The news comes from India: MoEngage, a marketing platform, argues that the future of the discipline lies in millions of autonomous assistants working in parallel. Beyond the headline, the underlying idea matters to any private company that wants to sell better: AI agents in marketing are not limited to answering questions, but rather execute multi-step tasks —segmenting, personalizing, sending, measuring and correcting— with minimal human intervention. Is this realistic for a Spanish SME today, or is it still investor narrative? Let’s separate the trend from the hype and see what can really be put into practice.
What MoEngage proposes and why it matters
MoEngage’s approach is that the next wave of marketing won’t be just another tool, but a constellation of agents that collaborate with each other: one analyzes customer behavior, another decides the message, another chooses the channel and the timing, and another measures the result to readjust the next action.
What’s relevant for the reader is not the Indian company itself, but the shift in mindset it represents. An AI agent is a system that acts, not just one that converses. And when several of these systems coordinate with each other, they approach what in jargon is called intelligent automation: combining traditional process automation with AI layers that provide judgment where a person was previously needed.
From the global promise to real business
Here it’s worth coming down to earth. A retail chain, a materials distributor or a group of sports centers doesn’t need "millions of agents": it needs to solve specific problems. Recover abandoned carts without anyone keeping an eye on it? Notify the gym member whose subscription is about to renew? Personalize a promotion based on purchase history? That’s where this technology stops being a headline and starts delivering measurable results.
The frequent mistake is falling in love with the tool before being clear about the use case. At TISA we insist on the opposite: first the business problem, then the technology that solves it. An AI project that can’t be measured is a project that’s hard to justify.
Why data comes before AI agents in marketing
There’s an uncomfortable truth behind all this enthusiasm: without organized data, no automatic assistant performs well. If customer information is spread across the ERP, a spreadsheet, the e-commerce platform and the salesperson’s head, no system will be able to personalize anything with judgment. The foundation of any serious initiative is data governance: knowing what data we have, where it is, who accesses it and with what quality.
That’s why the sensible path for an SME doesn’t begin by buying the most talked-about platform on LinkedIn, but by putting things in order at home. When sales, inventory and customer data live in an integrated way —for example, in an ERP like Microsoft Dynamics 365 Business Central connected to analytics tools like Power BI— automation has foundations to lean on.
How we put it into practice from custom development
In the custom development line we work on this trend with a pragmatic, phased approach:
1. Start with a measurable use case
We choose a specific and well-defined process: a recovery campaign, a welcome flow, a renewal notice. We define which metric has to move before touching a line of code.
2. Build on Power Platform
With Power Automate we model the automation flows and with Power Apps the interfaces that the marketing team needs. It’s the most direct route for a company that already lives in the Microsoft ecosystem: native integration with its data and with Copilot, without setting up a parallel system.
3. Add AI where it provides judgment
Not every task needs a language model. We reserve the generative AI layer for what really requires it: drafting message variants, classifying queries, prioritizing opportunities. And always with human verification, because a model can hallucinate and present as true something that isn’t.
4. Measure, correct and scale
A good automatic agent improves over time because it learns from results. We start small, check that the metric moves and only then expand to other processes.
What is realistic today (and what is not yet)
It’s worth being honest. The idea of AI agents in marketing operating on their own at large scale still has a lot of aspiration to it: it requires data maturity, integration and supervision. What is within reach of an SME today is the intelligent automation of repetitive tasks, personalization based on real data, and analytics that turns intuition into data-driven decisions.
In other words: you don’t have to chase the latest headline to get value from this technology. You have to start with a specific case, with data in good shape and with a partner who knows both the business and the tools. The difference between a project that works and one that stays a demo is usually right there.
The role of a partner with experience
Since 1987 we have accompanied private companies and SMEs in their digital transformation, with experience in retail, distribution, services, construction, food industry, hospitality and sports centers. That sector closeness is what allows us to distinguish the passing fad from the real opportunity, and to translate trends like this one into something that improves your bottom line.
If you’ve been wondering how to apply automation and AI to your marketing without diving in blindly, let’s talk. We help you identify the first use case, get your data in order and build a solution tailored to you. Call us at (+34) 971 305 885, write to us at info@grupotisa.com or visit grupotisa.com for a no-obligation assessment.