more with less — as Indeed’s leading AI expert Matt Burney pointed out in our instalment on ChatGPT and automation, recruiters are always finding new ways of streamlining tasks with automation. This can be done by finding new ways to manage unstructured data with the help of these tools.

Matt found, through an Indeed survey of Talent Acquisition leaders, that on average 14 hours per person per week was spent on manual tasks and processes. With language model GPT tools – that is, text creation and analysis tools – you could already save about 35% of time otherwise attributed to manual tasks.

HR teams are then free to work on tasks that require a more human, hands-on approach – dealing with staff face-to-face and addressing their concerns in a personable way. This might involve handling incidents of harassment and bullying, for example.

Next, we'll quickly move onto some other possible uses for AI in HR, before focusing on the main topic of how HR teams can use AI in order to better manage unstructured data.

Some other uses for AI in HR

As GrowthBusiness found, one of the biggest new trends in AI HR management will be performance management. This means being able to identify when staff require additional training, or support interventions – which could be identified far more quickly by AI than via managers manually identifying them. And here's another key point to take on board: in our guide to how employee development can boost company culture and retention, we found that 33% of UK employees said they would leave their jobs because of a lack of career progression. This makes performance management a top priority for any HR team.

GrowthBusiness also explain that AI can be used for ‘bread and butter’ HR management such as by creating a chatbot that’s able to answer staff questions about payroll or whether they’re owed any holiday time. 

One of the biggest pain points for HR is allocating time to managing unstructured data, because of just how difficult it is to manage. We’ll now look at this in our next section. 

What is unstructured data?

So what is unstructured data exactly? According to MondoDB, unstructured data is any data which is: ‘not arranged according to a preset data model or schema, and therefore cannot be stored in a traditional relational database or RDBMS’. This therefore includes any type of data that falls under text or multimedia, including formats like emails, videos, webpages, audio files, and photos. And MondoDB claims that 80% to 90% of data collected by organisations is unstructured. 

While much of the data that HR teams deal with is structured, like an employee or candidate’s name or date of birth for instance, other types of data, such as candidate interviews, emails, identification documentation, employee feedback survey responses, or PDFs will be unstructured. AI can create solutions to managing the above types of unstructured data, and how this could potentially free up time for HR teams. But firstly, let’s look at how to process unstructured data – which is somewhat more complex than structured data management.

Building AI tools to manage unstructured data

AI expert Matt Burney also expresses that it’s inadvisable to input sensitive company or customer data into ChatGPT type models (which is likely the predominant kind of data that your HR team will be working with), because it’s a data confidentiality issue. It might also run the risk of going against GDPR rules, but at the moment it’s somewhat of a legal grey area (as we discuss in our second article on the future of AI

Matt also explains in our ChatGPT and automation guide that one solution for companies is to develop AI tools to manage their own unstructured data without having to input data into AI tools owned by other companies – the data stays within the hands of the company, and doesn’t run the risk of being leaked. This makes AI a useful solution to HR recruiters who are dealing with plenty of unstructured data such as interviews. He says that you can use these tools to better organise interviews and learn about what did and didn't work from each one. Recruiters can also use this information to look at which behaviours make the most successful interviewee and why. 

How do you make unstructured data actionable?

Unstructured data processing involves taking data from formats like emails, videos, webpages and the like, and turning the data available in it into quantifiable data. HR teams planning on doing this will need the right data extraction tools to be able to extract unstructured data from its original file type. 

This then usually involves having to 'clean up' the data that you’ve extracted, before storing it in a management system which is easy and accessible for employees to use. Datasets extracted from unstructured data will usually need to be cleaned as they can often include errors like spelling mistakes or incorrect HTML tags. 

Managing unstructured data with AI

Once you have actionable unstructured data, you can then manage it more effectively with the help of AI tools. You might create or use a text analysis tool similar to that of ChatGPT, bearing in mind our points above about GDPR and handling sensitive data. You might use this to organise text by topic. For example, you might use AI in order to organise survey responses by specific phrases or words. This is particularly useful for HR teams which are dealing with qualitative rather than quantitative survey results, and need to better organise participant responses. 

As VentureBeat found, there are currently tools on the market which enable you not just to analyse text with AI, but to also use optical character recognition, voice recognition, and more. This means that you might be able to use AI in order to better organise sensitive security data on customers or employees used for identity verification. 

AI can help your HR teams focus on the more important issues

With the help of AI, HR teams can better manage unstructured data which is otherwise difficult to quantify. This includes interviews, videos, emails, or survey responses. As we’ve found, being able to streamline HR processes can help free up time on manual tasks, so that company teams have more availability to deal with responsibilities that need a more human approach.