AI opportunities discussed varied from operational efficiency use cases, a current favourite, to eventually embedding AI capabilities directly into the product itself.

One participant described how deploying large language models (LLMs) trained on the expertise of senior call centre agents significantly improved customer data accuracy and reduced call volumes.

Within the CFO function, AI is being applied to areas such as cash forecasting and contract reviews, illustrating how the technology can strengthen financial governance and enable faster, data-driven decision-making. One participant noted great success with training an AI agent on products, instead of spending on a dozen or more human sales trainers to acquire the same skillsets.

Emerging branches of AI such as agentic AI (proactive artificial intelligence systems that can operate autonomously to achieve goals without constant human oversight) and causal AI (which focuses on understanding root causes and consequences of actions) were topics of discussion at the roundtables.

Anecdotal evidence illustrated the growing popularity of agentic AI in achieving workforce efficiencies at an astounding scale. At one business, 120 developers were replaced with 10 agents, prompting a participant to note that “at the board level, we are not talking about going from a headcount of 1,000 to 100 in 10 years, we are talking about going from 1,000 to 200 in one or two years.”

Overall, participants lauded the growing set of proven use cases in back-office optimisation and individual productivity, while noting that these will not bring the greatest value creation in the near-term.  The next frontier lies in leveraging AI to make a step change in core business operations productivity and more complex, strategic decision-making.

Many business leaders are overwhelmed by the hype surrounding AI and struggle to approach it in a practical, no-nonsense way. Some may feel pressured to pursue AI initiatives simply because “everyone else is,” rather than because there is a well-defined business case.

To set up for success in AI initiatives, it is critical to assess and prioritise the areas for impactful value creation. As part of this assessment, business leaders must consider:

1. an organisation’s most material operational cost drivers and productivity levers
2. key strategic questions that the business is facing
3. latest AI capabilities that may unlock productivity or strategic decision-making in the most material areas identified.

A&M assists CFOs and wider leadership teams to establish these priorities and the benefits case, define the execution roadmap with clear resources and timelines, and provide implementation support where required.

While data quality remains paramount, the adage "garbage in, garbage out" ignores a critical variable: access speed. To operationalise AI, organisations must modernise the underlying platform, moving away from siloed storage tiers toward a unified data foundation. Such a platform must be capable of ingesting multimodal streams at scale while guaranteeing the security and low latency required for AI-enabled decision making.

This becomes even more relevant in the context of Agentic AI which, unlike other forms of AI, will be “always on” and therefore create unseen performance requirements for the underlying infrastructure. This architecture offers a distinct strategic advantage: it allows leaders to "leapfrog" gaps in traditional ERP maturity. By leveraging a high-performance data plane that unifies unstructured and structured data, businesses can extract AI value immediately, rather than waiting for legacy systems to catch up.

Algorithms are only a small fraction of a successful AI initiative. For real impact, technological innovation must go hand in hand with cultural adoption both at the consumer and organisational level. This means an appropriate mindset shift within leadership to drive implementation, along with the requisite upskilling of the workforce.

At one roundtable, participants observed a wide variation in the scale of AI adoption across businesses, with drivers ranging from industry to investor profile, to individual factors such as leadership enthusiasm.

Some participants saw a contrast between the stances taken by PE backers (more proactive in embracing AI tools for efficiency and innovation) and corporates (more cautious, facing more restrictions, especially in listed companies). Others disagreed, saying they have seen maximum traction in businesses helmed by AI-ambitious CEOs, regardless of whether PE is involved. Some noted a divergence in AI approach depending on the industry – for example, the IT industry’s focus is on efficiencies, financial services are looking at leveraging AI to make strategic choices, and so on.

Leadership engagement is a critical piece of the puzzle in driving change: only a CEO or CFO with a solid understanding of AI’s capabilities and limitations will be able to articulate a clear vision, ask the right questions, and drive adoption.

Starting with small productivity improvement use cases can demonstrate early wins and build momentum in cultural adoption. When in-house expertise is limited, engaging external specialists can accelerate implementation and reduce risk. Giving teams access to a hackathon to start using the tools and developing small use cases can be very valuable – we see leading PE operational teams getting impact from these sessions.  

As AI usage becomes ubiquitous and more AI platforms become available for various tasks, one common issue plaguing businesses is data leakage. At one roundtable, several examples were shared of employees using AI for professional purposes “on the side,” without proper oversight or authorisation from the organisation, leading to potentially valuable business data being used to inadvertently train LLMs.

While companies are attempting to enforce the use of specific or in-house platforms, anecdotal evidence suggests they are sometimes met with limited success. Establishing a sound AI approach that considers employee needs, behavioural changes around AI and ensures that the right training and protocol are communicated across the board, can help mitigate these risks.

Making the right hires or appointments is critical to drive the business AI strategy, implementation, and governance. According to ZRG, some large-cap PE funds have been encouraging their CEOs to hire Chief AI Officers and Chief Data Officers. These roles were not as clearly defined even two years ago.

However, the primary reason these newly created hires fail to deliver value is the lack of a culture that supports technology development and adoption. To ensure they execute plans faster, with greater returns on investment, boards must first work to identify and assess culture before focusing solely on hiring talent.

ZRG’s CFO practice has observed a marked increase in the demand for CFOs who can demonstrate a track record of early adoption (AI/Data/BI). There is a clear gap with supply being outstripped by demand, so Finance leaders are needing to upskill, re-train and focus more on embracing these changes to be more attractive for prospective roles.