AI for SMEs: How to Get Started Right

TL;DR: SMEs that implement AI pragmatically save an average of 5-10 hours per employee per week. The key: don’t start with the most expensive tool, start with the most concrete problem. A pilot project with a clear goal delivers more than a 50-page AI strategy.

Artificial Intelligence is no longer just for large corporations. Small and medium enterprises can also benefit from AI, if they start right. According to research from the OECD, fewer than 25% of SMEs across Europe currently use AI tools in their daily operations, despite the potential being relevant for all business sizes.

Where to Begin?

The biggest mistake: buying the most expensive tool right away. Instead, I recommend a pragmatic approach:

Analyze processes: Where do your employees spend the most time on repetitive tasks?
Identify quick wins: Which of these tasks can be automated with existing AI tools?
Start a pilot project: Start small, gather experience, then scale.

A concrete real-world example: A Swiss trading company with 35 employees started using ChatGPT to pre-structure incoming customer inquiries by email. Processing time dropped from an average of 8 minutes to 3 minutes per inquiry. With 60 inquiries daily, that translates to 5 hours saved per day, roughly one full-time equivalent per month.

Typical Use Cases

The following overview shows where SMEs achieve the fastest return on investment:

Area	Time saved/week	Entry barrier	Recommended tool
Customer communication	3-5 hrs	Low	ChatGPT, Claude
Document processing	2-4 hrs	Medium	Make, Zapier
Data analysis	1-3 hrs	Medium	Copilot, ChatGPT
Content creation	2-6 hrs	Low	ChatGPT, Canva AI

Customer communication: Chatbots and automated email responses reduce response times from hours to minutes
Document processing: Automatic classification and data extraction from invoices, contracts, forms
Data analysis: Pattern recognition in sales data and customer behavior, without data science expertise
Content creation: Texts, translations, social media content in minutes instead of hours

The AI Adoption Process at a Glance

Week 1-2:   Process Audit
             Which tasks consume the most time?
                    |
Week 3:     Tool Selection
             Which AI tool fits the problem?
                    |
Week 4-6:   Pilot Phase
             Small team, clear metrics, fixed deadline
                    |
Week 7-8:   Evaluation
             Before vs. after, honest assessment
                    |
From Week 9: Scale or Pivot
             Either roll out broadly or choose a different approach

Common Stumbling Blocks

From consulting conversations with dozens of SMEs, I know the typical obstacles:

Data privacy: Many companies hesitate because they’re unsure whether they can enter customer data into AI tools. The answer is nuanced: anonymized examples and internal texts are generally unproblematic. For personal data, Microsoft Copilot (within the M365 environment) or local open-source models offer good alternatives.

Employee resistance: AI tools are often perceived as a threat. Open communication that shows which tedious tasks will disappear removes most skepticism.

Perfectionism: Waiting for the perfect tool to be available blocks the start. Today’s tools are good enough to produce measurable results.

The Most Important Tip

Technology is only as good as the people who use it. Invest in training: your employees are the key to success. A half-day workshop with concrete exercises tailored to your business delivers more than an annual subscription to a tool nobody understands.

Conclusion and Next Steps

Start this week with a simple experiment: take a task your team regularly handles and try to replicate it with ChatGPT or Claude. Note the time it takes and compare it to the usual effort. This first step costs nothing and shows in 30 minutes whether AI is relevant for your use case.

Want to introduce AI in your organization? Get in touch for a free consultation.