How deep has enterprise AI penetration reached in Vietnam
Vietnam's AI market stood at USD 0.75 bn in 2024 [B-Company 2025], concentrated in IT services, financial infrastructure, and edtech. Of the country's roughly 900,000 registered enterprises, 170,000 (~18%) had deployed some form of AI by year-end [MST 2025]—a figure that understates adoption velocity: 47,000 new enterprises adopted AI in 2024 alone, a 39% YoY increase [MST 2025]. Sector distribution skews toward digitally mature industries: IT (31%), Finance (22%), Education (17%). Manufacturing, retail, and food service—the core verticals for KiotViet and Sapo—remain under-penetrated, which is where the next adoption wave is materialising through AI features embedded directly inside existing SaaS subscriptions.
The implication for SaaS economics is structural. When AI capabilities are sold as add-on modules at incremental pricing, adoption stalls due to SMB willingness-to-pay constraints. When they are embedded in base workflows—auto-classifying invoices, flagging anomalous inventory movements, routing customer tickets—they become invisible infrastructure and drive NRR expansion rather than new-logo growth. Vietnam's leading vendors have read this signal correctly.
Which AI use cases are generating measurable returns today
Six use cases are generating documented returns across the Vietnam SaaS stack. MISA's Agentic AI bookkeeping copilot—auto-classifying entries, reconciling bank feeds, drafting tax filings—shipped Q1-2026 [MISA 2026] and is the most consequential single product release this year: it compresses finance back-office labour in a market where accountants are expensive relative to software costs. Invoice OCR and e-invoice attachment from MeInvoice and Fast delivers approximately 70% labour compression in finance back-office roles [Kira estimates], a return that justifies adoption even at SMB scale.
In F&B, CukCuk and iPOS are deploying menu and inventory optimisation models that predict daily waste and adjust menu mix pricing dynamically. KiotViet's retail SKU intelligence layer—POS transaction data feeding AI pricing recommendations—carries a 6–9 month payback period for stores with 100+ SKUs [Kira estimates]. FPT AI Factory's Vietnamese-native customer-service chatbot delivers 40–60% ticket cost reduction across fintech and telco deployments. FPT Software also applies AI code-generation to offshore contracts for Japanese and US clients, improving developer throughput on legacy-modernisation projects [FPT AR 2024].
What structural barriers constrain AI-driven margin expansion
Three constraints bound the pace at which AI compresses unit economics across the Vietnam SaaS stack. Talent supply remains the most acute: the pipeline of engineers capable of productising foundation-model capabilities in Vietnamese-language contexts is thinner than in comparable ASEAN markets, driving salary inflation that partially offsets gross-margin gains. Local foundation-model compute is improving—FPT and CMC have commissioned AI factory infrastructure—but latency and cost for inference at SMB scale remain above global benchmarks, which matters for real-time POS and customer-service applications. SMB willingness to pay above the workflow baseline is the third constraint: Vietnamese SMB operators absorb AI only when it is invisible inside an existing subscription, not when presented as a separate AI product line at premium pricing. Market participants building AI-first SaaS products will find adoption slower than penetration data suggests; those embedding AI into incumbent workflow loops will capture the near-term economics.
Methodology note: AI adoption figures from Ministry of Science and Technology AI Adoption Survey 2025 (MST 2025). Vietnam AI market size from B-Company Vietnam AI Landscape Report 2025. Use-case labour-compression estimates and payback periods are KIRA in-house analyst estimates based on vendor case studies and primary interviews. MISA Agentic AI shipping timeline from MISA Q1-2026 product announcement. FPT code-gen details from FPT Corporation Annual Report 2024.