AI Must Start with Trustworthy Data

By Paul Scott, Chief Technical Officer at Matrix Booking

The modern workplace is about to be transformed by artificial intelligence (AI). Already, AI is being used to automate repetitive tasks like data entry and scheduling, freeing up humans for activities that require complex decision-making, emotional intelligence and creative thinking.

Not only that but AI is set to make our office lives easier by spotting problems that slow us down and simplifying our work processes. Imagine it automatically reserving meeting rooms and allocating necessary resources based on real-time needs. It’ll be like having an all-knowing office assistant who prepares everything behind the scenes, allowing us to focus on the more critical tasks at hand.

Paul Scott CTO Matrix Booking
Paul Scott

However, this rosy future hinges on one crucial element: trustworthy data. AI algorithms act like complex sieves, sifting through vast amounts of information to identify patterns and make suggestions for improvements. But the quality of these recommendations is directly tied to the quality of the data they’re based on.

Improving the workplace with AI

AI in workspace management could fine-tune every area and support employees in meaningful ways. Imagine a system where a person allows it to learn their working patterns. It would be able to suggest the best resources for when and where they want to work. With this personalised approach, employees would have what they need, when they need it, reducing waste and saving time.

Using real-time data, AI could make immediate adjustments to the workspace. If a meeting room is booked but remains unused, the system can free it up for others. It could allocate desks based on who is in the office that day so that departmental team members could sit next to one another.

On top of that, AI could be integrated with environmental sensors and other building systems to regulate energy usage and make the workspace more comfortable. For example, the system could manage air conditioning or heating based on both planned and actual occupancy. Using this information, it would pre-cool or pre-heat specific areas of the building accordingly. As the working day unfolds, the system would then make real-time adjustments based on occupancy data to maintain a comfortable environment.

In the long term, AI could analyse data trends to forecast future needs. It might predict peak times for resource use and suggest adjustments to meet demand. This would help businesses plan better and stay one step ahead. The potential for AI to become a powerful tool to optimise the workspace and support growth is clear.

However, if any of the underlying data is flawed, the AI systems will end up taking the wrong actions. As the saying goes – ‘garbage in, garbage out’.[1]

The hidden danger of bad data

Collecting accurate data is more challenging than it seems. Decision-makers might think they are getting the right data, but issues like employees booking resources and not using them effectively can create false records. These inaccuracies can make the data seem reliable when it’s not.

Imagine employees book desks for the day, but then spend most of their time in meeting rooms. The resource management system, seeing the desk bookings, might assume full occupancy and not free up those desks for others who need them. This can lead to inefficiencies and frustration, as available resources are misallocated. A clearer picture could be formed with more nuanced data that combines planned desk bookings with actual occupancy sensors. The system could then make better decisions and optimise resource usage more effectively.

This is just one example of how poor data can turn a well-intended AI solution into a source of problems. Trust evaporates quickly when AI recommendations consistently miss the mark, which is why prioritising data quality is paramount. We need accurate, complete and unbiased data to help AI become a valuable partner, not a source of wasted resources, increased costs and disgruntled employees.

Getting the right data for AI

The collection of trustworthy data starts with encouraging the proper use of booking and check-in systems. Employers must provide regular training sessions to emphasise its importance and make sure everyone understands best practices. Making the booking and check-in process as easy as possible is key. On top of that, technologies such as sensors can passively collect anonymous usage data, complementing booking information and making it simple for employees to effortlessly engage with the system.

Businesses can also use additional technology to verify check-ins and presence in the office. If someone isn’t present, the system can free up their booked resources for others. This helps maintain accurate data and enables AI to optimise the workspace effectively.

To get the most accurate and complete data, businesses should adopt uniform data collection methods across the entire organisation. This simplifies data gathering and makes AI insights reliable and consistent. However, large-scale data collection comes with risks, so it’s important to limit access to authorised personnel to prevent accidental or malicious alterations, or unauthorised people getting hold of booking data that could reveal an individual’s booking data and their whereabouts. Protecting data confidentiality and integrity is fundamental to the trustworthiness of AI applications.

Furthermore, clear communication about how data is used and stored builds trust and encourages responsible data management. When employees understand the importance of data security and accuracy, they are more likely to handle data responsibly. The use of the data can also improve employees’ engagement with the system. Both of which support the overall reliability of AI systems.

The AI-powered future of the workplace

There is no room for debate, AI is set to change the way the modern workplace operates, making resource management and organisation much easier. This means resources are booked ahead of time, employee engagement is improved, environments are personalised and no space is wasted.

However, for AI-driven systems to operate well, they need trustworthy data. Without it, chaos would reign supreme.

[1] https://en.wikipedia.org/wiki/Garbage_in,_garbage_out