How to Screen for Real AI Skills When Hiring in Vietnam

Answer first: a list of AI tools on a CV tells you what a candidate has tried, not what they can deliver. To screen for real AI skills when hiring in Vietnam, ask every candidate the same three things: what result changed, at what scale, and how they checked the AI’s output. Then score the answer on five criteria (measurable, size of change, scale, command and control, sustained) and ask for evidence you can open yourself, such as a repository, a portfolio or a short work test.

The pay-off is real. In ManpowerGroup’s Q3 2026 survey of 265 employers in Vietnam, 83% said they are willing to offer better pay to candidates who use AI tools effectively and responsibly (ManpowerGroup, 2026). “Effectively and responsibly” is exactly the part a keyword filter cannot see. This guide shows how to test it.

Why do AI tool names on a CV say so little?

Answer first: almost every candidate can list the same tools, but employers pay for the outcome and the judgement behind it.

The market rewards AI skills. PwC’s 2026 Global AI Jobs Barometer, built on more than one billion job ads, found that jobs requiring specific AI skills grew 69%, against 9% for the overall jobs market, and the average wage premium for AI skills reached 62% (PwC, 2026). Lightcast’s analysis of over 1.3 billion postings found that job ads including AI skills offer 28% higher salaries (Lightcast, 2025).

The same research points to judgement, not tools. PwC reports that AI is raising the emphasis on human skills such as judgement, creativity and leadership (PwC, 2026). In Vietnam, ManpowerGroup found that 75% of employers also value AI model and application development skills (ManpowerGroup, 2026). Both signals describe what a person does with AI. A line such as “ChatGPT, Copilot, Midjourney” does not describe either.

Practical consequence. If you filter by tool names, you reward the candidates who write the longest lists. If you ask for results, you find the people who change how work gets done.

What does BlockchainWork’s CV data show about AI skills in Vietnam?

Answer first: AI skills are already common in the Vietnamese candidate pool, and the number of candidates who can show measurable AI results is growing fast.

BlockchainWork is a Vietnamese blockchain, Web3 and AI talent community with 25,000+ candidates, 350+ hiring companies, 12+ educational institution MOUs and 200+ media mentions. Our CV data from July to October 2026 shows:

• More than 1,100 new CVs in October 2026, clearly above the monthly average of Q3 2026.
• About 60% of CVs list AI skills, and more than half of those are engineers who directly use and build AI features.
• Nearly half of profiles include a GitHub, portfolio or product demo link.
• The share of profiles with AI projects that report measurable results rose from 5.8% (Q3) to 9.9% (October).

Source: BlockchainWork CV data, July-October 2026.

What this means for an employer: AI skills alone no longer set a candidate apart, so the question to screen on is impact. The pool is moving in that direction. Candidates increasingly write results, not just tools, and many already attach links you can check before the first interview.

Which five criteria separate real AI impact from a keyword list?

Answer first: score each AI example a candidate gives, not the candidate as a person, on five criteria. A strong example scores well on all five.

Lightcast found that 51% of job postings requiring AI skills in 2024 were outside IT and computer science occupations (Lightcast, 2025). The criteria below therefore work for engineers, marketers, operations staff and analysts alike.

Criterion	What a weak answer sounds like	What a strong answer includes
Measurable	“AI made me faster”	A before and after figure in the same unit and scope, such as hours per module or tickets per day
Size of change	No figure, or a figure with no baseline	A clear change from the baseline, such as twice the output or half the time
Scale	The same task, done the same way	More channels, projects or languages, or work that used to need more people
Command and control	“I use AI every day”	How they instruct the AI, how they check its output, and an example of an AI error they caught
Sustained	A one-off experiment	A habit kept for three months or more and turned into a team process

A simple way to use the table: rate each criterion 0 (missing), 1 (partly there) or 2 (clearly there). An example that scores 2 on “command and control” but has no evidence still deserves a follow-up question, not an automatic pass.

How do you read an AI-impact line on a CV?

Answer first: look for six parts in one line: the task, who did what (person and AI), before and after, scale, how the output was checked, and evidence.

BlockchainWork suggests this formula to candidates: [task] + [my role, the AI’s role] + [before → after] + [scale] + [how I checked] + [evidence]. You can read CVs against the same formula. The two lines below are illustrations written to show the difference, not taken from any candidate profile or BlockchainWork data:

• Weak: “Proficient in ChatGPT, Claude and GitHub Copilot. Used AI to improve productivity.”
• Strong: “Smart contract testing: I specified the test cases and reviewed results; AI generated the test code. Went from 40 to 120 tests per module and from 3 days to 1 day, across 3 projects. AI missed a reentrancy bug, which I caught with fuzzing before audit. (Public pull request link)”

The strong line gives you five interview questions for free: Which projects? How did you measure? What did the AI get wrong? Who reviewed your work? Can I see the pull request?

Candidates are moving this way. In BlockchainWork’s data, the share of profiles with AI projects that report measurable results rose from 5.8% in Q3 to 9.9% in October 2026 (BlockchainWork CV data, July-October 2026). A CV that already follows the formula is a strong signal; a CV that does not can still be tested with the questions below.

Which interview questions reveal how a candidate really works with AI?

Answer first: ask for one concrete story and then walk through the five criteria. Good candidates answer with numbers, names of checks and specific mistakes.

