What are the Gen-AI capabilities your business needs this summer?

Explore the essential Gen-AI capabilities your business needs this summer, from tool investigation to data architecture and cyber security. Stay ahead of the curve with Grant McGregor's expert assistance.

Whether or not your organisation is already exploring the potential of generative AI, it’s likely that you will need to develop associated skills fast – or risk getting left behind or missing out on potential competitive advantage. We take a look at some of the most useful skills to get the most from gen-AI tools.

In a recent blog, we looked at which AI skills organisations need most . The surprising results showed that competence in a few specific tools is being sought after. However, as the use of Large Language Model (LLM) tools, like ChatGPT and Microsoft Copilot, becomes more widespread, there are a number of more generalised, transferable skillsets which every organisation will need to develop. 

#1. Investigation and exploration of tools and their capabilities

The starting point for most organisations will be deciding which tools are going to be most relevant for your people, processes and the customers you serve. This requires time and resources to be allocated on keeping up to date with new product launches, reviews and use cases as well as the time and resources to test out the tools that have most relevance to your operations. 

This doesn’t necessarily require technical skills. In many ways, it helps if these early explorations are employee-led, since they will have most hands-on understanding of how the tools can benefit your day-to-day operations. 

It does need enthusiasm and motivation to explore the tools, a willingness to experiment and some creativity when thinking about doing things differently and how AI could be used. Look for champions in your business who can lead on these discovery phases and be sure to have them report regularly to the rest of their team and the business as a whole. Seek feedback and engagement from the whole team.

#2. Prompt engineering

Your ability to construct effective prompts has a huge impact on the success of your use of any generative AI tool. This may seem counter-intuitive, especially when people talk about “natural language queries”. But asking something of a gen-AI tool is very different to asking something of a standard web browser. How you prompt the tool to respond will greatly affect the results you get back. Poor prompting can lead to problems of hallucinations and bias.

Even before you start using gen-AI tools in anger, it is a good idea to start building up a library of best practice, suggested prompts and user experiences to share throughout your organisation. Setting up a dedicated Teams channel where everyone can post their experiences, share ideas and access best practice tips and training resources is a good way to raise awareness, boost engagement and make it easy to share your successes.

#3. Governance and compliance 

The use of AI tools in an organisation presents many issues around governance and compliance which need to be addressed. It will require close working between your technical teams and your governance and risk managers. The implications of using gen-AI range from issues around data and information security; transparency about how the tools are used, especially in how they impact decision making; bias; the potential for error or hallucination and how this is identified. 

Of course, these risks will differ according to the gen-AI tools you use, the data which is exposed to them and how they are used. Identifying these risks and communicating them clearly requires a good understanding of the tools themselves. Public models, like Chat-GPT, shouldn’t be used with sensitive business data. There is a risk of that data being used to train models and ending up in the public domain. Microsoft Copilot, on the other hand, stops at your tenant, so your information security concerns are around ensuring that access privileges are carried across to your Copilot deployment – so that only those who have a right to access certain information will have their searches exposed to that information.

If your teams aren’t confident they can answer these questions, it is important to reach out to IT partners who can help you understand the risks and limitations before you start using the tools. Having the right governance in place isn’t something you can put off until later. 

#4. Data architecture and optimisation

Data architecture and analytics capabilities in the UK are not keeping up with demand. It is already difficult to recruit to these specialist roles. It is likely that, at least in the short term, the demands of gen-AI will exacerbate this situation. If you are developing your own in-house gen-AI models, perhaps based on Microsoft’s Copilot Studio or Azure OpenAI, then data architecture will be critical to your project’s success. The old adage of “rubbish in, rubbish out” is just as true for gen-AI as it is for any other software. 

Even with tools such as Microsoft Copilot for Microsoft 365 which offers generative-AI capabilities over the top of your everyday office and productivity tools like Teams, Word, OneDrive and Outlook, you are likely to require some adjustments to the way information is stored and accessed in your business. For small and medium-sized businesses, it is often better to work with an experienced partner or to bring in consultants on a project basis. This offers a more agile and cost-effective way to get the data architecture right before launching your use of gen-AI tools.

#5. Cyber security

As with any IT or data project, you need to understand the cyber-security implications of your technology rollout. Implications differ according to which tool you use, who is using it and how. One recent study showed how generative-AI systems may be vulnerable to new types of cyberattacks and malware, specifically a “zero-click” worm that does not require a person to make the mistake of clicking a suspicious link or file; rather, the malware is processed and propagated automatically by the AI system.

If you are making changes or introducing new tools and you aren’t sure of the cyber-security implications, it is always a good idea to work with your IT partner to explore and be clear about the risks so that they can be addressed and mitigated in advance. 

What now?

Is your business ready to explore the new possibilities created by gen AI? Would you like support putting the necessary governance, IT and data infrastructure and AI tools in place?

The Grant McGregor team can assist.

Call us: 0808 164 4142

Message us: https://www.grantmcgregor.co.uk/contact-us

For more information on What are the Gen-AI capabilities your business needs this summer? talk to Grant McGregor Ltd