As businesses look to drive costs down and value up, technology is a powerful enabling lever. There has been a focus on modern cloud-based infrastructure, automation and digitisation for some time already – but what of the buzz technology of the moment, AI, and in particular the new poster child generative AI (such as ChatGPT)?

AI brings a powerful cost lens

On the cost-out side, ‘traditional’ AI (if we can call it that) can be applied in some powerful ways. For example, in Finance where AI-enabled tools can scan invoices and contracts, look for overpayments, duplicate payments and amounts not billed in line with contract terms.

It also has a big role to play across Procurement and the supply chain in terms of running analyses and comparisons of pricing, discounts, and service and performance levels.

GenAI for value generation and productivity

When it comes to genAI, the applications are more on the value side – performing tasks such as content generation, information extraction and summarisation, language translation, code generation, and powering smart chatbots or virtual assistants.

GenAI drives speed, productivity and quality of output. It can save human time on laborious tasks (or those that would simply take too long) and free people up for more value-adding work – strategic and creative thinking, innovation, speaking to customers and addressing their needs.

The four horizons for GenAI

However, most businesses are still in the early stages of their genAI journey. We see four ‘horizons’ in terms of its adoption and deployment – and most organisations are still in the first and second of these:

Horizon 0: Procrastination - Not really understanding what the technology can do, knowing that they should be doing something, but not knowing where to start

Horizon 1: Colleague support – tech savvy employees play with prompts to enhance elements of their job roles and routine tasks e.g. research. Businesses undertake ring-fenced proof of concepts (PoCs) to enable genAI support in existing processes.

Horizon 2: Independent colleague – genAI enables the automation of certain tasks. Cloud-based/SaaS solutions have a genAI component and businesses benefit from improved intelligence, content generation and analysis as employees adapt to the new way of working.

Horizon 3: Business model disruption – integrated generative AI is deployed across all functions, spanning individual solutions and datasets. Processes are redefined, enabling operating model transformation.

Right now, it feels like there is something of a lull – or an intake of breath – as businesses step back to ensure they have proven a robust business case to scale their genAI deployment, and that they have the full backing of the Board. The Board will be hesitant if it hasn’t been fully educated and convinced around the value potential. At the same time, the risk and regulatory side of the equation is taking up more of the conversation – as it needs to, given the brand and reputational risks that ‘AI gone wrong’ can pose.

Moving ahead with intelligent implementation

To succeed, businesses need to be ‘intelligent implementers’ – linking AI to the enterprise-wide cost and value agenda, rather than setting off random pockets of activity. It needs to be:

  • Connected to enterprise capabilities, the business strategy and colleague/ customer demands.

  • Value-driven to make it self-funding (this should always be the mid to long-term goal).

  • Understood technically to scope out the art of the possible.

  • Controlled and trusted, e.g. through a Centre of Excellence approach that creates the governance and controls framework.

It’s clearly essential to have the foundational technologies in place too – risk readiness, cybersecurity, data management strategy, privacy & access controls. AI won’t paper over the cracks of a messy data architecture and the costs for increased capacity, for example Cloud hosting costs, also need to be considered.

AI needs to be owned at the C-suite level and, we would suggest, sponsored by the CEO. The cultural adoption element is also crucial to success – there needs to be a proper programme of upskilling, training and support for staff through the organisation so they are equipped to use the AI tools and outputs and share their success stories with their colleagues (rather than hiding their use of genAI tools) – otherwise the value will simply be lost.

Helping businesses move to the next horizon

At KPMG, we have supported many organisations with their AI strategies and advised on numerous PoC rollouts, for example:

  • embedding genAI in contact centres at a retail client helping to save them 20% in their customer service agent workloads with a projected multi-million-pound cost/value opportunity across the business.

  • working with a multinational consumer goods manufacturer to integrate gen AI capabilities into their central HR platform, allowing them to automate repetitive tasks, make better decisions through data and re-deploy people to higher value tasks e.g. case management.

  • designing and deploying a gen AI-enabled contract review tool for a telecoms client, saving them significant time and cost for a previously manual task and enhancing the accuracy of reviews, thereby mitigating risks and reducing potential losses.

We also have a powerful genAI rapid assessment tool that analyses an organisation’s maturity and readiness, benchmarks against competitors, and assesses the size of the opportunity.

This year, we’re going to see the leading players begin to move ahead on the AI journey, progressing into Horizon 2 and beginning to reap real cost and value benefits. Can you afford to be left behind?

Here are some questions to ask yourself:

  • Do we have clear visibility of where generative AI such as ChatGPT is currently being used in the business? How are we understanding and controlling the risks posed by this?

  • Which ‘horizon’ are we in?

  • Which current and/or planned projects will become a hinderance and can we freeze these to fund the AI work?

  • What does intelligent implementation mean for our business specifically? Have we identified, quantified, prioritised and tested AI use cases?

  • Do we have a roadmap to scale our deployment and unlock increased value?

  • Do we have C-suite ownership and accountability to drive this forward?