Question 1

AI transformation for the CEO: first use case, ROI and accountability

Accepted Answer

The first AI transformation decision is the use case: pick one with a clear goal, usable data already in hand and a tolerable cost of failure, such as quality inspection or predictive maintenance. The second is governance first: set up an AI policy, an AI system inventory and a review process before deployment, and set KPI baselines beforehand so the return can be tracked. The third is accountability: make clear who is responsible for human oversight and for the decision to switch a system off.

Question 2

How should we choose the first AI use case and calculate the return?

Accepted Answer

Rank candidates on three conditions, not on how new the technology is:
Condition / Question to ask
Data availability / Do we already have usable data? Are quality, access and definitions consistent?
Business benefit / Which metric improves? What is the baseline before deployment?
Cost of failure / Can we absorb the loss if it goes wrong? Can we fall back to a manual process?
Good starting points usually have a clear goal, ready data and a low cost of failure, such as AI quality inspection or predictive maintenance.
The key to tracking the return is setting KPI baselines before deployment and agreeing on a review frequency; without a baseline you cannot show afterwards what the AI investment delivered.

Question 3

Should we buy AI capability, co-develop it or build it in-house?

Accepted Answer

The three models can be combined as needed:
Model / Fits when / Clarify before signing
Buy a mature AI solution / Generic need, fast go-live wanted / Where data is stored, whether the vendor can use your data, whether you can exit later
Co-develop with a technology vendor / Specific need, no in-house data science / Ownership of models and data, acceptance criteria, who runs it after go-live
Build in-house capability / AI is a core competitive edge needing long-term iteration / Talent development, data foundation and governance processes
You can do AI transformation without data scientists; the point is to put data rights, exit terms and acceptance criteria into the contract and have legal review it.

Question 4

Will "governance first" slow us down? What is the minimum?

Accepted Answer

The depth of governance depends on the use:
Staff using general generative AI tools for everyday documents with no confidential or personal data: internal usage rules and training are usually enough.
AI that affects decisions about customers or employees, processes personal data, or goes into products for the EU market: formal governance and risk assessment are needed.
Formal governance is set up before deployment: an AI policy, an AI system inventory, security requirements for AI development and procurement, and an algorithm review process, which can map to an ISO/IEC 42001 AI management system. Risk assessment can follow ISO/IEC 23894 or the US NIST AI Risk Management Framework (govern, map, measure, manage).
Classifying before investing avoids over-governing low-risk uses and leaving gaps in high-risk ones.

Question 5

If AI makes a wrong decision and causes a loss, who is responsible?

Accepted Answer

Inside the company, responsibility must be written down in advance: who monitors model performance and drift, and who is responsible for human oversight and the decision to switch it off. When AI can issue refunds, commit delivery dates or change orders on its own, human oversight cannot be a spot check after the fact; define the scope of authority first.
External legal liability:
The EU AI Act itself contains no civil damages rules; breaching duties such as human oversight leads to administrative fines
The European Commission has withdrawn the proposed AI Liability Directive; harm caused by AI is mainly handled under the new Product Liability Directive (EU) 2024/2853 (covering software and AI systems, for products placed on the market from 9 December 2026) and national tort law
Allocate liability with AI vendors in the contract. Liability in a specific case is for legal counsel to assess.

Question 6

How do Taiwan's AI Basic Act and the EU AI Act affect our AI transformation?

Accepted Answer

Taiwan: the Artificial Intelligence Basic Act was promulgated and took effect on 14 January 2026, with the National Science and Technology Council as the central authority; most provisions govern the government, and concrete requirements for companies will mainly come through risk-based rules and industry guidelines from each sector regulator, which the Executive Yuan has asked to be issued before January 2028. The AI risk classification framework version 1.0, published by the Ministry of Digital Affairs on 7 July 2026, can be used as a reference for inventory and assessment.
EU: a Taiwanese company is within the EU AI Act when it places an AI system on the EU market or when the AI's output is used in the EU (Article 2(1)). First determine whether you are provider or deployer for each AI system and its risk level; Annex III high-risk systems are covered from 2 December 2027.

Question 7

Are there tax incentives for AI investment?

Accepted Answer

Article 10-1 of Taiwan's Statute for Industrial Innovation was amended in 2025 to add artificial intelligence and energy saving and carbon reduction, keeping smart machinery, 5G systems and cybersecurity. When the Industrial Development Administration issued the implementing rules on 27 November 2025, it gave examples of the new items such as inference servers with AI computing power and AI edge computing equipment; licensed software recognised as an intangible asset is also included.
Period: 1 January 2025 to 31 December 2029; only orders placed from 1 January 2025 qualify for the new items
Cap: the amount eligible for application rises from NT$1 billion to NT$2 billion
Procedure: delivery or technical service completed within 2 years from the day after ordering, with the credit claimed in the year of delivery
Whether specific equipment or services qualify and how the credit is calculated are for your statutory auditor to judge under the implementing rules and the facts.

Question 8

How does Winners Consulting help the CEO drive AI transformation?

Accepted Answer

VP- and director-level consultants lead the work end to end, never outsourced, with governance first: an ISO 42001 AI management system and EU AI Act compliance built alongside the AI technology.
Use cases and risk assessment: map core business processes, identify use cases such as predictive maintenance, quality inspection, customer service automation and sales forecasting, and assess each AI use's risk level.
Governance first: AI ethics principles, AI system inventory management, security requirements for AI development and procurement, and an algorithm review process before deployment.
Data: a management mechanism for AI training data.
Post-launch monitoring and regulation: model drift detection and continuous improvement, with an assessment of EU AI Act applicability.
Results and capability: baselines set before deployment; help choosing between buying, co-developing and building.
The service page lists a usual consulting period of 7–12 months or more. Read more: AI transformation consulting, Choosing an AI transformation consultant, ISO 42001 implementation, Choosing an ISO 42001 consultant.
Request a free mechanism diagnosis