What should SMEs prepare before adopting AI?
You can't just buy AI and expect it to work. Five concrete steps to get your data, processes and team ready first.
After 3 years advising 50+ Vietnamese SMEs, I've learned: most AI projects fail not because of the technology — but because the business wasn't ready. Here are 5 practical steps that have worked for us.
1. Start with the process, not the technology
The wrong question 80% of business owners ask: "What AI should I buy?" The right question: "Which process costs us the most time?" Spend a week observing and you'll see immediately: invoice data entry, reconciliation, weekly reports, templated customer responses. Each is a clear AI use case.
2. Data must be ready — and you have more than you think
AI needs data to learn. But you don't need big data — just enough samples to recognise patterns. For document OCR, 50-100 real samples are enough to start. For an internal AI assistant, a 30-question FAQ + 100 real conversation snippets is enough to fine-tune. Don't wait for "enough data" — start and let the AI learn over time.
3. Pick one process, don't try five at once
The most common mistake: wanting AI to do everything. Result: nothing gets done. Our rule: 1st AI project = 1 process + 1 team responsible + 1 clear KPI. When the first one runs smoothly (usually 4-8 weeks), then scale.
4. Integrate with existing software, don't create new ones
Most Vietnamese SMEs use MISA, KiotViet, Sapo, or Excel. AI doesn't need to replace them — just integrate. OCR reads documents → pushes to MISA via API. AI assistant reads Excel → replies on Zalo. The less you change existing workflows, the higher the adoption.
5. Measure from day one, not at the end of the project
Before starting, measure the baseline: how many hours/month on this process, how many errors, what cost. After 4 weeks running AI, measure again. Without before/after numbers, you won't know if AI actually helps — and the team will lose motivation.
AI is not magic. It's a tool — and tools only work when users know how to use them. For Vietnamese SMEs, the key is: start small, measure honestly, and scale from what's already working. That's how we've helped 50+ Vietnamese businesses save billions of VND in operating costs every year.