Kamal Ghorui
Gone are the days when clients provided developers with the requirements of a solution. If we are not consultative, we are not in the game anymore. If we don’t foresee business challenges and help our customers solve their problems, we should go back 20 years.
All of us have come across the joke where one follows the elaborate recipe instructions for Payasam step by step, but doesn’t have the Payasam ready because the instructions don’t start by asking us to switch on the stove. A developer does not need to be told that Payasams can’t be made without heating. Today, we are required not only to know that the stove needs to be switched on but also to advise on alternatives to make the Payasam tastier.
Non-AI solutions are equally important
There’s a high chance customers will ask you to use AI, especially Generative AI, to solve their problems. They are not at fault. They are also under pressure not to miss the AI bus. However, an AI consultant must be able to recommend a non-AI solution if the business problem can be solved as such. AI cannot be used as an approach unless the organisation’s data foundation is strong.
The AI vision
Frame an AI vision that cannot be misaligned with the organisation’s vision and values. The AI strategy must be built upon the data strategy. AI goals must help achieve business goals, not the other way around. An AI vision for an organisation is a guiding strategic document that outlines how the company plans to leverage artificial intelligence to achieve its goals, create value, and drive transformation. It typically includes a concise core statement (one or two sentences) plus supporting details on purpose, principles, target impacts, and ethical guidelines.
The AI programs
The consultant must identify AI programs based on the data captured and available. The programs or sets of use cases must include appropriate prioritisation and feasibility studies, and must determine the set of tools to be utilised based on risk assessment.
Measuring ROI
One must predefine how to measure the value generated by AI projects. At least, we must know whether the AI initiative achieved its objectives. Business ROI (Return on Investment) metrics must be set. Having said this, attributing revenue generation or cost savings to an AI project is easier said than done.
Adoption of an AI solution
The consultant and the customer must deliberate on ways to increase adoption of the AI solution. A solution that is brilliantly developed and deployed but no one uses is one of the worst kinds of solutioning. Resistance to change, especially when it comes to job losses, is not surprising. We have seen that everywhere and every time.
AI risks
Risk identification and mitigation are not optional. We must assess and be aware of bias, ethical issues, the costs of inaccuracies, exposure of sensitive information, and inappropriate content generation. When the dataset on which AI models are trained inherently under- or over-represents a specific class of people (for example), predictions will be biased. Technical approaches to handling imbalanced datasets must be employed.
Additional data capture
The consultant must create a roadmap to capture additional data and enrich the organisation’s data foundation. The additional data can help enhance existing solutions and can enable entirely new AI solutions. This is one of the most overlooked tasks in any consulting assignment or customer discussion. Unless we think long-term, alongside short-term gains, we will hit the limits of what we can do with existing datasets.
The future belongs to consultants who think beyond algorithms—those who blend empathy, strategy, and technology to create lasting impact. True AI success lies not in deploying tools but in shaping data-driven, ethical, and adaptive ecosystems that empower businesses to evolve continuously. If your focus is only on coding and developing applications, you are already replaced. Add “software applications” to your quick commerce website alongside groceries. I will deliver it to you within ten minutes.
END OF ARTICLE