ManpowerGroup’s wording, “effectively and responsibly” (ManpowerGroup, 2026), and PwC’s emphasis on judgement (PwC, 2026) both point to the same interview goal: find out whether the person stays in charge of the result. Sample questions:

• Measurable: “Pick one task where AI changed your output. What was the number before, and what is it now? How did you measure it?”
• Size of change: “What would this task look like today without AI? What part of the gain came from AI and what part from other changes?”
• Scale: “Did AI let you take on more work, new channels or new projects? What did you stop doing, or hand over, as a result?”
• Command and control: “Show me how you instruct the AI for this task. What do you always check by hand? Tell me about a time the AI was wrong and how you found out.”
• Sustained: “How long have you worked this way? Did anyone else on your team adopt it?”
• Responsibility: “What information do you never put into an AI tool? How do you handle confidential code or client data?”

Listen for specifics. “I check every figure and every proper name before publishing, and last month that stopped two errors” is a strong answer. “I always double-check” is not.

What evidence can you verify before making an offer?

Answer first: ask for something you can open yourself, then confirm it with a short task or a reference.

Evidence comes in four levels, from weakest to strongest:

Level	What it is	How to check it
Self-reported	Figures written on the CV only	Check that units and scope make sense; ask follow-up questions
Verifiable link	GitHub repository, pull request, published work, portfolio, product demo, certificate with a verification code	Open the link and confirm the work matches the claim
Work test	A 30-45 minute task of the same kind, with AI tools allowed	Watch how the candidate instructs the AI and checks its output
Third-party confirmation	A former manager, client or your own probation review confirms the result	Reference call, or review after the first months

Links are often already there. Nearly half of the profiles in BlockchainWork’s data include a GitHub, portfolio or product demo link (BlockchainWork CV data, July-October 2026), so the verifiable level is often a click away.

AI-allowed work tests are now mainstream. CodeSignal launched AI-assisted coding assessments in May 2025 that give hiring teams session replays and transcripts of every AI interaction during the test (CodeSignal, 2025). HackerRank’s AI-assisted interviews let interviewers watch in real time how candidates work with an AI assistant and review the full chat transcript afterwards, which HackerRank presents as insight into a candidate’s AI fluency (HackerRank, 2026). You do not need a paid platform to copy the idea: give a short, realistic task, allow AI tools, and ask the candidate to share their screen or their prompt history.

For a full view of the Vietnamese hiring process, from roles to onboarding, see our Hire Blockchain Developers in Vietnam: 2026 Employer Guide and Onboarding a Vietnamese Engineering Team: 2026 Guide. Market data for Q3 is in the Vietnam Blockchain Talent Report 2026.

Looking for engineers in Vietnam who can show real AI results? Register your company at blockchainwork.net/employer-signup to reach 25,000+ candidates across blockchain, Web3 and AI roles, or email our team at contact@blockchainwork.net to discuss the AI skills your role needs.

Frequently asked questions

1. Should I filter CVs by AI tool names?

No. Tool names show what a candidate has tried, not what they delivered. Use them only as a starting point and screen on results, scale and how the candidate checks AI output.

2. What is the fastest way to check whether an AI claim on a CV is real?

Open the evidence. Nearly half of the profiles in BlockchainWork’s CV data (July-October 2026) include a GitHub, portfolio or product demo link. If there is no link, ask for a short work test with AI tools allowed.

3. Should candidates be allowed to use AI in technical tests?

For most roles, yes, because that is how they will work. Platforms such as CodeSignal (2025) and HackerRank (2026) now record AI interactions so interviewers can review how the candidate used the assistant.

4. Do these criteria apply to non-engineering roles?

Yes. Lightcast found that 51% of job postings requiring AI skills in 2024 were outside IT and computer science (Lightcast, 2025). A marketer or operations lead can show before and after figures, scale and quality checks just like an engineer.

Sources

• ManpowerGroup Vietnam, Employment Outlook Survey Q3 2026, reported by Nhan Dan (2026)
• PwC, 2026 Global AI Jobs Barometer (2026)
• Lightcast, Beyond the Buzz: Developing the AI Skills Employers Actually Need (2025)
• CodeSignal, CodeSignal Launches AI-Assisted Coding Assessments and Interviews (2025)
• HackerRank, AI-Assisted Interviews (2026)
• BlockchainWork, BlockchainWork CV data, July-October 2026 (2026)

BlockchainWork

** This content does not constitute investment advice, investment strategy recommendations, or sponsored advertising. Please refer to our Disclaimer and use the information at your own discretion.

BlockchainWork – Vietnam’s leading blockchain talent community, connecting passionate candidates with pioneering companies in blockchain and Web3. With 25,000+ registered candidates, 12+ educational institution MOUs and 350+ hiring companies, BlockchainWork is the specialized platform for blockchain careers, from Vietnam to global opportunities.

• Partnership inquiries: contact@blockchainwork.net
• Learn more: blockchainwork.net
• Latest blockchain jobs: blockchainwork.net/viec-lam-moi-nhat
• Employers, connect with top talent: blockchainwork.net/employer-signup
• Candidates, find your next opportunity: blockchainwork.net/candidate-signup
• Join the community: Vietnam Blockchain Community – BlockchainWork

>> You may also like:

• Hire Blockchain Developers in Vietnam: 2026 Employer Guide
• Vietnam Blockchain Talent Report 2026: Q3 Job Market Trends
• Why Japanese Firms Hire Vietnamese Engineers (2026 Data)
• Onboarding a Vietnamese Engineering Team: 2026 Guide
• Build, Rent or Hire a Blockchain Team in Vietnam (2026)