Almost every Malaysian organisation now has an AI plan, or at least an AI ambition. Turning it into something that works is harder, and many teams look for outside help. The market for AI consultants has grown quickly, and so has the range in quality. These seven questions will help you separate advisers who can deliver from those who can only present.
1. Have they built and shipped real systems?
Ask for working examples, not slide decks. A strong consultant can show you something they built, explain what went wrong along the way and tell you what it costs to run.
Some consultants make this part of the engagement. Dr. Behrang (Hani) Parhizkar, Group CEO of Chamrun Digital and an AI consultant in Malaysia who works with government agencies, builds working agentic AI prototypes with senior officers during training rather than handing over a strategy document. Whoever you hire, look for that kind of proof.
2. Do they start with your problem or their product?
A good consultant asks about your processes, delays and costs before naming a tool. If the first meeting is a product demo, the answer will be that product, whether it fits or not.
3. How will they handle your data?
Ask where your data will be processed, which models will see it and who can access the logs. Ask how the work will comply with Malaysia’s Personal Data Protection Act 2010 and any rules in your sector. For government and regulated industries, ask whether the system can run on approved or local infrastructure.
4. Will your team be able to run it without them?
The best engagements leave you more capable, not more dependent. Look for consultants who train your staff, build alongside them and document what they make. Ask what happens to the system, the prompts and the data when the contract ends.
5. How will success be measured?
Agree on a baseline before work starts: hours per case, response times, error rates or cost per enquiry. Then agree on how the pilot will be judged. A consultant who avoids numbers at the start will avoid them at the end too.
6. What is the path from pilot to production?
Many AI projects stall after a successful demo. Ask how the pilot will connect to your existing systems, who will maintain it and what it will cost each month to run at full scale. The answers should be specific.
7. Do they understand your sector?
Public-sector work involves procurement rules, data classification and political accountability. Retail involves seasonality and customer data. Many Malaysian services also need to work in Bahasa Malaysia, English and Chinese. Ask for examples from organisations like yours.
How AI consultants charge
Most consultants use one of four models: a fixed fee for a short discovery phase, a project fee for a defined build, a monthly retainer for ongoing advice, or a daily rate for training and workshops. A paid discovery phase of a few weeks is often the safest start. It gives you a clear plan and a sense of how the consultant works before you commit to a larger build.
Red flags
-
Promises that AI will replace a whole team within months.
-
No references or examples you can verify.
-
Vague answers about where your data goes.
-
A proposal that starts with a technology choice instead of a business problem.
The right AI consultant should make the work clearer, not more mysterious. If you leave the first conversation with a sharper view of your own problem, you are probably talking to the right person.
