Can CMOs guide enterprises out of the AI rut?

Time of India

06-08-2025

Today, modern
CMOs
face multiple challenges. The average CMO tenure is on a downward spiral, currently at 4.3 years. Organisations increasingly view the CMO as a direct business and revenue driver, with 64% of CMOs saying they are responsible for profitability and 58% saying they are responsible for revenue growth. The consumer is 'always on', constantly generating data that signals their intent to discover, purchase or seek assistance.
A recent CMO study published by
IBM
claims that the successful marketing organisations and CMOs will not be the ones with bigger budgets or shinier tools but the ones that reinvent the operating model that delivers business growth. The reality is that CMOs are being undermined by operating models of a bygone marketing era and need to become interpreters and integrators between different siloed functions to deliver value to customers.
Tuhina Pandey
, director - APAC communications and marketing, India and South Asia, IBM, in conversation with ETBrandEquity.com, shared why AI holds the key to enabling marketers to deliver on customer expectations, organisational demands and unlock growth. However, marketers need to tackle operational realities before they can successfully implement AI.
Edited Excerpts
Q. How is the outdated marketing 'operating model' holding back marketers from delivering what their organisation demands?
There is not a new or old operating model, but how you respond to the needs of the marketing environment today. Marketers are grappling with a harsh paradox. At one end, there is an expectation to champion topline (revenue) growth, achieve a higher degree of profitability, increase return on investment (RoI), and improve expense-to-return ratios. AI is the way to do it. Bring in automation, cut down on repetitive tasks, and generate content on the fly so you can respond to clients with agility. Even if marketers accept this reality, organisations are not structurally or culturally prepared to leverage AI.
Q. Let's talk about the current marketing environment. What are the customer expectations?
Clients want messaging that is relevant to them and hyper-personalised. You have to intercept the client's journey; you can't just push your products onto them. You need to meet them in a moment of need or a moment of intent with the best solution.
However, I can say this of most B2B enterprises: 99% of data is sitting untapped by AI. If you layer AI and
MarTech
tools onto data that is siloed, where there is no single source of truth, that isn't collaborative, isn't current, or isn't of high quality, then the results are not going to be great. We are solving for the wrapper and not the core.
Q. Where are enterprises in their data journey?
Most enterprises' data journey is a 'work in progress'. The way enterprises like banking, finance and retail utilise their data is different depending on customer expectations and the regulatory environments, which determine much of how they leverage their data. Regulation, compliance, and the law of the land mandate that you have certain governance frameworks for your data.
The good part is that India Inc. is experimenting with AI and has been able to derive outcomes. There are purpose-built AI models where you can backward integrate to try and solve a business problem and pull the requisite data.
But if you really want to scale and derive the benefits of AI, then you have to put in the hard yards. That's going to be a journey of a year or two before we start seeing outcomes in the Indian business environment. The good part is that the industry attitude has changed to 'how can we drive real impact with AI' versus 'let's experiment with AI'.
Taking our own example, the IBM Client Zero story is pretty compelling. We've gone on record to say that we've saved $3.5 billion in bottom line because we are using AI automation in our processes and workflows.
Q. How can AI unlock value in the marketing function?
On the marketing side, we're seeing disruption in many areas.
One is the line of business or the functional side, such as workflows and processes. When you're integrating AI into your workflows and processes, you have to overhaul them and find the most efficient way of doing it. If you just automate with AI without changing the processes, you won't get the outcomes.
On the MarTech stack, it has become very important to have a customer data platform or unified data platform to get the source of truth right and break silos. You can't get an integrated truth about your customers if your data isn't integrated.
Additionally, aligning functions like marketing, sales and operations has become equally important. The report talks about a potential 20% revenue unlock if there is tighter alignment between these three functions. You need to synchronise internally between these three really well to ensure continuum in the client's journey, so that there is no breakage or seepage.
We're seeing a lot of applications of AI in garnering insights, analytics, segmentation, content generation, and being the marketer's creative assistant on a day-to-day basis. But what we're really conscious of is the responsible AI part of it.
Organisations should be able to explain their AI models in a transparent way. This is especially true for marketing functions which deal with customer data, pain points and needs. Not working with responsible AI is like running very hard but in the opposite direction.
At IBM, we're disproportionately focused on creating foundational datasets. We believe in AI augmentation but are also conscious that we need to scale responsibly.
Q. Today, consumers are 'always on' or in a state of perpetual demand. Even if organisations recognise that they need to leverage AI to cater to consumer expectations, will they be comfortable with completely autonomous interactions? What about the human in the loop?
Consumers don't care who is on the other end as long as they are comfortable with the interaction and trust the outcome. Ultimately, you will have automated systems marketing for you. Autonomous agents already exist. The real question is: who is going to connect the dots between the AI and the biases of your data, your creative vision, and the relationship between human and AI? It is not an 'either-or' question, but rather an 'and' question.
Will you have completely autonomous environments? Yes, in bits and pieces, but on a longer continuum it will be both humans and AI. We will have digital colleagues in the near future and certain micro-processes will have autonomous agents, but humans will have a role to play for this to succeed.
Q. While AI pilots are happening across enterprises, your report points out that only 19% are able to deliver the expected RoI. So, what should organisations and marketers focus on?
This is a big reality check when it comes to AI. So, it is equally important for enterprises to figure out what's most important to them. Having clarity of thought means they can apply resources and AI budgets to those problems.
Also, if the AI model eats out of the same data pool as every other enterprise, then you will get the same results. That's why AI needs to be fed with your proprietary data.
The good news is that enterprises are saying, 'Let's look at the problems that we can solve using AI' and not look at it as a silver bullet. The phase of piloting AI projects for six or twelve weeks is over. Now, organisations are putting in the two years of hard work into AI so that they can start reaping the benefits.
AI is just like the mobile phone or the internet. We need to be real about its applications and embed it at the core. It needs to become a part of how enterprises do things rather than become the thing itself.
Q. There has been consumer backlash to AI-generated content. A NielsenIQ study found that consumers found AI ads less engaging. Is AI doomed to be relegated to backend processes and not customer-facing communication?
There is a sea of sameness surrounding AI-generated content. If every organisation is using the same data, then the outcome is not going to be very different. Every corporation has unique institutional knowledge, lived experience of that enterprise, expertise and character. Unless you're patient in building foundational data models, putting in the guardrails, being responsible and accountable, then you will not be unique in your communications. Every human being has the same set of organs and body parts, but we're different in the way we perceive the world, our formative years and experiences. So, the question is: how are we curating AI?
(This interview is part one of a two part series.)