LALIT SARASWAT looks beyond whether AI has heard of your company, to the more important question of whether it understands what your business does, whom it serves and why a customer should consider it
Try a simple experiment. Open ChatGPT, Gemini or another AI assistant and type the name of your company. Unless your business is very new or has almost no presence online, you may get a fairly impressive answer. It may know what you do, where you are located, when you started and perhaps even who runs the company.
That can be reassuring. It can also be misleading.
The fact that AI knows your company exists does not mean that it understands your company in the way you would like a potential customer to understand it. More importantly, it does not mean that your company will appear when somebody asks for a recommendation without mentioning your name.
There is a considerable difference between recognition, understanding and recommendation.
The company you run and the company the internet knows
Businesses keep changing. Their digital footprints are often slower to catch up.
A manufacturer that built its reputation supplying relatively standard products may have spent the last five years developing sophisticated capabilities for much larger customers. A consulting business may have developed expertise in a particular sector while still being described across much of the internet as a generalist. A Goa company that once worked primarily within the state may now have customers across India or overseas.
Ask the management team and they will describe the business as it exists today, and perhaps more importantly, as they want it to be tomorrow. The internet may still be describing the business of five years ago.
This matters because an AI model trying to answer a customer’s question does not attend your strategy meetings. It has to make sense of the information it can find. Your website is an important part of this, but so are old news reports, directories, interviews, industry listings, reviews, social profiles, videos, customer references and a variety of other material that has accumulated over the years.
If those signals tell different stories, the resulting picture may not be the one management expects.
Knowing what you sell is not enough
Suppose a company describes itself as an engineering business. That tells us remarkably little. Does it manufacture products or undertake projects? Does it cater to small factories or large industrial plants? Is it particularly strong in pharmaceuticals, food processing or mining? Does it compete on price, engineering depth, reliability, speed or service support? Does it operate only in western India or can it execute projects nationally?
These details determine whether the business belongs in a particular customer’s shortlist.
The same problem exists in almost every sector. A hotel may be excellent for families but not management conferences. A software company may be suitable for a fifty-person organisation but a poor choice for a multinational. A digital agency may be very good at local lead generation and completely unsuitable for a complex industrial brand.
Humans understand these distinctions instinctively when sufficient context is available. AI has to reconstruct them from the evidence available online.
This is where a curious situation can arise: the machine may know perfectly well what your company does and still fail to understand when it should recommend you.
Sometimes AI exposes an older marketing problem
It is tempting to call all of this an AI problem. Often it is not. If your website says one thing, your LinkedIn page another, old directories describe services you no longer prioritise and the strongest independent references to your company relate to work you did ten years ago, AI has not created the confusion. It has merely exposed it.
The more useful question for management may therefore not be, “How do we optimise our company for ChatGPT?” It may be: Does the information available about our company accurately reflect the business we are trying to build?
That is a much broader question. It involves positioning, marketing, public relations, sales, reputation and corporate communication. AI Discovery sits across all of them.
It also explains why simply producing more content is not necessarily the answer. If the positioning itself is unclear, producing fifty more articles may simply give the internet fifty more ways to be unclear about you.
Ask a different question
Instead of asking an AI assistant to tell you about your company XYZ, try asking: “What type of customer is XYZ best suited for, and why?” Then ask your CEO, marketing head and sales head the same question. The words do not have to match. What matters is whether the substance does.
You can take the experiment one step further and ask, “Why should a customer consider XYZ instead of its alternatives?” The answer may be flattering. It may be wrong. It may highlight something you stopped focusing on years ago. Occasionally it may identify an advantage that your own marketing has not articulated particularly well. All of those outcomes are useful.
In my previous column, I suggested asking whether AI would recommend your brand. There is a question that comes before that. Before expecting a machine to recommend your company, check whether it understands the company you have become. Being known is useful. Being understood is considerably more valuable.