There is a new line appearing in business budgets across Bangladesh. AI.
It may be hiding inside the marketing budget, technology budget, customer-service budget, HR budget or consulting budget.
Some companies are paying for ChatGPT, Gemini, Copilot and other AI platforms. Some are building chatbots. Some are experimenting with AI-generated campaigns. Some are training employees. And some are still trying to understand what everyone else is talking about.
But there is one question that eventually reaches every boardroom:
What are we actually getting for all this AI investment? That may be the most important AI question facing business leaders today. Not whether to use AI. But whether we are using it well enough to change the business.
The AI Experiment Has Already Begun
AI adoption in Bangladesh is no longer something happening only inside technology companies. Marketing teams are generating content. Sales teams are analysing customer data. Finance teams are automating reports. HR teams are screening information and drafting communications. Customer-service teams are experimenting with conversational AI. Executives themselves are using AI to prepare presentations, analyse documents, test ideas and challenge assumptions. The technology is spreading faster than most organizational policies can keep up.
Bangladesh’s draft National AI Policy acknowledges that the country is at a pivotal point in AI adoption, while identifying talent shortages, infrastructure gaps, data governance and cybersecurity as important challenges. That creates an interesting paradox. AI adoption may be happening bottom-up while AI strategy remains top-down.
Employees are experimenting. But the board may not know where. Teams are subscribing. But procurement may not know how many tools are being used. Managers are saving time. But nobody has calculated how much. This is where AI can become expensive without necessarily becoming valuable.
Questions Every C-Suite Should Be Asking
Rather than asking whether the organization is “using AI”, I would suggest asking eight harder questions.
- Where is AI actually creating measurable value?
If a marketing team produces a presentation 50% faster, that is interesting. If the same team uses AI to improve conversion, reduce campaign wastage or accelerate product launches, that is business value. The board should distinguish between activity and impact. Because generating more content is not automatically generating more business.
- Are we buying too many tools?
The AI marketplace has created an unusual corporate problem. Too much choice. Different employees may be using different platforms for similar tasks. One department has one subscription. Another has another. A third is using a free version. Eventually, the company may discover that it has built an unofficial AI ecosystem without building an AI strategy. The question is not: “How many AI tools do we have?” It is: “Which capabilities do we actually need?”
- Is our data ready for AI?
This may be the least glamorous part of the AI conversation—and one of the most important. AI cannot compensate for poor data indefinitely. If customer information sits across disconnected systems, product databases are incomplete, sales information is inconsistent, or different SBUs define the same customer differently, AI will simply process the confusion faster. Garbage in, garbage out was yesterday’s technology warning. Today, it is a boardroom issue.
- Who owns AI?
Is it the CIO?
CTO?
CMO?
CHRO?
Business heads?
Or the CEO?
The answer may be: all of them.
AI is not another software installation. It affects how people work, how customers are served, how products are developed, how risks are managed and how capital is allocated. That makes AI a business transformation issue. Not an IT project.
The Bangladesh Question: Productivity
This is particularly important for Bangladesh. As companies face pressure to become more competitive, improve productivity and operate with greater efficiency, AI offers an opportunity to do more with the resources we already have.
But we should be careful with the word “productivity.” If AI allows an employee to write ten reports instead of five, have we doubled productivity? Perhaps. But what if the business only needed three? Then we have increased output without necessarily increasing value.
The real opportunity lies deeper. Can AI reduce the time from consumer insight to product launch? Can it help a sales team identify the customers most likely to buy? Can it forecast demand more accurately? Can a manufacturing business predict equipment failure? Can a bank identify unusual transactions earlier? Can a retailer reduce inventory wastage? Can a customer-service operation resolve problems without increasing headcount? Those are business questions. And those are questions the board should be asking.
Then Comes the Data and Cybersecurity Problem
There is another conversation that needs to happen before AI moves deeper into our organisations. What information are we giving it?
In June 2026, Bangladesh Bank instructed its officials not to put confidential or sensitive banking information into public AI platforms and required approval for certain official AI uses.
The message is relevant far beyond banking. A marketing executive may upload a confidential product launch plan. A finance executive may paste sensitive numbers. An HR executive may upload employee information. A sales executive may share customer data. The intention may be completely innocent. The risk may not be.
As AI becomes embedded in everyday workflows, companies will need clear rules around data classification, approved tools, access, human review and accountability. AI governance cannot arrive six months after AI adoption. It has to arrive with it.
The People Problem
There is also a human dimension. Some employees embrace AI. Some are indifferent. Some are quietly worried. And some may be using it every day without telling their managers. That last group should concern leaders—not because employees are doing something wrong, but because it may indicate that the organization has not created a safe framework for experimentation.
Fear rarely produces innovation. Neither does forced adoption. People need to understand:
What can I use AI for?
What shouldn’t I use it for?
What happens if AI gets something wrong?
Who checks the output?
And perhaps most importantly:
How will AI change my role?
The answer should not always be “It won’t.” Because it probably will.
The better answer is: “Let’s redesign the role so that human capability becomes more valuable.”
From Prompting to Rebuilding Work
We are still fascinated by prompts. Which prompt should I use? How do I get a better answer? How can I generate a better image? These are useful questions. But they represent only the beginning. The bigger opportunity is to redesign workflows. Instead of asking an employee to use AI for one task, ask:
Can AI become part of the entire process?
From research to briefing.
From briefing to creative development.
From customer inquiry to resolution.
From sales data to forecasting.
From market intelligence to management reporting.
This is where AI agents and automated workflows could become more significant.
The future may not be an employee asking AI one question at a time. It may be an AI-enabled workflow continuously performing defined tasks—with humans setting direction, applying judgment and taking accountability.
The ROI Conversation
Eventually, every CFO will ask the question. “How much did we spend on AI?”
That is a useful starting point. But it is not enough. The better questions are:
What can we do now that we could not do before?
What became faster?
What became cheaper?
What became more accurate?
What customer experience improved?
What revenue opportunity did we unlock?
What risk did we reduce?
And perhaps the most important: What did our people become capable of doing because AI was introduced?
This is where AI moves from being a technology expense to becoming a strategic capability.
The Boardroom Should Not Wait for Perfect AI
Bangladesh is already moving toward a formal national AI framework. The current policy process emphasizes responsible, inclusive and innovation-driven AI, while recent discussions involving government and UNESCO have highlighted human-centered AI, data governance and information integrity. Businesses should move with similar discipline. We do not need to wait until AI is perfect. But we should stop treating experimentation as strategy.
Start small.
Measure.
Learn.
Build guardrails.
Scale what works.
Stop what doesn’t. And keep the human being accountable for the decisions that matter.
The Chronicle Continues
Every major technology wave eventually creates the same temptation. We become fascinated by the technology itself. We forget the problem we were trying to solve. Bangladesh’s AI journey should not become a race to see who can deploy the most tools. It should become a race to see who can create the most value from them. The winners may not necessarily be the companies with the largest AI budgets. They may be the companies that understand where AI belongs—and where it doesn’t. Because the future of AI in business will not ultimately be measured by the number of prompts we write. It will be measured by the number of meaningful business problems we solve.
And perhaps the most important question for every C-suite today is remarkably simple: If we stopped paying for AI tomorrow, what measurable business capability would we actually miss?
If the answer is difficult to articulate, perhaps the AI strategy needs another board meeting.
Author: TOWFIQUE HASAN FCIM
Chief Business Development Officer
Akij Resource
