Only 12% of companies believe their data quality is actually good enough for AI.
That number, from a Gartner study, has been making the rounds lately, and it’s worth sitting with for a moment. Because Marketing Cloud Next is, at its core, a data-first platform. Your segments, flows, agent interactions, customer profiles, all of it runs on the quality of what you’ve put in.
Marketing Cloud Next has matured significantly, and the case for moving is stronger than it’s ever been. And a more capable platform also means more ways to get the implementation wrong. Our Marketing Cloud team has identified 6 pitfalls. Don’t migrate without reviewing these first.
Table of Contents
Marketing Cloud Next implementation pitfalls
1. Poor data quality
2. Treating transition as a big bang migration
3. No vision for agentic governance
4. Underestimating Salesforce Flow learning curve
5. Ignoring credit consumption planning
6. Skipping the AMPscript audit
Are you ready for Marketing Cloud Next?
Marketing Cloud Next implementation pitfalls
1. Poor data quality
The garbage in, garbage out phenomenon is well-understood. It’s hard to believe that marketers don’t realize the importance of data quality.
More so, now that AI pervades marketing.
And there’s the rub; we’re on a different terrain now, one that is dominated by agentic AI.
Here it is useful to refer to IBM’s effort at distinguishing data quality understood in the traditional sense, and data quality in the context of agentic AI. This is how the two differ.
| Traditional data quality | AI data quality |
| Focuses on data correctness and cleanliness | Focuses on how data impacts model performance and behavior |
| Static, point-in-time validation | Dynamic, lifecycle-based (training, deployment, monitoring) |
| Checks completeness and accuracy of records | Ensures coverage of real-world scenarios, including edge cases and bias |
| Primarily supports reporting and operations | Directly influences model accuracy, fairness, and robustness |
Arguably, that’s the proper way to think about data quality going forward.
According to IBM, data in agentic AI has to be evaluated across these dimensions:
- Accuracy: Correct data with minimal label or measurement error
- Completeness: Covers all scenarios, including edge cases
- Integrity: Traceable, reproducible, and secure across pipelines
- Consistency: Uniform across sources, pipelines, and time
- Timeliness: Up-to-date and resilient to data drift
- Relevance: Directly improves model performance and reliability
Salesforce’s acquisition of Informatica is essentially a nod to these dimensions.
The IBM framework reframes data quality as a test for strategic readiness, not hygiene. Audit before you configure. Decide what your data has earned the right to automate; start there.
Refer to the pre-migration rundown on Marketing Cloud Next.
2. Treating transition as a big bang migration
We’ve said this before, and we can’t stress it enough.
Some organizations, excited by the potential of Marketing Cloud Next, try to move everything in a single-phase rollout. This approach consistently fails. Marketing Cloud Next’s content migration tools from Marketing Cloud Engagement are still maturing, thus a big-bang transition introduces significant operational risks.
Keeping that in mind, we recommend a phased approach:
- Build new campaigns natively in Marketing Cloud Next
- Keep existing journeys running in Marketing Cloud Engagement+
- Gradually migrate legacy assets as the platform evolves and your team ramps up.
Refer to the checklist for migrating to Marketing Cloud Next.
3. No vision for agentic governance
Did you know the global AI governance market is projected to reach $4.83 billion by 2034? That signals just how critical governance is turning out to be.
At the same time, marketers are still in the Wild West phase of agentic governance.
Organizations that deploy agents in Salesforce Marketing Cloud without governance frameworks are going to run into brand and compliance issues. It is important you ask yourself:
- Who reviews agent-generated content before it goes live?
- What brand voice guidelines must agents follow?
- How do you handle an agent that sends a tone-deaf message during a public crisis?
Agentforce is designed with human oversight in mind, but it requires you to define the guardrails. What that means is you need to configure product knowledge, messaging boundaries, outreach frequency limits, escalation logic, etc. before the agents go live.
One wonders if it will take years of trial and error, and perhaps a few high-profile agent mishaps before standardized governance frameworks are fully adopted. Governance is new territory for marketers, and it can feel like a bother. At first glance, it may seem abstract, but it becomes more clear when reframed in terms of metrics. With that mind, feel free to refer to these six Agentforce governance metrics.
4. Underestimating Salesforce Flow learning curve
Unfortunately, Journey Builder expertise doesn’t directly translate to Flow proficiency.
For teams coming from Marketing Cloud Engagement’s Journey Builder, Salesforce Flow is a different tool with a different logic. Journey Builder is visually intuitive and marketing-native. Flow is more powerful and flexible, but it was originally built for system automation, not for marketers. The Summer ’25 release introduces improvements for marketing use cases (including scenario templates and wait-until-event elements), but the learning curve can be a bit steep.
You might want to read when to use Salesforce flow vs Agentforce.
5. Ignoring credit consumption planning
The shift from Marketing Cloud Engagement’s Super Messages to Marketing Cloud Next’s credit model requires careful planning. Different activities consume credits at different rates—email sends, SMS sends, agent sessions, Data Cloud activations, and segmentations all draw from the same pool. Without a clear consumption model, teams can run out of credits mid-campaign.
Use the Digital Wallet dashboard proactively and model expected credit usage before launching any significant agent-driven workload.
6. Skipping the AMPscript audit
As per the Summer ‘26 release notes, Marketing Cloud Next finally has AMPscript support. The release of this support makes a compatibility audit the most important step of migration.
An AMPscript audit should catalog:
- Every template using AMPscript
- Every CloudPage with scripted logic
- Every Automation Studio query feeding scripted personalization
- Every API call embedded in AMPscript functions
Without a thorough audit of all AMPscript, you can’t assess migration complexity, set realistic timelines, or identify which templates can use the Marketing Cloud Engagement+ bridge versus those that require a complete reengineering.
Marketing Cloud Next is a fast-evolving platform. Features announced at Connections ’25 have been rolling out through 2025 and into 2026 on a phased schedule. Some capabilities are GA, some are in beta, and some are on the roadmap. Always verify the GA status of specific features before building implementation plans around them, and consult the Salesforce release notes for the current state.
Are you ready for Marketing Cloud Next?
To ensure a successful setup, prioritize these three actions immediately:
- Audit before you move: Do not migrate legacy technical debt. If an AMPscript block or a data extension hasn’t been used in 12 months, leave it behind.
- Budget for the learning curve: Your team’s current expertise in Marketing Cloud Engagement will not fully cover Salesforce Flow or Agentforce governance. Allocate specific time for retraining before the platform goes live.
- Monitor consumption daily: Because credits are shared across segments, sends, and AI agents, an unoptimized flow can deplete your budget quickly.
You don’t have to navigate this architectural maze by yourself!
Book a “Marketing Cloud Next Complexity Assessment” with a Marketing Cloud architect.
