When I speak with SME executives about AI, what interests them most is not the technology itself. What they want to know is: what can I do with this this week? Which specific task can I improve or automate without hiring anyone new or making a large investment?
This article answers those questions with real use cases that Spanish SMEs are implementing right now. I am not talking about future projections or laboratory experiments. These are concrete applications that work today.
Why SMEs are adopting AI now
Three factors have aligned in 2025 and 2026:
- More accessible tools. ChatGPT, Claude, Copilot and Gemini offer business plans from EUR 20 to 25 per user per month. No development needed: they are used like any other tool.
- Mature AI agents. Agents are no longer prototypes. They can execute complete tasks from start to finish: find information, process it, generate responses and act on external systems.
- Regulation that drives training. The EU AI Act requires AI Literacy since February 2025. Companies that get ahead form their teams and, in the process, discover use cases that improve their operations.
According to data from the Spanish AI Observatory (RED.es), 42% of Spanish SMEs already use some form of AI in their daily operations. Not all of them call it that, but it is there: customer service chatbots, content generation tools, recommendation systems in e-commerce.
Automating customer service
This is the most widespread use case, with the fastest return. An AI agent can handle 60 to 70% of recurring customer queries without human intervention.
Real example: an online design products store with 12 employees implemented a ChatGPT-based customer service agent. The agent answers questions about orders, delivery times and returns by querying the CRM database directly. The customer service team went from spending 4 hours a day on repetitive queries to less than 1 hour.
What you need to get started:
- A documented knowledge base (FAQs, return policies, updated catalogue)
- API access to your order management system or CRM
- A conversational AI provider (Dialogflow, OpenAI Assistants, or solutions integrated into your CRM)
Implementation cost is around EUR 500 to 2,000 for an SME, and the return is noticeable within the first few weeks.
Content and document generation
SMEs generate more documents than they realise: commercial proposals, client reports, product descriptions, follow-up emails, meeting minutes, internal presentations.
A team that uses AI to generate drafts of these documents saves between 40% and 60% of the time they previously spent writing from scratch. The point is not for AI to write the final document: it is for AI to produce a first draft that your team reviews and adapts.
Real example: a 20-person consulting firm uses Claude to generate the first draft of audit reports. The consultant reviews, corrects and personalises it. What used to take 4 hours now takes 1.5 hours.
What you need:
- A ChatGPT, Claude or Copilot account (business plan recommended for data protection)
- Templates of the documents you generate most frequently
- A clear process: AI generates the draft, your team reviews and signs off
Data analysis and automatic reports
If your company generates data (sales, web traffic, production metrics, satisfaction surveys) but nobody has time to analyse it, AI can help.
Current tools allow you to connect a spreadsheet or a database and get analyses, summaries and visualisations in minutes. No programming knowledge required.
Real example: an electrical materials distributor with 35 employees used Google Sheets to track sales. They now use an AI agent that generates an automatic weekly report: product trends, regional variations and alerts when a product falls below minimum stock. The commercial director spends 30 minutes reading the report instead of 3 hours building it.
Tools that help: Copilot in Excel, ChatGPT with Advanced Data Analysis, or solutions like Power BI with integrated AI.
Supplier and procurement management
Comparing quotes, negotiating terms, tracking orders and evaluating supplier performance are time-consuming tasks that are well suited to partial AI automation.
An AI agent can:
- Compare prices and conditions between suppliers automatically
- Generate personalised quote request emails
- Alert you when a regular supplier raises prices above the market average
- Summarise the key terms of a supply contract in bullet points
Real example: a corporate catering company with 8 employees receives quotes from 15 raw material suppliers. They use an AI agent to standardise the data from each quote and generate a comparison table. What used to require half a day of manual work now takes 20 minutes.
Fraud and anomaly detection
This use case is particularly relevant for SMEs handling financial transactions: e-commerce, insurers, accounting firms, professional offices.
AI models detect anomalous patterns in transaction data: duplicate invoices, amounts outside the normal range, new suppliers with inconsistent data. They do not replace the control team, but act as a filter that flags what is worth reviewing.
Real example: an accounting firm with 15 employees implemented an AI system that reviews incoming invoices before they are recorded. In the first month it detected 3 duplicate invoices that would have generated erroneous payments totalling EUR 4,200.
Anomaly detection tools are available in platforms like Xero, Sage or QuickBooks with integrated AI modules.
How to start without a big investment
The most frequent mistake is trying to do too much at once. I recommend a gradual approach:
Identify your team's most repetitive task
The one that takes the most time and requires the least judgement. That is the first candidate for automation or AI improvement.
Test with a standard tool
ChatGPT or Claude can solve most cases without custom development. Spend a week experimenting.
Train your team
A team that knows how to use AI with criteria gets more out of it than one that improvises. Training does not need to be long: 10 well-focused hours make a clear before-and-after difference. And you can subsidise it with FUNDAE.
If you want to understand how FUNDAE subsidies work so training costs nothing, I recommend our step-by-step guide on subsidised AI training.
Practical training for your team
The difference between a company that experiments with AI and one that uses it as a competitive advantage is usually training. Not everyone needs to be an expert. What the team does need is to understand what AI can do, how to use it safely and what its limits are.
At Delbion we offer two courses designed for SME teams:
Secure AI Application in Business (10h)
Safe use, privacy, responsible prompts, EU AI Act compliance. Ideal for the whole organisation. View programme →
AI Agents in Business: Real Use Cases (10h)
Concrete use cases in sales, operations, HR and finance. Designed for executives and managers who want tangible results. View programme →
Both are 100% online, subsidisable with FUNDAE and we handle all the documentation.
You can view dates, prices and upcoming start dates on our training page.
AI Agents Training
Want your team to start using AI with real criteria?
The AI Agents course shows real use cases by sector. 10 hours, online, subsidisable with FUNDAE. Your team leaves with a concrete action plan.
View AI Agents Course →Your team needs secure AI training
The EU AI Act requires AI literacy for all staff from August 2026. Our courses cover compliance, AI agents and governance. FUNDAE can subsidise 100% of the cost.
