Differentiating your businesses in the era of commoditised AI

If any nation has the potential to be a global AI powerhouse, it is the UAE. Long before ChatGPT’s landmark launch, it was the UAE that created the world’s first AI ministry.

In the years since, the government has never missed an opportunity to encourage private-sector enterprises to adopt the latest and greatest, and to do so itself. And it shows. Last year, the country’s AI market was estimated at nearly a billion dollars ($949.8m), according to one estimate, and is projected reach $4.3bn by 2030 — compound growth of more than 28 per cent across just six years.

The AI assets race has already begun, spurred by the hype around generative AI that has dominated boardroom confabs for the past two years. DeepSeek’s announcement at the start of the year only emphasised the urgency to get ahead, firmly convincing business leaders that they must earnestly investigate how to maximise returns from inevitable investments. Some, it is projected, will continue to devote budget to Generative AI, which is set to be worth more than $200m this year and grow to $2bn by 2030. This is a CAGR of 46.47 per cent, far outpacing AI as a whole. In pursuit of AI maturity, and a companywide AI culture, decision makers are starting to become more pragmatic about where they route their AI dirhams.

The problem lies in differentiation. AI methods — GenAI, natural-language understanding (NLU), predictive analytics, and so on — are standard, and as DeepSeek recently proved, can essentially be commoditized. If a supermarket invests $1m in trying to better understand its customers, it may receive a revenue bump for a short time, but its competitors need only duplicate the investment to duplicate the advantage.

Competitive edges end up canceling one another out and everyone is a million dollars poorer with no ongoing benefit. The only way around this apparent paradox — which can occur in any industry — is through the development of a technology nobody else has or a business model nobody else has considered.

Stand out, move up

First, let us look at technology-based differentiation. Yes, the models are standard and some powerful ones even available as open-source. But what is unique to each business is its data. The data itself — customer, operational, financial, security — may even suggest a use case, and the model can be built around it. The winners will be those who find ways to go beyond the simple automation that saves on labour and differentiate themselves though applying their unique data to more advanced models. They could also use AI to create richer views from their data, which may be comprehensive but may also be stored messily.

On the business front, organisations can look at ways to deliver ubiquitous AI without eating too deeply into budgets. Generative AI is relatively costly, but not every use case requires GPT-4. If one of the supermarkets mentioned earlier were to deliver 360-degree views of their customers at the cost of, say, $700,000, it would have a cost advantage over its competitors, and may come out on top even if it was not the first to invest. In the real world, simple decisions like slimmed down products, or even selecting investment in CPU hardware over GPUs, can make all the difference. Then there are operational decisions. Do you train and maintain AI models in the cloud or on premises? Given GenAI’s costs of compute, storage, and bandwidth in the cloud, in-house hosting may be the better way to go.

But there is no doubt that when pursuing Universal AI, non-technical employees benefit from more advanced tools built on generative AI. Modern solutions allow more employees to add value. Additionally, the organization solves the problems presented by the regional AI skills gap.

Recruitment is time-consuming and expensive, and newly joined data scientists and AI specialists will also take more time on projects than business-oriented employees. Requirements gathering will no longer be a bottleneck when those who know the business best are tasked with implementing their own ideas with the help of generative AI.

AI potential: Increment versus revolution

Already, industry pundits are expressing the potential of AI by making a comparison between 10 per cent gains and tenfold gains. AI can increase the efficiency of supply chains, enhance customer service, decrease manufacturing faults, speed up R&D, and more. The 10 per cent gains are the marginal boosts of old; the tenfold gains are the stuff of which industrial revolutions are made. What automobiles did for transportation, cloud computing did for the software segment, and smartphones did for payments. Now generative AI has arrived to repeat the disruption across a range of industries.

One stumbling block remains to those who choose to embark on a GenAI expedition, and that is regulatory obligation. Only through the right training and cross-disciplinary expertise can UAE companies successfully navigate the legal landscape. Governance must be designed, and it must be designed up front. Waiting until legal issues arise can be expensive, not just because of the potential fines incurred, but because of the expense of redesigning and redeploying solutions, retraining staff, and possibly procuring further tools.

Fortunately, with the right AI platform, governance can be applied consistently and accurately across the organisation. Everything from role-based data access to predefined dataset views can be established by steering committees before a single project is greenlit. Modern AI platforms can even automate monitoring of live solutions for model shift and emerging trends that may throw off results and lead to negative impacts.

Building an AI business takes will and skill. Never let it be said that diving into the next powerful LLM with arms flailing will guarantee results. Mountains of operational, financial, and legal shocks await those who face our AI future like this. Cool heads must prevail. Choose the AI path that fits your individual business and leverage your individuality to make that path unique in your market.

The writer is the Area VP and GM – Middle East, Turkey & Africa at Dataiku