AI Agent Development Services We Offer in Ho Chi Minh City
AI agent demand in Ho Chi Minh City splits between three buyers, and we ship for all of them. Vietnamese fintechs (MoMo, VNPay, ZaloPay, Cake Digital Bank, Timo, TPBank LiveBank) deploy customer service and KYC agents handling Vietnamese-first conversations under SBV digital banking regulations and PDP Decree 13/2023. E-commerce platforms (Tiki, Lazada Vietnam, Shopee Vietnam, Sendo) run agentic merchandising and customer support that must coordinate with Vietnam's e-invoice mandate and the Cybersecurity Law 2018 data localisation rules. Enterprise IT teams at Vingroup subsidiaries, Masan Group, Saigon Co.op, and FPT internal operations pilot agent copilots for procurement, internal IT, and HR workflows in Vietnamese with optional English fallback. Our agents ship with proper Vietnamese language model selection (Sailor, VinaLLaMA, PhoBERT for classification, plus GPT-4 and Claude when cross-border is acceptable), structured tool use, and evaluation harnesses tuned for Vietnamese diacritic handling and Saigon-versus-Hanoi dialect variations.
Our AI Agent Development Development Process
Discovery opens with a Vietnamese language coverage workshop, a SBV regulatory classification if financial workflows are in scope, and a Cybersecurity Law 2018 data residency assessment to determine whether the agent must run on FPT Cloud sovereign infrastructure or whether AWS Singapore (ap-southeast-1) is acceptable for the workload. Build sprints are two weeks, running on ICT (UTC+7) standups with Vietnamese-language QA pass at each milestone. We pair Vietnamese linguists for prompt engineering and evaluation harness curation because most agent failures in Vietnamese production come from diacritic mishandling, dialect ambiguity (Northern Hanoi versus Southern Saigon vocabulary), and code-switching with English that off-the-shelf evals miss. Deployment includes monitoring, prompt injection defences, hallucination logging, and a documented rollback plan that SBV and MIC (Ministry of Information and Communications) auditors accept without a second vendor review.
Process Discovery
1-2 WeeksWe sit with the people doing the work in {city} and record the real process — including the exceptions they handle by instinct, which are exactly what kill naive automations.
Tool Surface Design
1-2 WeeksEvery system the agent touches gets a typed, permission-scoped tool with its own rate limit and rollback path. The agent gets a narrow set of verbs, never raw admin access.
Build & Evaluate
3-6 WeeksThe agent is built alongside its evaluation suite from day one, using real tasks from your business with verified outcomes. Every change is scored before it ships.
Shadow Mode
2-3 WeeksThe agent runs against live traffic but commits nothing. We compare its proposed actions to what your team actually did and tune until agreement is high enough to trust.
Staged Autonomy & Run
OngoingAutonomy is released by risk band — reversible actions first, irreversible ones keeping a permanent human gate. Then we monitor completion rate, escalations, latency and spend.
Technologies We Use for AI Agent Development
Agent stacks for HCMC clients almost always need Vietnamese data residency for SBV-regulated or Decree 13/2023-scoped workloads, so we default to FPT Cloud (the Vietnamese sovereign cloud headquartered in Hanoi with HCMC presence) for primary inference, with AWS ap-southeast-1 (Singapore) for non-sensitive auxiliary services where Vietnamese privacy counsel approves the cross-border flow. Agent orchestration runs on LangGraph or CrewAI with Vietnamese-tuned models — Sailor (open-weight Southeast Asian language model from Sea AI Lab), VinaLLaMA (Vietnamese Llama fine-tune from VinAI Research at VinUniversity), PhoBERT and PhoGPT (also from VinAI Research). For frontier capability we use Claude 4.5 Sonnet, GPT-4o, and Gemini 1.5 Pro via APIs when the data classification allows cross-border. Vector databases ship on pgvector inside Vietnamese-hosted Postgres or on managed Pinecone / Weaviate when residency permits. Evaluation harnesses include Vietnamese-specific test sets covering diacritic normalisation, Bac-Trung-Nam dialect handling, and code-switched Vietnamese-English customer inputs.
Other Services We Offer in Ho Chi Minh City
Looking for a different service? Explore our full range of technology solutions available in Ho Chi Minh City.
Explore Our AI Agent Development Specializations
Dive deeper into our specialized ai agent development offerings.
AI Agent Development in Other Cities
We deliver ai agent development solutions across 45 cities in 24 countries. Find a location near you.
Latest Work
Drag to explore or use arrow keys