What's ahead in 2025?

Written by Sven Martin (CEO – Australia)

We’ve all got a couple of years under our belt of considering, prodding, and, in some cases, actually putting AI to work. 

I predict that in 2025, businesses will start seeing real use cases with meaningful service or efficiency impacts across their business processes. This will shift their thinking and perception of the importance of using AI to enhance their workflows and productivity - and they will realise the value of looking at their AI strategy.

What’s changed?

Business maturity and confident use cases will be the big enabler for AI in 2025. 

While last year was the time for many to watch how other businesses made progress with AI, I believe this year belongs to those who have done their due diligence. They now have a clearer vision of where and how AI can deliver an operational ROI and are ready to “make it so.”

Tangible use cases will be the biggest motivation for investing and adopting AI. These use cases will give confidence to the cautious and demonstrate the real-world viability of AI to the tentative. 

What adopting AI does require is a change of mindset. AI delivers best (and transforms most) when you’re not trying to replicate existing processes – but are prepared to rethink them. And when you are willing to fail fast and start again if needed. 

Some use cases that impressed the heck out of me

There are more and more examples of how businesses and organisations can leverage AI in the real world to change how they can deliver services, improve operational performance – and even save lives. One (timely) example is that of Alberta Wildfire, who fight Canada’s record wildfires using an AI-powered tool to help duty officers make decisions and use resources more strategically. Trained on a massive amount of historical fire data, the tool makes predictions based on regional weather and forest conditions – and can accurately predict the likelihood of a new wildfire 80% of the time. 

In another example, our own team demonstrated a high-value AI use case for e-commerce organisations. Online stores often struggle with the effort and time it takes to compile and present the content for hundreds, if not tens of thousands, of products needed for their shopfronts. Each listing requires a photo and description, a SKU, dimensions, colours, technical specifications, price, and more. 

We used AI tools, including natural language generation tools, in conjunction with Microsoft Dynamics 365 Finance & Operations (it could be any ERP) to match the F&O product lists against a library of product manuals, pull out all the relevant product descriptions, photos and technical specs from them, rewrite them to fit the website format, and automatically populate the entire shop front. 

Did it work? Yes. And it took just minutes to populate an entire store - not days, weeks, or months – so it’s a game-changer. Best of all, it took little time to deliver an AI tool that offers significant user and business value.

I firmly believe use cases like these will pave the way for more confident adoption this year by those keen to reimagine how (well) they do business. Watch this space!

Written by Jacob Gunzburg (CFO)

Despite all the hype, AI (artificial intelligence) is definitely more than just another buzzword; it's a critical component for driving efficiency, enhancing accuracy, and generating insightful data trends. And I believe that the use of AI will significantly contribute to our ability to reshape business operations and strategies this year. 
Here are my thoughts on the opportunities and challenges it presents to those in the CFO role.

Lessons learned: Don’t reinvent the wheel.

Internally, the potential for AI to streamline Fusion5’s processes and provide actionable insights is immense. Our AI initiatives will be hugely important to us – and our customers - from an efficiency, effectiveness, and accuracy perspective.

So, it’s no surprise that we’ve already started to harness AI for various operational functions where we can support our ability to look forward (which most CFOs would agree is something we don’t do enough of). Which is great.

However, we are mindful that without a coordinated approach, we risk increasing operational siloes - leading to wasted resources and missed opportunities for business-wide success. We quickly learned the importance of establishing champions within our teams to oversee our AI initiatives so that what we learn (and have success with) is communicated with others – rather than learning from scratch for each project.

Turning AI into tangible value

Several months ago, I was part of a forum with around 20 other CFOs, where we discussed our respective visions of how and where AI would deliver value for each of the businesses we represented. It didn’t disappoint - some of the ideas were truly exciting, unique, and innovative.

Was there consensus about where and how AI would deliver the most tangible value? No. But we all agreed that we could and would ‘do better’ (yes, even Fusion5!) when it came to our pace of adoption. We were missing out on opportunities and risking not being first to market with our innovations – a competitive disadvantage which translates to lost potential income in anyone’s language.

While there’s evident clarity of how products such as Microsoft Copilot will improve both personal and business productivity and efficiency, I believe the bigger AI projects will drive an ROI and cement your position in the market. 

Where to next?

From the CFO's perspective, AI is a long-term investment that will yield measurable returns. However, any expenditure should have a clear justification and result in demonstrable benefits to both your business and customers.
That said, what can’t be quantified is the leap of faith it takes to overcome the paralysis of indecision that stops progress in its tracks. What is certain, though, is that standing still in a changing world is never a good growth strategy. 

Written by Kris Jackson (GM - Cloud, Infrastructure & Security)

My prediction for 2025 is that deepfake technologies will be the ones to watch. And this is important because while we can defend you from many other attacks, spotting attempts to use deepfake on a one-to-one basis is more challenging - and needs your help.

