ARC: A framework for successful AI adoption

Treating AI as something that can be bolted onto existing systems is not enough. Rather than merely adding AI to existing workflows or products, the real value comes from building entirely new business models, experiences and sources of value with AI at the core.

The ARC framework – Acceleration, Replacement and Creation – describes the main ways to strategise for successful AI adoption. By identifying AI use cases in these three categories, organisations can increase their odds of successful deployment. 

You can accelerate existing tasks and processes, cutting costs through automation and optimisation. These are the simple tasks that AI can make easier and more efficient, and this is the place where most organisations start when adopting AI. 

Replacement occurs when the extent and reliability of automation means AI-based systems can effectively take on the whole of a job that previously involved a human.

Creation is about taking the opportunity to reimagine your business, create entirely new forms of value, and set yourself apart in ways that were never possible before.

But whichever phase an organisation chooses to prioritise, the groundwork has to be done. As Sebastian Heinz, CEO of Statworx, recently put it in a LinkedIn post: "AI isn't failing. Companies are. They chased hype… without laying the foundation: No data management. No operating model. No enablement. No strategy." 

That could signal the beginning of what Gartner's famous Hype Cycle calls the Trough of Disillusionment: a phase where overinflated expectations give way to reality. Some might even call it the onset of another AI Winter. But that's not a bad thing and that's not necessarily a setback. It's a turning point that draws a clear line between those chasing buzzwords and those delivering real outcomes and value.

Let ARC guide, not prescribe

While ARC framework is used to describe the progression of transformative technologies, the three phases are not necessarily sequential. In the short to medium term, SMEs may never get beyond acceleration, or possibly replacement once the software they use becomes sufficiently AI-infused. At the other extreme, some companies – especially startups and certain enterprises – may choose to leap directly into the creation phase, bypassing the earlier stages altogether. A word of warning: remaining in the early stages for too long can leave you trailing behind competitors who are quicker to innovate and differentiate.

The mistake too many companies made 

There was a rush to experiment with AI after ChatGPT launched. Boards demanded pilots. Every vendor promised AI-enhanced tools. Prompt engineers were suddenly in demand. Demos were flashy. And yes, some even launched their own "CompanyGPTs."

But many skipped the foundational work. They rushed to showcase AI without investing in the essentials such as a robust data infrastructure, a scalable operating model, organisational enablement, or a clear strategy tied to defined outcomes and value. Instead, they placed bets on surface-level wins, hoping momentum alone would carry them forward.

And now questions are starting to be asked. Why is our AI pilot stuck? Why is adoption so slow? Where is the return on investment (ROI) in AI?

Automation is table stakes 

Using AI to improve customer service, automate support tickets, or streamline workflows is essential, but doesn't provide differentiation. Nearly every organisation is already doing it or will be soon. Such capabilities are necessary for survival, but won't set you apart. 

Real leaders are pushing beyond automation and asking a transformative question: How can AI help us deliver something fundamentally new that we couldn't offer before? That might mean reinventing a product, reimagining how customers engage with the brand, or unlocking new business models.

That's where the outsized returns are and where durable competitive advantages can be created. 

Building for AI value

While AI's impact at the macroeconomic level has yet to spark widescale productivity bursts, a more compelling picture emerges at the enterprise level.

For example, US-based grocery chain Kroger has invested in AI-powered solutions that help them keep better track of inventory and expiration dates. When combined with network-connected freezers and fridges that can alert store managers that there may be a refrigeration issue, the grocery chain has reduced food waste and can better ensure that customers are always getting the freshest products possible.  

Such projects aren't about incremental improvement or layering AI onto legacy systems. The activities of Kroger and other companies signals a full entry into the creation phase of the ARC framework, where AI fuels not just efficiency but strategic reinvention. Whether it's accelerating product development, redefining customer experiences, or outpacing peers in earnings performance, these organisations are using AI to build what others haven't even imagined yet.   In fact, The Tech Council of Australia's Future Ready: Australians and AI Workplace Tech report launched in August 2025 found that 93% of Australian workers believe AI will impact jobs by augmenting them, not replacing them.

The right way to invest in AI

If you're leading an AI initiative, use ARC as your compass. Identify whether your investment is augmenting, replacing, or creating. You need all three, but you need to aim for differentiation. So innovate, don't just automate. Use AI to create unique experiences. Demand measurable ROI, and if a project isn't delivering results, move on. Test fast and learn faster. But remember to lay the groundwork, because strong data, governance and enablement are essential for AI success.

And think holistically: AI isn't a feature, it's a foundational capability that should span your business.