> ## Content Index
> Fetch the complete content index at: https://www.martechtherapy.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# How ready are you for the agentic CDP?
- URL: https://www.martechtherapy.com/how-ready-are-you-for-the-agentic-cdp/
- Published: 2026-11-16T07:00:00.000Z
- Updated: 2026-11-16T06:59:59.000Z
- Description: Five parts in, the only variable that matters is not the vendor. It is the organization you point it at. Here is how to score your own.
- Author: Matthew Niederberger
- Tags: Agentic CDP, Agentic CDP series

I promised five parts. Part 4 went up a month ago, and then I went on a well-deserved summer holiday. And got distracted, again.

I spent the past few weeks writing [a practitioner's guide to who owns the identity match](https://www.martechtherapy.com/identity-resolution-a-practitioners-guide-to-who-owns-the-match/) and [what happens when two systems disagree about the same customer](https://www.martechtherapy.com/which-context-wins-when-two-systems-disagree-about-your-customer/). Both of which, it turns out, are what this final part leans on hardest.

Poor scheduling, or research? I will take either.

Anyway. The closing piece.

## Almost nobody has actually done this yet

![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/08/adoption-gap-chart.svg)

In Brinker and Riemersma's [State of Martech for 2026](https://newsletter.chiefmartec.com/p/here-s-your-ungated-copy-of-the-state-of-martech-2026-report?ref=martechtherapy.com), 90.3% of marketing and martech leaders say they use AI agents somewhere. Only 23.3% have them in full production, and most of the rest keep the agent on a short leash, suggesting while a human decides.

So if you are not running an agentic CDP today, you are not behind.

You are early.

Early is the best place to be, because the work that matters is mostly work you can start now, long before the contract or the migration, and well before some keynote at a conference somewhere convinces your CEO the tool will sort out a decade of neglected data on its own. As if...

This last part is a check on how ready you actually are, across the three things I have been circling since June: **structure**, **capability**, and **process**.

You do not need three greens.

*You need to know *which one is red*.*

## Readiness is the variable, not the tool

Let me say the thing these five parts have been building toward.

> An agentic CDP does not fail or succeed on the vendor you pick.  
> It fails or succeeds on the organization you point it at.

I might be preaching to the choir here, but I have watched this play out for years, long before agents showed up.

A team buys the platform that topped the shortlist, and eighteen months later the platform is fine and the results are not. The data was a mess, or nobody owned the goals, or the one person who understood the segments took another job.

Agents raise the stakes on all of it.

A capable team gets more out of a modest tool. An unready team hands a powerful one the keys and watches it drive, confidently, in the wrong direction.

I wrote about the distance between what Martech can do and what an organization can absorb in [the Martech Readiness Gap](https://www.martechtherapy.com/the-martech-readiness-gap/). That distance is Martec's Law, and agents pull on it hard. I have argued [elsewhere that AI might finally close that gap](https://www.martechtherapy.com/will-ai-close-the-gap-in-martecs-law/), and I still think it can, but only for the teams ready to use it. For everyone else it widens.

Use this check to work out which side you are standing on.

## Three questions, and where the other parts fit

![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/08/agentic-readiness-check.svg)

### **Structure**

**Can the agent reach clean data, and does the tool fit you?**

Where your data lives, whether identity resolves, whether the platform suits your real needs. The fit-to-need scorecard from [part 3](https://www.martechtherapy.com/how-to-evaluate-an-agentic-cdp-on-your-own/) sits here. So does identity, which is why I went and wrote [a whole guide on who owns the match](https://www.martechtherapy.com/identity-resolution-a-practitioners-guide-to-who-owns-the-match/) instead of finishing this piece on time. An agent that cannot join a customer across web, email, and the store will act on a broken picture and never know it.

