For years, marketers have been promised that artificial intelligence would make life easier. We’ve seen smart recommendations, automated subject lines and predictive analytics. But 2025 feels different.

This year, the conversation isn’t about AI as a tool. It’s about AI as a teammate.
And that’s where agentic marketing enters the scene.

The concept, popularised by IBM, Salesforce and Braze, is simple but radical. These companies are developing AI systems that don’t just assist marketers but actually act independently toward goals like engagement, retention and loyalty.

In other words, marketing just got a new colleague. One that never sleeps, A/B tests continuously and learns faster than any intern after six coffees.

What Exactly Is Agentic Marketing

Think of agentic AI as the next evolution of generative AI.

Generative AI can create emails, images and social media posts.
Agentic AI can decide which version to send, when, to whom and even how to adjust your ad spend based on live results.

According to IBM, agentic AI refers to systems that can sense, reason and act independently to achieve marketing objectives. These are not pre-programmed automations; they are adaptive agents that tweak campaigns dynamically based on performance data.

That means your AI agent could decide to send Product X offers to high-value customers in Sydney this week, pause underperforming creative in Brisbane and test a new headline, all without you lifting a finger.

Yes, it’s exciting and slightly terrifying at the same time.

Salesforce, IBM and Braze: The New Power Trio

Salesforce’s Agentforce: Making Marketing Autonomous

In September 2025, Salesforce unveiled Agentforce, part of its Einstein 2 platform. The goal is simple: allow AI to handle entire workflows, from lead qualification to campaign orchestration.

Marketers can assign AI agents tasks like:

• Optimizing email journeys in real time

• Choosing which channels to use for each segment

• Generating and deploying creative variations automatically

Agentforce connects across Sales Cloud, Marketing Cloud and Service Cloud, giving brands a unified AI employee that executes campaigns end-to-end.

This is part of Salesforce’s broader promise to make every marketer ten times more productive. You can read more about how predictive tools are already reshaping loyalty in AI and the Future of Loyalty.

---

IBM’s Forecast: Agentic AI Pilots Will Go Mainstream by 2025

IBM’s Marketing Intelligence Outlook 2025 predicts that over half of enterprises will run agentic AI pilots within the next year.

Their reasoning is straightforward. CMOs are under pressure to do more with less, facing shrinking budgets, rising ad costs and increasing demands for personalization. Agentic systems can auto-optimize campaigns, freeing human teams to focus on creative and strategic thinking.

But IBM also warns that around 40 percent of projects could fail because of “agent washing,” where brands rebrand basic automation tools as agentic AI without real autonomy behind them.

In other words, there’s real potential, but plenty of marketing hype too. Success depends on proper guardrails and transparency.

Braze’s Take: From Journeys to AI-Driven Orchestration

Customer engagement platform Braze, known for powering loyalty and engagement for brands like Burger King and HBO Max, has jumped headfirst into the agentic era.

Their 2025 trends report highlights “agentic orchestration,” where AI systems decide how to engage each customer, when to reach out and which channel will hit best.

For example, Braze’s system might detect that your app users open push notifications mostly after 8 p.m. and that email performs better on Fridays. It then automatically adjusts timing and messaging for each segment.

That’s no longer predictive analytics. That’s self-optimizing marketing.
You can explore how this trend fits within broader engagement strategies in The New Rules of Customer Engagement.

What This Means for Loyalty and Retention

Here’s where it gets interesting for loyalty marketers.

Agentic AI isn’t just about running ads. It’s about creating micro-loyalties in real time. Imagine this:

• Your AI agent identifies customers likely to churn next week and triggers surprise offers before they leave.

• It personalizes rewards dynamically, such as doubling points for product categories that specific customers actually love.

• It tests retention messages automatically and kills off weak ones within hours instead of weeks.

In this world, AI becomes a relationship manager at scale.

However, the best loyalty experiences still come from human empathy. AI can identify patterns, but it can’t feel joy or surprise. Brands like Nike and Sephora that combine automation with emotional storytelling are already proving that the most powerful marketing blends data and humanity.

How Mid-Sized Brands Can Start Without a Big Budget

You don’t need Salesforce’s enterprise software or IBM’s research lab to start experimenting with agentic marketing. Here’s a practical way to begin:

1. Start small. Try a micro-agent for specific tasks like subject line testing or ad bidding.

2. Keep a human in the loop. Let AI recommend actions, but always review results before full automation.

3. Focus on clean, ethical data. Poor data leads to poor decisions.

4. Define success early. Give your AI a clear goal, like improving engagement rate or reducing cost per acquisition.

5. Audit outcomes weekly. Treat your AI as a junior team member. You trust it, but you still supervise.

If you’re exploring how to build confidence around automation, you’ll find useful strategies in How to Build Customer Trust in the Age of Automation.

The Risks You Need to Manage

Agentic marketing isn’t risk-free.

• Brand safety: Autonomous systems might deliver messages that feel off-brand if context is misread.

• Ethics: Who takes responsibility when an AI agent acts wrongly?

• Data bias: AI can amplify biases if your data isn’t diverse or balanced.

This is why experts like IBM’s APAC Marketing Director Angela Chien advocate for “guardrailed autonomy.” The idea is to let AI act freely within defined limits while logging every decision for accountability.

It’s like giving your intern a corporate card but setting spending limits. You trust, but verify.

The Future: From Assistants to Colleagues

Fast forward a few years and marketers may be working alongside digital colleagues that brainstorm, test and launch campaigns on their own.

Salesforce calls this collaborative intelligence. IBM calls it autonomous orchestration. For most marketers, it simply means less repetitive work and more time for creative problem-solving.

When done right, agentic marketing won’t replace people. It will free them to do what humans do best: build stories, empathy and connection.

Because no matter how smart AI becomes, it will never truly understand why a customer smiles when they open your email.

Final Takeaway

Agentic marketing isn’t about robots taking over. It’s about giving marketers superpowers.

Salesforce, IBM and Braze are showing what’s possible when you combine smart automation with human oversight, feed your systems high-quality data and measure outcomes ruthlessly.

The brands that master this balance won’t just automate marketing.
They’ll reinvent loyalty itself, turning every interaction into a living, evolving conversation between customers and intelligent systems.