Not-the-real Musk?

Forbes recently estimated that deepfake has generated US $12 billion in fraud losses globally. While Elon Musk may be the most popular persona used by scammers, it’s increasingly easy to be fooled by a digital interaction with someone you thought you knew well. Forbes cites one high-profile case where cybercriminals used voice cloning to impersonate a CEO and convinced an employee to transfer US $243,000 to a fraudulent account. 

By using hyper-realistic visuals, voice cloning technology, and leveraging the easy accessibility of AI tools, deepfakers are playing for high stakes – and too often winning. While you may think we can fight AI with AI, Forbes observes, “Top deepfake detectors only catch the phonies 75% of the time.” This means that your personal vigilance is critical. 

It can’t happen to me, or can it? 

Here are two deepfake close calls we talk to our customers about.

First, an example from New Zealand. In late 2023, Christian Pellone, CFO of Zuru, was targeted by a deepfake video conversation with his boss, Nick Mowbray. Fake Mowbray tried to dupe Pellone into transferring cash during a WhatsApp video call. Luckily, Pellone was suspicious about (primarily) Mowbray’s use of the chat function instead of voice during the call, and he didn’t action the transfer. 

Next, in mid-2024, the mayor of the Sunshine Coast in Australia was impersonated in a live video call on Skype. Fortunately, although she looked great, she didn’t sound ‘normal’ to her caller, and the deepfake attempt was foiled. 

In both cases, human vigilance saved the day—not AI. And as AI advances in sophistication, it will become even harder to spot these attempts.

What can you do?

If you rely heavily on digital communication, you’re especially vulnerable. Criminals can tailor their attacks using readily available data from social media profiles and corporate websites, making their phishing attempts appear all the more credible. 

No one will be exempt from these targeted attacks. A moment's inattention can lead to a critical breach that compromises your organisational integrity. This is a wake-up call for senior management to adopt a proactive stance on cybersecurity and recognise that traditional training methods may no longer suffice. 

Sadly, trust in human goodness won’t save the day. When it comes to deepfake, adopting a culture of scepticism towards unsolicited communications is vital, as is encouraging your employees to verify information through multiple channels.

Written by Becky Rutherford (COO)

I’m going to put this out there: My prediction for 2025 is that Fusion5’s focus on using AI efficiently and effectively in our business will pay tangible dividends for both us and our customers. 

If you’ve been anxious about where and how to start adopting AI, you’d want your partner to have firsthand expertise in developing use cases in their own business. And quite rightly so.  

So, we’re on a mission to use AI to improve our performance and processes and add more value to our customer relationships and projects. 

So, how, why, and where will we use AI? 

We aim to use AI primarily to free up our consultants to do what they do best: consult. This is part of a larger initiative to completely revamp our consultancy enablement program – across the entire business. 

What can you expect from us in the near future? Rather than using our consultants’ time to write extensive documents, take meeting notes, and perform other ‘perfect-for-AI’ tasks, we will be able to leverage their experience, knowledge, and ability to provide higher-value outcomes for you. Not only will using AI make our people more efficient, but it will also speed up the engagement as you’ll get what you need more quickly. 

By the end of 2025, we should be using AI to generate the first draft of your business requirements using the content we gather from workshops and Teams meetings. We also plan that by year’s end, we’ll be able to use AI to create initial drafts of user stories and test criteria and any other documentation we traditionally prepare to align and define needs, expectations, and deliverables. So, watch this space. (Also, you’ll note the use of the word ‘drafts’ – all reviewing will be human.) 

As a business with over 20 specialist divisions ranging from solutions to services and support, the challenge for us will be to agree on how to do this and align the outcomes so they are usable by everyone.   

What’s in it for you, and what can you learn from us?

We expect the biggest future impact of our use of AI on our customers will be our ability to pass on the benefits. 

For example, by using AI to streamline how we produce your project documentation, more of your budget will be spent on high-value engagements - not on the hours and hours required to generate the supporting paperwork. So, by the end of 2025, you should be getting more bang for your buck, and our consultants will get to focus on what they love – delivering great outcomes. 

What can you learn from us? Well, we’ve been taking a bottom-up approach by asking everyone at Fusion5 to contribute their ideas on how AI can improve their jobs, our business, and the customer experience. 

Our people have come up with some great suggestions, such as using AI to review contracts. Practical, easy-to-implement ideas like these will add immediate value and efficiency to our business processes. (If we’d restricted input to a select few in the business, we’d be the poorer for it, so we recommend this strategy.) 

We’re adopting many of the small AI initiatives suggested. If they don’t gain rapid traction within the business, we can agree to fail fast. Then, we can pick ourselves up and move to the next idea with minimal time wasted or damage done. 