### **Capability**

**Can your people direct an agent?**

The move from [operator to director](https://www.martechtherapy.com/from-operator-to-director-the-marketers-job-after-the-agentic-cdp/) in part 2\. What pays now is writing a brief an agent can act on, then judging what comes back without flattering it. If the only person who can tell good output from confident nonsense is you, your capability is one deep.

### **Process**

**Have you agreed what good looks like, and who lets the agent act?**

[The measurement work from part 4](https://www.martechtherapy.com/your-agentic-cdp-will-optimize-what-you-measure/), goals and guardrails, plus the autonomy line: *what the agent settles alone and what it brings back for sign-off*.

It also covers the case where two systems hold different versions of the same customer and something has to break the tie. *Process is the least glamorous of the three*, and the one most teams have never written down.

## Score yourself, honestly

Rate each one green, amber, or red.

**Structure** is green when your data is unified and identity resolves, you know where it lives and who owns it, and you have scored any tool against your real needs rather than a maximal spec.

**Capability** is green when more than one person can write a goal-and-guardrail brief, someone can look hard at an agent's output and catch the mistakes that sound right, and your team has delegated small decisions and reviewed them before.

**Process** is green when your goals are written as outcomes with measures and limits, the numbers come back fast enough to steer on, and someone owns the line between what the agent does alone and what needs a human.

*Three ambers is a normal*, healthy place to start.

*Do not wait for three greens*, because honestly, that team does not exist.

A red in structure means fix the data before you shop, because no tool fixes it for you. A red in process means an agent will amplify a disagreement you have not settled. Pick your red, or your darkest amber, and put a quarter into it.

## One team, three colors

Let's walk through an example.

Take a mid-market team. Not a real one, a composite, but you will recognize it. 

- **Structure**, amber: the warehouse is fine, but the same customer still shows up three times and nobody has been asked to fix it.
- **Capability**, red: one person can brief an agent properly and she is also the only one who reviews the output.
- **Process**, red: the goals live in a deck, and whoever is not in a meeting signs off on the send.

Their instinct is to fix structure, because data work feels like the serious move (and because there is usually budget for it).

I would point them at capability first. Not because the data does not matter, but *because a second person who can brief and judge costs a fortnight of deliberate practice, and it changes what the team can safely do with everything else*. 

Structure without capability gets you clean data nobody can direct an agent to use.

![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/08/readiness-check-worked-example.svg)

That is the value of coloring in three boxes. Not the colors, the argument you have afterwards about which one you touch first.

## Two kinds of agent are coming

These five parts have pointed one way, at agents that work for you.

Brinker and Riemersma keep pointing at the other: agents that work for your customer. Buyer-side agents that research, compare, and negotiate on someone's behalf, and that you will not control.

I have no idea how that plays out, and anyone who tells you they do is selling something. *But getting ready for the first is how you prepare for the second.*

An organization that can state a goal, trust its own data, and judge an agent's work is one that can hold its own when the customer turns up with an agent of their own.

The readiness you build now pays twice.

## Where this leaves you

Back in June, the agentic CDP was a label two vendors reached for within a day of each other.

Now you have a definition you own, a picture of the job it changes, a scorecard to buy by, a measurement check, and a readiness gauge across structure, capability, and process.

None of it needed the platform.

That was rather the point.

The category is early. The teams that come out ahead are the ones doing this work before the tool enters the architecture, not after.

I run this exercise with teams in [the Martech Workshop](https://workshop.martechtherapy.com/?ref=martechtherapy.com), if you want help turning three colors into a plan. And the framework underneath it lives in [the Martech Readiness Gap](https://www.martechtherapy.com/the-martech-readiness-gap/) and [Finding the Martech Goldilocks zone](https://www.martechtherapy.com/finding-the-martech-goldilocks-zone/).

Thanks for coming this far with me. Now go find your red.

Do you have any questions after reading this article?  
Or need support with your Martech projects?

[Contact me today >> ](https://www.linkedin.com/in/matthewniederberger/?ref=martechtherapy.com)