Episode 50 - Agentic AI Part 1: Definitions & Growth

Did someone forward this to you? If so, click to subscribe

📰 Top Story: Agentic AI part 1 - Definitions & Growth

Back in Episode 11 of this newsletter, I floated the idea that AI agents could soon drive entire CRM marketing journeys - from ideation to execution - with nothing more than a single prompt.

That’s no longer hypothetical.

Martech vendors are already doing this (see Netcore Cloud, Airship and others) and new players are entering. And the race is on to build tools that don’t just help marketers do CRM better - but that do CRM for them. Entirely.

This next wave is powered by what’s called agentic AI - and it’s going to completely change how retention strategies (not just CRM) are designed, deployed, and scaled.

If you're serious about retaining customers in the age of AI, In this episode I break down what you need to know about what agentic AI is, while in part 2 we will dive in specific applications in customer retention. 

Agentic AI is artificial intelligence systems that can act autonomously to achieve specified goals. 

You’ve seen how AI can act as a co-pilot - helping you write emails, segment users, or even tweak copy based on performance.

That’s useful. But it’s still reactive. It still needs you to give it instructions.

Agentic AI is different.

It doesn’t wait to be told what to do.

It defines tasks, makes decisions, and adapts its strategy - all autonomously.  

This means: 

1. The AI decides which tasks are required to accomplish goals 

2. The AI plans sequences of actions to accomplish goals 

3. The AI makes decisions without human guidance

4. The AI learns from experience  

5. The AI interacts with other AI Agents and systems to gather information 

6. The AI adapts its strategies and changes courses of action as information changes 

7. The AI has the ability to perceive its environment and take action accordingly 

8. The AI is able to develop and deliver different experiences to different customers based on previous experiences 

9. Essentially, the AI that can act on behalf of a person, and make decisions that a human would make, most likely even better decisions than a human would make, given the depth of information access and the lack of human biases 

Agentic AI is a fundamental shift in how retention strategies are built - moving from rule-based workflows to outcome-driven systems.

Agentic AI includes tools that can proactively analyse complex situations, identify emerging problems, develop innovative solutions, and implement those solutions autonomously. 

This will fundamentally transform customer retention because it will enable clear identification of multiple interdependencies in how relationships are managed. I’ll explore this further in future episodes of this newsletter.  

Agentic AI will operate in dynamic environments, making real-time decisions based on contextual data and learned experiences, enhancing efficiency and effectiveness. 

Everything that has gone before in AI, like recent advancements in generative AI and co-pilots, like ChatGPT, are simply a precursor to the huge change we are about to experience. 

To truly understand what agentic AI is, it is essential to be clear on definitions 

Agentic AI is not automation or AI workflows. Alexandre Kantjas does a great job of breaking it down in this LinkedIn post. 

To summarise 

• Automation is about leveraging AI to execute pre-defined tasks, e.g. build this CRM Marketing journey to drive repeat e-commerce purchases 

• An AI workflow is a program that calls an LLM (Large Language Model) via an API for one or more steps, for example: write copy for each communication tailored to each segment for the same repeat purchase e-commerce CRM journey 

• An AI Agent is different to a workflow because it has the autonomy to build the solution in whatever way it wishes in order to drive the best outcome e.g to maximise e-commerce repeat purchases for different cohorts with different experiences delivered for each unique customer. 

AI agents are, therefore, non-deterministic and very adaptive, while automation or AI workflows powered by AI are much more deterministic and the rules are directly set by humans. With agentic AI human set the parameters, not the direct rule set.

For agentic AI to be effective, AI agents needs to be able talk to other AI agents

In a lot of contexts, to be successful AI agents will need to be able to interact with other AI agents, and this is where MCP (Model Context Protocol) comes in.

With MCP, different AI agents - across marketing, product, billing, customer success - can communicate in natural language, via LLMs, without needing complex integrations.

That means:

• A retention agent can ask a billing agent if a customer's payment failed
  

• A CX agent can flag to a product agent that feedback sentiment is dropping
  

• An onboarding agent can pause a CRM automation agent if the user is stuck in support
  

In other words, agents will not work in silos - something most organisations still haven’t cracked with human teams.

Anthropic (creators of Claude) who built MCP ask us to “Think of MCP as a USB-C for AI applications’. It is essentially a protocol that allows AI models to connect with outside data sources and services without requiring unique integrations for each service. 

Agentic AI will be the engine of AI growth

The agentic AI market is expected to grow 10-fold over the next 5 years as adaptation accelerates. 

The implications for customer retention are huge, think dynamically adapting experiences delivered at a cohort level: 

1. Proactive Churn Prevention 

2. Highly Intelligent Chatbots 

3. Dynamic Loyalty Program Management

4. Retention-Focused Pricing Optimisation

5. Digital Twin Customer Account Manager 

6. Competitive Intelligence Shield

7. Customer Success Autopilot

8. Relationship Health Management

9. Cross-Organisational Retention Intelligence

10. Hyper Personalised Contextually Evolving Experiences 

In episode 51 of this newsletter and part 2 of this agentic AI series, dropping next Friday, May 30th, I’ll explore these retention use cases in much more detail. 

Until next week, 

Tom  

Do you need help with Customer Retention?

When you are ready, contact me to discuss consulting, my fast-track retention accelerator, courses, and training. Or if you are interested in sponsoring this newsletter, get in touch via [email protected]