Generative AI for SMEs: how to use it without breaking the budget

Most SMEs do not need to train their own models. The difference between good and bad AI spending is decided at the integration layer, not in the underlying model. GPT-4, Claude or Gemini are available via API for less than the cost of a management software subscription, and they handle the vast majority of business use cases.

The pattern that works: identify three or four high-frequency repetitive tasks, connect a model via API to the relevant internal data and measure time saved week over week. Starting with a specific case, measuring it and scaling once return is demonstrated is consistently more effective than trying to transform the entire operation at once.

• Automated responses to frequent customer queries (support, sales, logistics)
• Internal report generation from data already in the ERP or CRM
• Classification and routing of incoming emails, forms or documents
• Sales assistance: account summaries, proposal preparation, pipeline follow-up
• Extraction of key information from contracts, invoices or technical documents

The most common mistakes are buying all-in-one platforms that cover use cases the company does not have, or implementing AI in processes that first need to be organised. A platform with 50 features of which 3 will be used is not an AI investment — it is an investment in a nice interface.

• Do not buy platforms for their feature catalogue but for the 2-3 cases you will actually solve
• Do not implement AI in processes that are still chaotic or lack reliable data
• Do not pay enterprise licences when an API integration costs 10% of the price
• Do not scale without first measuring the return of the first use case
• Do not assume AI will solve a process problem — first the process needs to be organised

Generative AI does not require a data department or an enterprise budget. It requires clarity on which problem you want to solve, data that is organised enough to feed the system and real time from someone on the team to manage adoption. That is within reach of any SME.