AI doesn’t need to be a big-bang approach; what’s important is starting - and constantly looking for ways to go beyond what you and your customers expect. 

Written by Troy Gerber (CTO - Conversational AI and Copilot Microsoft)

For me, the technology that shines mostly brightly for 2025 is agent-based AI.

Yes, just when you were coming to terms with generative AI, agent-based AI has emerged from the depths as the one to watch this year. It may sound like souped-up generative AI - but I’m confident that agent-based AI will change everything. Yet again. 

Why is agent-based AI hot?

You’re forgiven if you roll your eyes when I say that agent-based AI will redefine how we interact with technology. And fair enough. 

But in truth, agent-based AI does represent a major leap forward from traditional AI assistants. While gen AI can provide answers and suggestions, agent-based AI goes further to actively execute tasks on your behalf.

As well as responding to questions, your digital assistant (aka agent) can autonomously manage your schedules, send emails, and fill out forms based on your instructions – and more. Using any or all the task-oriented agents available (each has specialised skills) will instantly help you get more done.

Microsoft has already begun integrating agent-based AI capabilities into their platforms, promoting some of their Copilots to agent status. These agents aren’t loners either, they play nice with other agents to execute a series of steps. For instance, your coach agent pulls key points from a presentation, then hands them over to the writer agent to turn into a blog – without you lifting a finger after your initial request. Laptop playing up? Your agent can automatically create a service ticket – so there’s no need to log the issue manually. What I love about this? This hands-off approach saves time and means less to worry about in an already busy day.

Who will it impact?

Everyone. However, business users are especially set to win as they gain their own executive assistant. For example, an agent can take care of the routine Monday morning email requests for updated sales figures, project updates, supply chain status, etc. As well as requesting and collecting the information, your agent can return it to you in your preferred format (say, a spreadsheet) via your favoured messaging platform.

Suddenly, admin chores (and more) become less of a thing. 

What next?

I believe the key to leveraging agent-based AI lies in early adoption and integration, so it’s important to get a handle on how the technology will benefit your business sooner rather than later.

The downside? It’s honestly hard to find one. Yes, you’ll need to train your staff to work with agents (as with any new technology). But once they’ve got to grips with them, you can expect to see (or hear about over a coffee or a beer) some great stories about improvements in personal productivity and work-life balance. 

Written by Grant McKeen (Technical Director – Integration)

My prediction for this year is that integration will be all about platforms, not products. 

Let me explain. 

The typical approach to integration (until now) has been to buy multiple and often disparate products - and then engage a technical team to configure the logging, monitoring, security, policy management, etc. 

So, what’s changed? Let’s start with the people. 

Today’s organisations are increasingly supplanting technical teams with cross-functional product teams. 

If you haven’t encountered the concept before, a cross-functional product team is a diverse group of individuals from various departments (think marketing, analysis, development and customer support) collaborating to deliver a single product. 

How does this impact your integration technology?

To enable a cross-functional product team to be effective, the technology they work with should be provisioned to give them autonomy. 

This needs to happen without the team having to learn and certify in using a new product or relying on a technical team to configure it for them. And aside from this, most cross-functional product teams don’t have enough seats in them to afford the luxury of an integration product specialist. 

Supporting the cross-functional structure demands a shift to a robust platform that centralises capabilities like integration, monitoring, and security, eliminating the need for (any) teams to juggle multiple products and their complexities.

What platforms are we talking about here?

Platforms such as AWS, Azure, IBM Cloud, Red Hat, Oracle, Heroku, and Google Cloud are among those leading the charge. 

Each offers an integrated environment where a cross-functional product team can quickly deploy necessary resources without the hassle of configuring multiple disparate products. This capability allows businesses to focus on delivering value rather than managing technology.

Importantly, a true platform has everything a cross-functional team needs ready-baked into it. The environment comes with all the integration components, telemetry, logging, monitoring, security, and other utilities they need – so it’s instantly usable. 

Is this the death knell for standalone integration products?

Probably – it’s been a downward trend to go standalone for the last five or so years. But I’m not alone in saying ‘platforms, not products.’ 

In Gartner’s 2024 whitepaper, Adopt Platform Engineering to Improve the Developer Experience, they say: “Successful platform engineering reduces the cognitive load on product teams by providing self-service tools, capabilities and processes.”

And I couldn’t agree more. 

Written by Rebekah Pine (GM - Customer Success & Employee Engagement)

Big news: we’ve developed a use case for AI in employee engagement in 2025, and it’s becoming a reality! And my prediction is that we won’t be alone in using AI this year to revolutionise how we listen to and understand our people. 

