AI will create a new divide between companies
Two extreme approaches to artificial intelligence can be seen among SMEs. Some supposedly treat AI as a breakthrough comparable to the birth of the internet, yet implement it with great caution. Others decide that it is enough to buy a few subscriptions to ChatGPT, Claude or Copilot and the question of transformation is dealt with. Both approaches can lead to wrong decisions.
In my view, the biggest change that AI will bring will concern not technology but the way a business is managed. And this is exactly where a new division of the market will emerge, one that will considerably strengthen some companies and significantly marginalise most of the others.
The divide will not be between rich companies and poor ones, or between large and small, but between businesses that use AI to rebuild their business processes and those that confine themselves to generating emails, presentations and notes with the help of artificial intelligence.
AI today is cheaper than many people think. Looking only at the cost of a single query to a language model, you could get the impression that artificial intelligence is becoming almost free. The cost of generating a response from a model is falling very fast. At the same time, more and more high-quality open-source models are putting price pressure on the largest providers of AI technology. That is good news, but there is another side to the coin.
The companies developing the most advanced models are spending gigantic sums on data centres, GPUs, electricity, research and development, and hiring the best specialists in the world. That is why the profitability of the AI sector as a whole remains an open question. We do not yet know whether today’s prices really reflect the full cost of providing these services. It is possible that part of the market is currently driven more by the logic of winning market share than of maximising profit. The rising cost of energy and access to water, both of them essential for running data centres, are unknowns.
Will AI become more expensive?
A question that often comes up in public debate is: “In a few years’ time, will using AI cost many times more than it does now?” In my view, there is no clear-cut answer. I do not expect a sharp rise in the price of basic services. Technological progress is too fast. Models are becoming more efficient, and the cost of processing data is falling steadily. The market will probably begin to segment clearly, however, much as in aviation.
There will be:
basic models available to the mass market (the low-cost airlines),
business-class solutions for companies,
advanced agentic systems (the private jets) offering far greater capabilities.
Basic AI intelligence will get cheaper. Expert AI intelligence will get more and more expensive.
And this is where an important question arises for business owners: why will the money a company spends on AI not, in itself, decide who has the advantage in the market?
Spending on AI – more, or smarter?
Many entrepreneurs assume that the advantage will go to the organisations that spend the most on artificial intelligence. Business history, however, shows something different.
Personal computers – In the 1980s the computer was an expensive tool. It seemed that only the largest businesses would be able to derive real benefit from it. Things turned out differently. The greatest advantages went to the companies that changed the way they worked, not to those that bought the most equipment, which then went unused.
The internet – In the mid-1990s, having your own website was a competitive advantage. A dozen or so years later it had become standard. The winners were not the companies that spent the most on internet technologies, but those that learned to use e-commerce, marketing automation, customer data and new business models.
ERP, CRM and similar systems – Simply implementing expensive systems gives no advantage if they are barely used.
It will be much the same with AI. Let us imagine two businesses:
Company A – Pays PLN 5,000 a month for AI solutions.
Its employees use AI mainly for:
writing emails,
creating presentations,
summarising documents,
translation.
As a result, individual productivity rises by anything from a few per cent to a dozen or so, and employees’ time is freed up – time that unprepared leaders will not be able to put to use. Surplus staff may be laid off, but that in itself will not give a competitive advantage in the market. Others will, after all, do the same.
Company B – Also pays PLN 5,000 a month, for:
rebuilding the sales process,
automating the preparation of quotations,
using AI in financial controlling,
analysing all meetings,
deploying agents that support employees or replace some of them.
As a result, productivity may rise by several dozen per cent, and in some areas even double. This will give a real competitive advantage and create a new elite of companies using AI, which will make use of:
advanced AI agents,
their own knowledge bases,
integrations with ERP,
integrations with CRM,
AI-supported analytics,
automated decision-making.
In organisations like Company B, AI will work around the clock. Alongside them there will remain a large group of businesses like Company A, using AI mainly to generate text, notes and translations, which will make work more efficient but will not give an advantage in the market.
Opportunity and threat for SME owners in the age of AI
Paradoxically, the biggest beneficiaries of AI may turn out to be small and medium-sized enterprises, because the owner of an SME often has no access to the specialists that large organisations have, such as a financial controller, a data analyst, a strategist, an HR director or a process expert. Artificial intelligence will not replace these skills entirely, but it can significantly widen access to knowledge and analysis that only a few years ago were beyond the reach of most smaller companies.
An entrepreneur who learns to manage a team of specialised AI agents will be able to make decisions on the basis of information of a quality comparable to that once provided only by large teams of experts. In practice, this may reduce the advantage of large corporations, although at the same time it will make smaller companies dependent on the big tech firms that build AI systems.
The greatest threat to SMEs is using AI superficially and starting the process too late relative to the competition. Business owners often think they still have plenty of time before they implement AI comprehensively, in the operational layer of the company. That is like believing you can start a 100-metre race later than the other runners and there will still be time to catch them up. Staying in the starting blocks will be costly, even though it is not 100 per cent clear how to run. Companies will learn AI on the move – at a jog, perhaps, but unmistakably forwards. SME owners, run ragged by day-to-day operations, have no time to think about the future and to look at their business afresh.
If, however, the owner of an SME uses AI to rebuild the processes of:
sales,
financial controlling,
customer service,
knowledge management,
reporting,
decision-making based on historical data and simulations of future scenarios,
then they can achieve results that bear no comparison with a mere increase in the productivity of individual employees.
So the question is not “Do you use AI?”, because most entrepreneurs will answer yes, but:
“Which key processes in your company will be designed or rebuilt with the help of AI (and when), so that they deliver significantly more value than today to your customers and to your company?”
If you are thinking about implementing AI step by step in order to improve your results and your liquidity, please get in touch.