Richard Owen, Director of Digital Workplace Solutions, Unisys, on avoiding a ‘bring your own AI culture’ to work.
AI’s promised to be one of the most disruptive and transformative forces in the modern world. By the end of 2023, there were over 30,000 mentions of AI on earnings calls, compared with 500 mentions the previous year – showing its quickly increasing prominence.
Most industries want to integrate AI into their ways of working, with 55% of organisations using AI in at least one business unit, up from 20% in 2017.
However, mastering the deployment of AI and doing so safely and ethically is paramount.
Not all AI copilots are created equal. It can be challenging to determine which copilot is best for your organisation’s team, whether they be part of your IT, sales or marketing departments. For example, Microsoft takes pride in positioning M365 Copilot as ‘your copilot for work’ due to its compliance with AI data standards and ability to integrate across all your day-to-day applications in the Microsoft ecosystem, such as Outlook, Teams, PowerPoint and Word. Consequently, your teams can access AI assistance consistently as they weave through a busy day of tasks.
Unlike ChatGPT, which uses AI and machine learning technology to generate text-written responses to questions, Copilot is present throughout your working day—assisting with organizational tasks, prompting and analysing data and providing integration into Office applications.
Your organisation should build a wish list based on AI business objectives and strategies. Relevant considerations include: What is your end goal? Are there workplace inefficiencies that require correcting?
Your organisation must also be certain they can achieve this adoption. To avoid delay or setbacks to deployment, your team needs to be able to demonstrate the value and impact of your adoption.
To do this, you’ll need a strategy that outlines a personalised set of outcomes, showing how this surplus of time and workforce will be better used. The benefits from this reallocation must not be lost or miscommunicated as the message ascends through management, so keep these straightforward.
After deciding to invest, it is time to deploy your AI assistant to your users. First, who will get access? Will it be rolled out indiscriminately to the whole workforce? Or might you look for a specific employee persona who can benefit the most? A few prerequisites are also needed before your copilot can be fully integrated, including important administrative tasks such as assigning the appropriate base license. IT teams will need to make sure applications are updated across the entire team structure beforehand to support integration and avoid the risk of slowing down the end-user experience.
Firstly, AI governance needs to be up to date – establishing safe and ethical AI systems within your organisation should be a top priority for your team.
Guardrails and legal frameworks must be followed to ensure that a high level of diligence is introduced and maintained when using this technology.
Those related to data security are also high on the list of risks to watch out for. AI-generated information is accessed through the sharing of data repositories. Within these repositories, leaders need to ensure that data is being properly managed to get the most accurate results. To put it bluntly, ‘rubbish in, rubbish out’ means if you are feeding your AI legacy or inaccurate data, it will use this to form the basis of its generated work and responses.
Highly versatile document management solutions like SharePoint, which have great value as a data repository, have become a ‘dumping ground’ for all types of data due to a ‘never throw away’ mentality and poor records retention management. This creates a considerable risk to the accuracy of the results and response you would get from an assistant like Copilot, as this is typically one of the primary sources of information Copilot can draw from.
Lastly, not all AI is created equal, and organisations should be wary of choosing the wrong AI for the wrong person. This gives rise to the threat of Bring Your Own AI (BYOAI), where employees utilize commercial solutions instead of those authorized for the enterprise. While there are undeniable advantages to training your team to deploy your AI assistant to your users, it is time to proactively explore the world of AI and the various solutions available, including a positive impact on return on investment (ROI), using authorized products is deemed the best practice to avoid data breaches from less reliable tech and protect your company IP.
Once properly deployed and used effectively, your team will reap the full benefits of AI. This can look like AI-generated meeting notes by Teams, which can later be developed in Word, with the ability for users to dictate format and writing style based on previous documents. It could look like using an AI like M365 Copilot in Outlook to prioritise and identify specific tasks within hundreds of emails and generate draft responses to incoming emails. You can even analyse data to create projections and business plans that already outline pressure points and successes, e.g., in sales, where revenue is down and costs need to be cut back.
Many may feel discouraged when they realise an AI assistant won’t do all their work. Remember that with tools such as M365 Copilot, this is a copilot, not a pilot, and should be treated as such. It provides a jumping-off point, removal of writer’s block or a helping hand as a digital virtual assistant, so you do not start from scratch or take up valuable time on the mundane. The future of AI assistance is sure to see rapid development in the coming years.
If you follow the steps above, your organisation will be well-positioned to deploy AI assistants across all business areas.
