Hurry up AND wait: The case for a three-speed AI strategy for SMEs

Despite the unrelenting release of ‘game-changing’ AI tools, the game remains remarkably unchanged for many traditional SMEs. Yes, AI will transform how we all work, eventually. In the meantime, organisations need a plan to move fast, but not break things.

Much like the dot-com hype cycle (for those of us who were there!), AI will have a more profound impact than we imagine, but also take longer than we think. If it were all as easy as the dot.AI startups claim, organisations would switch AI on and live happily ever after.

Yes, it’s mind-blowing for generating documents, videos, research, songs, code, automations, real-time queries and slightly odd-looking images – and is improving by the day – but has AI (yet) fundamentally changed the core small and medium enterprise market?

The divide between the AI haves and have-nots

On one hand, the Apples, Microsofts, Metas, Amazons and corporates of the world have the budgets to bet on the moonshots that will define the future. Hundreds of AI startups also attract funding and eyeballs as they dream up the next big thing.

On the other hand, back in reality, small and medium enterprises – you know, the ones powering Australia’s economy and employment – are caught in the middle, trying to grow profitably today while adapting to the new age of AI opportunities and security risks.

SMEs can be hamstrung by the dual challenge of navigating bureaucracy and legacy systems without the deep-pocketed freedom to disrupt and break things.

Act now, think long-term

From Nexon’s experience working with SMEs across industries, AI has already proven to be a powerful change agent when integrated strategically and methodically, building on the digital transformation that has been happening for a decade.

However, that doesn’t mean we recommend moving slowly. A three-speed combination of short and sharp pilots, existing process optimisation and ongoing data readiness work can fast-track the outcomes of AI. You can look at this in three stages:

NOW: Run targeted pilots in short sprints to prove results

Pick a critical pain point, defined use case or low-hanging opportunity in one department, and deploy an AI tool to address that. As low-budget and low-risk initiatives, they can bypass bureaucratic delays, demonstrate a result, and succeed or fail fast.

Success in one department builds momentum and buy-in to move to the next and creates a compelling case for a wider rollout. Don’t overthink these; the test-and-learn process is valuable even if it fails. Perfection can be the enemy of progress. Sitting on the bench for six months trying to design and sell in significant AI transformation can sacrifice quick wins and learnings.

NEXT: Unlock AI features emerging in existing platform

Start with the tech stack you already have. For example, if you use any major Cloud platforms like Microsoft, AWS or Google, or modern Cloud-based solutions like Genesys or ServiceNow, you’ll find a lot of the heavy lifting is being done for you. They are all rapidly embedding AI capabilities into their software that you can tap into.

In addition, the latest integration options and APIs enable you to share data and functions between legacy platforms to unlock quick AI wins without reinventing the wheel.

Once you have maximised your existing tech stack, you can overlay new AI solutions or even build your own where it makes sense.

ALWAYS: Smarter data is the real intelligence

The old cliché ‘garbage in, garbage out’ gains new meaning with AI. Access to quality data is the true source of its intelligence and automation power.

Consolidated and up-to-date data will win the AI race regardless of your choice of platform or business challenge. Dedicate resources to ongoing data cleaning, tagging, privacy settings, permission levels, deduplication, verification and integration across multiple sources. Of course, there are AI tools for scrubbing data, too.

At Nexon, we’ve had outstanding results using AI-powered customer experience (CX) and employee experience (EX) tools, but only if they can access the correct information. For example, if you have disconnected data from three websites and four databases in various formats, even the most intelligent CX/EX tools won’t deliver fit-for-purpose answers to the right question on the right channel at the right time.

There will forever be shinier, smarter and faster AI tools. But even if you pick the wrong technology or find something better, the more organised your data the better equipped you are to take advantage of AI innovations.

Let us know if you need help accelerating a pilot program or want guidance with your AI readiness, strategy or data consolidation.