AI in SMEs: 5
productive use cases.

Most AI projects in SMEs fail because they start too ambitiously. Building your own language AI? Hopeless. A ChatGPT integration for everything? Burns money. Here are the 5 applications we have successfully implemented with SMEs — with measurable ROI.

1. Incoming invoice OCR with booking suggestion

What it does: Incoming invoices are scanned or captured via email. An AI extracts supplier, amount, VAT rate, booking account, and suggests a completed booking. The employee only needs to confirm.

ROI: With 500 documents/month, this saves 25–40 hours of accounting time. Based on a half-time employee, approximately €1,500–2,500 in personnel costs/month. Setup typically costs €8,000–15,000 one-time, with amortization in 4–8 months.

Risks: Minimal — booking suggestions are confirmed. There may be training needs for very individual accounts.

2. Contract analysis for compliance & risk

What it does: Incoming contracts (supply, service, NDA contracts) are automatically checked for risk clauses: liability limitations, notice periods, presence obligations, exclusivity exclusions. The result is presented in a traffic light report.

ROI: For B2B medium-sized companies with 200–500 contracts/year, this replaces half a lawyer. Setup costs €12,000–25,000, with amortization in 6–12 months.

Risks: AI verification is not legally binding — it provides a good preliminary check, but a lawyer must review decisions ready for signature.

3. Customer inquiry triage

What it does: Incoming emails and tickets are automatically categorized (offer, complaint, technical question, routine). Frequent inquiries are answered directly (with confirmation), complex ones routed to the appropriate department.

ROI: With 1,000 inquiries/month: 30–50% volume reduction for processing. Response time is halved. Setup costs €10,000–20,000, with amortization in 6 months.

Risks: Tone must match the brand, ensure escalation in sensitive cases.

4. Custom RAG assistants (Retrieval-Augmented Generation)

What it does: An AI assistant that accesses your own documents (contracts, manuals, FAQs, knowledge base) and answers employee questions. "What does our supply contract with Schmitz Cargobull say about penalties?" in 5 seconds instead of 30 minutes.

ROI: For each employee with knowledge needs, 30–60 minutes/day time savings. With 50 employees and a €50 hourly rate: €1,000–2,000 daily. Setup costs €15,000–40,000, with amortization in 2–4 months.

Risks: Data silo consolidation is necessary in advance. Updates are required when documents change.

5. Forecast models for sales and inventory

What it does: Based on sales history, seasonality, marketing activities, and external factors (weather, economic conditions), expected sales are predicted. Orders are optimized automatically.

ROI: For consumer goods SMEs, typically 15–25% fewer out-of-stock situations, 10–20% less inventory. For a €5 million warehouse: €500,000–1,000,000 released liquidity.

Risks: Data quality is everything. A clean sales history over 24+ months is a prerequisite.

What DOES NOT work (in SMEs)

• In-house LLM development: Too expensive (8 digits), too slow, without competitive advantage.
• 'Let AI run my business model': AI is a tool, not a substitute for strategy.
• ChatGPT with customer data: GDPR issue, data leakage. An own EU-compliant solution is necessary.
• Generic marketing content: Works briefly, long-term it damages the brand.

How we proceed at TABAK

Every AI project starts with a 2-week readiness assessment: Which use cases have the highest impact in your operation? Where is the data? What does the solution cost? Based on the analysis, we jointly decide on a pilot — and after 6–8 weeks, you will see the first productive results.