Practical implementation and governance: building an AI-ready business

If you haven’t caught our previous AI-related blogs – which you can find here ‘the reality of change’ and ‘shifting mindsets away from AI fear’ – one thing is clear: AI is becoming an increasingly important tool.

With this in mind, business leaders must consider how they can best set up their organisation and people for this shift, so the tools can be used successfully.

To better understand ethical AI leadership and technology change management, we hosted a webinar with HR Director Lizzy Barry, Director of HR Consulting Dan Grace and Senior L&D Advisor Vanessa Aradia.

In this blog, I’ve summarised some of their conversation around how you can get your business AI-ready.

Creating an AI policy and guidance 

During the webinar, Lizzy covered the importance of having strong AI governance, saying: “We previously covered the lawsuit that arose when applicants were unsuccessful due to AI bias.

“This is a really salient example of some of the risks we face by using AI without putting in place proper checks and balances – these are not small things or minor legal issues.

“Ethically, it’s also not the right thing to do, as it’s people’s lives that can be shifted by this technology.

“We all have a moral, as well as a legal obligation, to make sure we are using these tools effectively.

“Some of the things we’ve been doing at IRIS, guided by best practice and upcoming legislation in a range of different locations, is making sure we have those essential data privacy, data security and AI-usage policies in place across the organisation.

“Is it clear to everyone in the business what sanctioned AI tools staff can use at work, and the sort of information they can put in these tools?”

“Are we teaching our employees how to use AI so they understand that not everything it outputs is true?

“I’m sure many of you have heard of the situation where a legal firm decided AI could write a lot of their summaries and do the background research on some of the cases they were preparing for, and the AI made up all the citations and laws that were referenced.

“Nobody checked the information before it went out, and it was an enormous embarrassment that made international news, highlighting not only the failure of AI to do something correctly, but ultimately the failure of the people using it.”

AI literacy

Lizzy also shared some insights into IRIS’ AI literacy programme, detailing the areas we cover, such as:

• How to assess AI-generated content
• How to fact-check the information
• Reviewing images and videos that could be AI-generated
• What information is safe to share with AI

She added: “We must make sure we are managing all these risks in the same way we manage other areas like finance and data protection.

“We need rigorous safeguards in place, and to ensure that people understand how to keep the business safe.

“Employees also need to know where to go if they have questions or concerns.”

Practical steps to get your business AI-ready

When it comes to getting your business ready for AI, Vanessa shared some practical advice:

• Start small: identify high-volume, low-value processes ideal for automation 
• Apply the SAFER framework: sustainability, accountability, fairness, explainability and resilience
• Cross-functional governance is key: AI strategy must involve HR, IT, compliance and operations
• Build explainability into every AI deployment: users should understand what the system does and why.

Vanessa explained further, saying: “Governance is your guardrail for trust.

“A transparent audit trail and clear ownership model protect both the business and the people behind the data.”

“If your AI can’t explain itself, it shouldn’t make a decision. 

“At IRIS, we’ve built guidance for teams around what data is safe to share, when to involve human oversight and how to validate outputs.

“That kind of clarity protects trust because people know where they stand, and trust is the foundation of every successful technology rollout.” 

Avoiding AI bias

AI bias is becoming an increasingly common issue, which Dan touched on in the webinar, saying: “The uncomfortable truth of AI is that because the systems are trained on historical data, they’ll have the historical biases that exist in everyday media, especially if it’s available on the internet.

“The internet is a non-safeguarded treasure trove of information.

“There are a lot of opinions that aren’t unbiased – in fact, they’re normally heavily biased.

“When it comes to practical mitigation strategies for using AI, firstly, I recommend that if you’re looking at bringing in an AI tool, always look at pre-implementation due diligence, which can include requesting a bias audit from the vendor.

“A bias audit is different from their marketing material.

“It should be a real audit result, covering what the data training sources are, how they verify it and what the data demographics are – this will help you establish biases in the tool.

“Secondly, you should use diverse representative training data, applying fairness and awareness algorithms.

“If you’re not technically-minded, the vendor should be able to help with that, or if you have an internal software engineering team, they could also support.

“With AI, the non-negotiable is that you should have ongoing monitoring.”

Webinar: Beyond the hype – ethical AI leadership and change management for HR and payroll

If you want more guidance, good news! Our no-nonsense, practical session is now available to watch on demand.

In this webinar, you’ll learn how to:

• Transform employee fear into technology adoption with ethical leadership
• Build an AI-ready organisation that amplifies human potential
• Create practical governance frameworks for data privacy, security and company policies
• Develop employee training and AI literacy programs
• Shift your HR function from reactive to strategic

This session will give you the framework and tools to lead confidently through AI transformation, without losing the human touch that makes your organisation special.