AI is the biggest topic of conversation these days. Even though AI comes with especially useful applications and some very practical downsides, it’s important to know how to frame the conversation with CIOs or CEOs before a company considers investing in it.

By Eric Knipp, vice-president of solutions engineering at Cisco

Here are four things to think about when weighing the conversation possibilities:

* Determine your AI strategy – While many companies might be asking what the AI strategy is in order to better inform employees, customers, partners, etc., the question they should be asking is what existing functions within the company can be improved by AI. The right question differentiates how AI complements existing functions versus how AI serves as an overlay strategy. For example, “what is the customer service strategy enabled through AI?” vs “what is the R&D strategy enabled by AI?”

* Know that employees are already using AI – This technology has taken the world by storm and employees are naturally curious about trying AI using programmes such as ChatGPT. This is why companies need to be very intentional about the policies they put in place to protect their IP internally. Instead of telling employees not to use public AI services, consider partnering with a provider to develop internal AI capabilities that leverage ecosystem partners and the company’s own internal platform capabilities. While some companies have chosen to block employees from using AI services on their corporate network all together, if it is impossible to restrict usage, at least give employees an AI alternative. As with security, when restrictions make it hard to find a good tool, employees may rely on shadow IT to download the tools they want. Give them an AI alternative that is backed up by internal policies and InfoSec protections. For example, Cisco’s recently launched Motific provides a central view across the entire GenAI journey to help central IT and security teams deliver trustworthy GenAI capabilities securely and responsibly across their organisation.

* Demonstrate how AI ties to your core business – Let’s face it, AI is expensive and requires a greater technology investment. For example, Reuters estimates that a ChatGPT search costs 10x more than a standard Google or Microsoft search. In this economy, companies are protecting their core business right now and not taking on projects outside their IT budgets. If it can be shown that AI is going to help the core business generate revenue, fuel growth, reduce risk, reduce costs, or better utilise resources, then it will be budgeted for. If it can’t, it will probably be a no-go.

* Be creative with AI training and hiring – We’re experiencing a massive skills and resourcing gap in the IT landscape – there are approximately 10 jobs available for every skilled professional just in the security space. Because AI is a brand-new skill set, many industry companies are looking to university students or “early in career” candidates to fill these positions. However, AI can be taught to, and learned by, people who don’t necessarily have a university degree. With available training and programmes new skill sets can be learned and these skills can be leveraged within in organisations.

While there are seemingly hundreds of decisions to make about AI, the wisest course of action is to start slowly. Start by assessing what AI tools can and can’t be used by employees within the organisation. Then determine which, if any, internal AI projects can be controlled to protect IP.

If a business can’t develop or buy AI technology, consider partnering with a company who can provide it. Ensure that the right employees with relevant skillsets can understand and manage it, then see how AI can be applied to improve both the employee experience, customer experience, and the company’s overall profile.

Since 2015, Cisco has been a strategic partner with NVIDIA, the inventor of the GPU that fuelled the AI boom. We are combining our expertise in Ethernet networking and our extensive partner ecosystem with NVIDIA’s GPU technology to help customers navigate AI transitions with highly secure Ethernet-based infrastructure.

Cisco already has Nvidia’s GPUs inside our Collaboration portfolio, now we’re expanding that capability into the data centre. Recently Cisco and Nvidia announced an enterprise AI workload solution – we’re putting Nvidia’s GPUs into our UCS servers and wrapping them with Cisco Validated Designs (CVDs) to make it easier for enterprises to deploy and manage their own secure AI infrastructure.

Essentially, we’re creating AI superpowered networks.