Two experts consider important considerations business leaders should keep in mind when using business leaders to drive business transformation projects.
According to Nina Barakzai, an AI ethics and data privacy expert and qualified accountant, transforming your business to leverage artificial intelligence (AI) comes from the idea of project management. “It's all about the steps and checks you've introduced to help you navigate the journey,” she says. “As a finances representative, I want to ask, 'Where is the value? What problem am I trying to solve? What kind of problem is it that I can allocate money to this and measure what was spent wisely?” ”
Speaking at ICAEW's annual conference for its sincere promotion of AI adoption at ICAEW, Barakzai said: “You need to implement these value judgments, particularly when you are responsible for highlighting triple constraints on scope, time and cost,” he said.
Technical debt
Initially, it is important to form a clear business case, says Peter Beard Aca, director of Genfinance.ai and partner in creating ICAEW's Genai Accelerator. “When you're using Generator AI (Genai), you need to think about your ultimate goal first,” he points out. “In general, there are three options. Want to make something faster, better or cheaper? Getting the “Holy Trinity” is not necessarily viable. However, applying one or two of those criteria can help clarify the approach. ”
Beard will share real-world examples of AI early adopters at the annual conference, and when considering how to identify use cases, he explains that team skills in typical finance functions grow at linear rates and AI capabilities grow exponentially.
So there are important areas in the business that exacerbate “technical debt” and that will only get deeper if left unaddressed. Therefore, by assessing where the most important technical debt is rising, and where the costs of omission are at the most serious, you can know where your business will apply AI and associated upskills at the earliest opportunity.
When it comes to pinning the right tools for the job, Beard says the solutions usually come from one of the “Big Six” foundation models: ChatGpt, Perplexity, Claude.ai, Grok, Microsoft Copilot and Google's Gemini.
Barakzai's valuation shows that every model on the market, whether it's major or emerging, has its own quirks and fo. Therefore, employers must be discerning. “The main questions you should ask are: Has the model got the right dataset to address your problem? And looking at how it is trained, does it have the correct intake?”
She warns that if the tool doesn't have the right data type for your question, it won't give you consistent, robust, resilient and relevant answers. “It just expresses the most popular answers. It also has the money and effort challenges that enable businesses to keep their IT systems up to date and meet their business needs.”
Learn by doing
Before and after the onboarding process, senior management should ensure that employees are taking their journey. Tandem must be open to the purpose of AI adoption channels with external stakeholders, such as customers, suppliers and investors.
On the first point, Barakzai states: “It's clear that coming from a compliance and regulatory background, it's essential to understand and explain what's going on effectively.”
She focuses on the key questions employees want answers: What is the tool going to do? Personally, how does it help? And how does it help us as an organization? “In a way, this trip is more important than a destination and people tend to learn from doing it. So, help staff engage in using this tool together and use it and make sure everyone is together.”
Barakzai emphasizes that the joint approach also supports external messaging. In particular, some business relationships can provide a foundation for rebooting and moving forward from an entirely new set of principles.
“AI offers opportunities for customization,” she says. “Some of your relationships go back many years, and the stakeholders themselves repeat to create new products and services. So, explain what you're doing and ask, “How does the way we're innovating with AI tools help us achieve some of your own goals?”
According to Beard, investors are important to be as transparent as possible to ensure backing of AI projects. “It's about expressing how AI fundamentally and systematically transforms the game,” he says. “Because of its versatility, the scale and range of waves for this transformation are different from what we saw before.”
Higher stakes
Inevitably, AI tools cast occasional curveballs or surprising results due to unknowns about functionality. Beard explains that many AI models come with “Reason Mode,” allowing users to see not only the answers but also how the tool has reached it.
In terms of ensuring ethical hygiene around the solution of your choice, Barakzai encourages users to consider implementation and use in the context of five widely accepted responsible AI principles. In other words, the tool is:
- Focusing on humans,
- Privacy and security are enabled,
- Comprehensive,
- Robust, and
- Resilience.
However, protection against errors and their knock-on effects requires a comprehensive risk management approach. In the case of a beard, the three biggest factors that can lead you to get things wrong are inappropriate platforming, inappropriate training and inappropriate use. Therefore, businesses must be particularly sensitive to these risks. In parallel, he says they must balance the triad of effort, impact and risk. This includes segmenting low-risk and high-risk tasks, taking into account the potential impact of errors in these categories and allocating risk management efforts accordingly.
“For example, a low-risk prompt might be “create a summary of the oldest customers who are behind the bill.” It doesn't necessarily have to be accurate to pounds and pennies. Conversely, “producing and sending business VAT calculations” offers a much higher stake on it, as the cost of errors becomes more serious.
So, is there any point that businesses can say is “end” of the AI transformation project? “It's constantly and repetitive,” says Barakzai. “We need to reevaluate our solutions very regularly, but we need to be careful not to automatically assume that the choices we made a while ago will serve as guidelines for today and tomorrow. Companies have developed an agile approach to addressing changing regulatory situations.
