An Executive's guide to generative AI

Artificial Intelligence, or AI, has quickly become the watchword of 21^st century business. The game changer that could transform virtually every aspect of our lives, not least the way businesses operate, develop, and grow. But amidst the hype, acronyms, and complexity, what exactly is AI and what could it mean for business operations?

In this first of four articles on generative AI's business potential, we'll explore the foundations of AI and its potential to transform the operational landscape. 

AI refers to the ability of computers to perform tasks that typically require human intelligence, such as recognising speech, identifying objects in images, and making decisions. AI systems can learn and improve over time by analysing large amounts of data and identifying patterns, enabling them to make predictions and recommendations. These AI systems can automate repetitive tasks, analyse complex data sets, and optimise business processes. 

Generative AI is a specific kind of AI that can create new and original images, music, or text from user prompts. 

In the generative AI space, Large Language Models (LLMs) have emerged as possibly the most exciting innovation for businesses since the development of the internet itself. 

LLMs are deep learning models trained on massive amounts of text data. The best-known is ChatGPT, but others are available, including Google's Bard, Bing AI, and many more. LLMs are programmed to generate new content that mimics human language in a chatty conversational style, hence the nickname ‘chatbot’. 

LLMs basically store a large amount of language knowledge in a model that can be easily referenced. The real power comes when you connect LLMs to other data sources, or ‘knowledge banks’, such as an ERP or CRM system, or files and folders across your company network. Once an LLM has access to your data, it can generate answers to your questions based on this data, so it can be original, specific, and meaningful to your business.

While generative AI systems could eventually replace many roles currently performed by humans, for the moment, it is likely to be a complimentary tool to support existing teams, with use cases needing to be created on a firm-by-firm basis.

For instance, firms that are market leaders could improve their performance even further if they can properly implement generative AI to support value-adding or differentiating areas of their business. Conversely, competitor firms could close the gap on market leaders if they can implement generative AI faster and more effectively, for example, to reduce time to market, increase margins or improve customer service.

There is also a clear opportunity to quickly automate repetitive tasks or gain new insights by using generative AI to quickly spot patterns across your organisation.

Although there are generative AI tools available now, none of the tech giants have yet fully integrated generative AI into their software and made the service easy to use and deploy for individual users. Once this happens, adoption is likely to follow very quickly.

Previous technology-driven ‘disruptors’, like computers, the internet, and mobile phones, tended to have initial high barriers to entry and a long adoption timeline. In contrast, generative AI will have low barriers to entry and a short adoption timeline.

Figure 1.1 – (left) Illustration of cost and adoption rate for previous technology disruptors, e.g. computers, internet, etc. and (right) the potential cost and adoption rate for Generative AI

Generative AI tools will have a cost, but it won’t be at the same high level as the first computers or mobile phones. And in today’s cloud-software ecosystems, AI solutions can be rolled out overnight into corporate Office365 and Google Workspace platforms.

This speed of deployment poses its own risks, with some organisations likely to be caught off-guard, due to inadequate internal controls, poor implementation, or lack of planning. 

To avoid these risks, and to maximise the business transformation opportunities that AI could deliver, we have identified four key areas for businesses to focus on when preparing for generative AI:

1. Data access. Companies must be able to connect and access previously unconnected data sources to create a proprietary knowledge bank. Doing so will enable valuable insights into future generative AI use.
2. Skills and training. Upskilling staff is essential to ensure they can make best use of these tools. Think about how you will talk to staff about generative AI and encourage them to start exploring generative AI websites now.
3. Risk management. Establish mitigation control frameworks to manage the risks around issues like IP loss, data accuracy and security, and ethical usage.
4. Planning. Planning how to develop, engage, and adopt generative AI at pace is crucial to stay ahead of the competition. 

Any advisor that tells you they have all the answers about generative AI probably isn't being completely honest. However, KPMG has a successful track record of helping businesses integrate technology into their operations that could aid in developing generative AI integration and adoption plans. Our Strategy Consulting, Connected Tech, and Digital Ninjas teams can provide expertise and support to help you exploit generative AI quickly, build out test cases, and enhance digital learning across your organisation. Ultimately, we believe in learning with our clients to provide customised solutions that meet the unique needs of you and your business.

To learn more on this topic, see our next article on the strategic implications of generative AI for businesses.