When it comes to AI for Marketing - I am a glass half-empty kinda guy

Don't get me wrong, Generative AI is magical. Shakespearean rap about gerbils - no problem. 1000 image variations of your brand of jelly skiing with kittens - absolutely. Precise text that is compliant to medical rules in Albania - you got it.

The possibilities are endless. And marketers like myself are having a ball tinkering with it. And every minute of free time that Gen AI tools were supposed to give us is now spent in tinkering with new ones that emerge every hour.

The marketing industry is definitely going through a transformation.

• Poor freelancers are suddenly out of a job, as their output is replaced by bots
• CMOs who just signed an expensive Gen AI software contract are demanding their teams to generate twice the output in half the time
• CEOs are expecting their CMOs to do twice in half the budget
• And boards are expecting their CEOs to launch more products and more services than ever before

And of course this shows in recent surveys, with 70% of CMOs saying that they are "experimenting" with Generative AI and they are "confident" about the tech's future impact on their efficiency.

That's the problem, isn't it? CMOs are trying to say "my cup runneth over with cool Gen AI tech" - but the way I see it, these statements don't do justice to the true potential of this amazing tech. Hence - glass half empty.

And after being involved in multiple Gen AI pitches to CMOs, I realize that the issue is that most people can't get over the "coolness" of the tech, and are refusing to see that like any technology-driven transformation, it begins with process transformation.

Let me demonstrate through the example of a simplified campaign launch process:

What you see here is different teams using different tools to achieve interconnected outcomes. What's missing here are the hundreds of sub-steps in each hand-off. For example, the creation of the brief will take multiple interactions between the product, sales and marketing teams. Micro-segmenting and planning the data for the campaign can take weeks; and we have not even got into the creation yet. And sure enough, everyone is using some form of AI or Gen AI to "help" them in their tasks.

The way Gen AI is being used today reflects this siloed approach, and I emphatically declare that this approach is scratching the surface of AI's true potential for marketers.

AI and Generative AI have the potential to achieve large scale industrialization of marketing and all its functions, creative or non-creative.

If that is not what the CMOs are aiming for, then their glass will forever remain half-full; and enlightened competition which takes the "industrialization" view will rule the market.

The agents are here.

"I'd like to share a revelation that I've had during my time here..." is one of the many memorable lines uttered by Hugo Weaving's Agent Smith - the human manifestation of a software bot - in the Matrix.

I've had a revelation too after two decades of running marketing teams.

Effective marketing depends on the industrialized dissemination of creative ideas to appropriate audiences. And industrialization is a function of predictability and consistency, not of creativity.

The eagle-eyed among you will notice that there is a contradiction baked into the statement. If creativity is antithetical to industrialization, then is industrialized creative marketing a myth? It was until 2024. You see, the biggest impact of Generative AI is exactly this: it bridges creativity and industrialization.

But not in the way it is being used today. Because the other part of industrialization is the breaking down of silos through integrated processes and sharing of insights and data. Other functions such as Manufacturing, Operations or Client Servicing achieved this through bots.

But bots are dumb and efficient. They have defined inputs, defined logic and defined outputs. They stay in their lanes and do what they are supposed to with 100% accuracy.

Agents are bots that can think like humans do

There is a reason why Salesforce is going all in with Agentforce.

(link opens the latest Agentforce ad that is completely confusing many about what Agentforce actually does)

Agentic workflows represent a new frontier for Marketing Industrialization

So what is an Agentic Workflow in Marketing?

It would be interesting to understand first what an Agent can do.

An Agent can take human-like decisions through "Situational Awareness". It does this by ingesting structured and unstructured data and correlating it with specially trained large language models (LLMs) through intelligent prompts or a series of chained prompts. (In non-geek speak, this basically means an Agent can "talk" to an AI model to infer the best decision or next action.)

Let's now apply this to a real situation in Marketing & Sales:

So in this case, the Direct to Consumer Agent, detects Instagram activity by an influencer, and uses it to generate an "on-the-spot offer" for a buyer on Instagram who happens to follow that influencer. Now imagine 1000s of such bots working 24-7 to "hyper-target" millions of potential customers across Instagram, Facebook, WhatsApp, Mobile ad-networks etc.

If it sounds too good to be true, then it is - and it is not.

Sure enough, more agile companies are already doing it. Amazon and Meta are already doing it in their "walled gardens". But for most of us, this remains a dream, simply because we believe that we don't have the infrastructure for it.

Marketing organizations are more ready for agentic industrialization than they think

The good news is that all the tinkering that your team is doing with Gen AI tech; won't go to waste. This is how general populations learn new technology. They play with it.

But the play phase is over now for Marketers. Time to get serious.

Here are 5 things that CMOs and Marketing Tech leaders should consider in 2025.

1) When considering Gen AI - look at the process and not just the task. You will be surprised how many marketing teams out there don't have their processes mapped out in detail. Proper documentation of processes is the first step before any potential transformation.

2) Hire a professional agency (like the one I work for) to evaluate the "Gen AI readiness" of your current marketing organization and processes. It won't be all good news, but it won't be all bad either - and you might be surprised by how small improvements and integrations can help greatly.

3) Invest in Gen AI awareness and education for your team. Create an Gen AI innovation fund. Do hackathons. Invite vendors to do free presentations. Yes there will some money to be spent, but it will be well worth it.

4) Evaluate RAG (Retrieval Augmented Generation) applications in your current marketing workflows. In all probability this will require a tech-upgrade (both in terms of tools and skills). Don't be afraid of it...it is less complicated that it sounds; but is the basis of most personalization.

5) Improve your data quality. This will likely be one of the recommendations from your "Gen AI readiness assessment". But you don't need to wait for that. Clean your data because successful industrialization depends on data quality (and quantity).

Summary

2025 will be the year that will separate the early movers (and early winners) from the followers.

If you are rolling your eyes, I get it. Every new year carries with it the promise of something amazing. But what I have seen with the new OpenAI and Anthropic models in terms of advancement towards "Artificial General Intelligence (AGI)" is scary and exciting. Google's Whisk has the potential to be game changing. And Nvidia's Blackwell announcements show that the big companies out there will continue pushing the hardware boundaries of raw inference speed.

If as a Marketing leader, you have been sitting on the fence on deploying Gen AI to transform what your team can achieve, it is time to get off it.

This is as real (and urgent) as it gets.