The integration of AI in the employee engagement model represents a significant opportunity for Fusion5, especially given the rapid increase in our staff numbers over recent years. Yes, growth is a wonderful thing. But as you add more people, it becomes harder to understand how they feel, to promote useful and open dialogues, and to create a more engaged and happier workforce. 

AI will enable us to do this and more, and I’m exceptionally excited by the prospect.

How will AI make us a better employer?

The more people you have, the harder it is to listen to them all. As we’ve grown, the traditional (i.e. manual) methods of gathering and gauging employee feedback have become too cumbersome and slow – which is frustrating for everyone. 

The highlight of using AI for ‘listening’ will be our ability to collect and efficiently analyse employee sentiment quickly. We’ll be able to process our survey data within 5-10 minutes and immediately identify what our employees feel is working well - and where there’s room for improvement. 

But our shift towards AI-driven analysis isn’t just about speed, it’s also about depth. AI tools can sift through vast amounts of data to uncover more nuanced insights. Without those insights, your capability to foster a culture of openness and continuous improvement is limited.

I expect AI's impact on employee engagement will be felt across all levels of Fusion5. For our leaders and managers, it will provide valuable feedback on team dynamics and specific areas for improvement. And it will allow us to highlight areas where our managers excel or might need some additional support  (at around triple our previous cadence). 

Learnings from our AI journey (so far)

My recommendations if you’re considering building AI into your own employee engagement programme.

• Start small. Begin with pilot projects to test AI tools in specific teams or departments before scaling up.

• Educate your leadership team. Buy-in is critical to ensure they understand the value of AI in enhancing employee engagement.

• Focus on anonymity and privacy. AI tools must be designed to protect anonymity while providing actionable insights.

• Champion continuous communication: Your employees should feel that their voices are heard and that their input leads to tangible changes.

• Make sure you have enough runway! Don’t go overboard. Set a sensible survey schedule so you have enough time to follow through with actions.

• Evaluate and flex: After implementation, continuously assess your solution’s impact on employee engagement and adjust as necessary.

What about people?

Yes, AI is amazing. But it’s essential to remember that successful employee engagement still depends on genuine human connections and one-on-one conversations. People will always be at the heart of any great employee engagement programme.

Written by Kristy Brown (CEO – New Zealand)

My prediction is that if you’re not already gearing up to leverage the power of AI in 2025, you’re already falling behind your competitors. 

AI-driven collaboration tools, particularly those with advanced capabilities, aren’t just another software enhancement. They represent a significant shift in how your workflows can be optimised, productivity enhanced, and success measured – and they’re evolving at breakneck speed. 

Why AI-driven collaboration tools are the one to watch

The significance of AI-driven collaboration tools lies in their ability to fundamentally change how your teams work together. Seriously, tools like Microsoft Copilot allow for real-time insights and collaboration that we’ve only dreamed of until now.  

Two AI tools that are particularly exciting to me are Copilot Pages and Copilot Analytics. (Note: I’d love to talk about agent-based AI, too, but our CTO beat me to it).

Satya Nadella, Chairman and CEO of Microsoft, recently posted about Copilot Pages on LinkedIn, saying: “I’ve been using Copilot Pages to prepare for meetings and it's been a game changer for me and my team. It puts all of us on the same "page" before, during and after meetings, creates a first draft of our agenda and meeting notes, and can be shared in real-time with everyone involved, enabling us to ideate with AI and collaborate with each other.” So yes, it’s pretty impressive! 

Copilot Analytics finds ways for you to truly measure the ROI of your Copilot licensing and transformation. Until now, most businesses have looked at how many emails Copilot has helped Employee X write or how many meetings it’s captured for them and tried to extrapolate that data into an efficiency gain that translates into an ROI – but that doesn’t truly measure how effective they are at doing performing the core role you’ve employed them for.

Copilot Analytics turns that approach on its head by looking at the KPI of a standard role (say, a contact centre worker or salesperson) and asking how - and to what degree - Copilot accelerates their measurable performance - not the other way around. 

In my book, that’s transformative. The focus on role-specific KPIs means that users won’t have to become AI experts but can instead focus on their core job functions. This tool proves that the actual role can meet its KPI better with Copilot than without. It’s a mind-shift that allows you to understand the true value of your technology (and makes it easier to justify the investment).

Given these new innovations, what’s the problem?

I can guarantee that AI-driven collaboration tools will have a widespread and positive impact on your organisation and its stakeholders. But only if you act sooner rather than later. 

If you’re living in a state of “POC paralysis” - bogged down in proof-of-concept stages – you can and will stifle innovation, which leads to lost opportunities. By delaying AI adoption, you risk falling behind your competitors and constantly playing catch-up.

Where to start? Your journey should begin with understanding potential impacts, defining role-based use cases that can drive immediate benefits, and working with an experienced partner (that’s us) to enable this transition. 

Now is not the time to hesitate. 2026 could be too late.