Episode 20
Why AI Can’t Fix a Broken CRM: Marketing Operations Lessons from Danielle Balestra
Danielle Balestra
She has more than 20 years of experience in marketing operations and MarTech. She has led marketing operations teams in financial services, healthcare, legal services, and publishing, and now consults independently on how companies use their MarTech systems to scale. She was named a Marketo Fearless 50 and was a Gartner Innovation finalist.
AI is not the strategy. That is the central point Danielle Brestra, a marketing operations and martech leader with more than 20 years of experience across financial services, healthcare, legal services, and publishing, made on this episode of Ops in Motion. Organizations are buying AI tools faster than they are fixing the data and process problems those tools sit on top of, and the result is AI generating insight from exactly the same stale contact records and disconnected systems that have undermined marketing automation for two decades.
Danielle built and ran marketing operations teams through the first wave of marketing automation, the shift to full martech stacks, and now the move into AI, giving her a direct comparison point most people in the room do not have. Her position: AI can speed up the work dramatically, but only once the underlying data, process, and team alignment are in place. Skip that step and AI just makes bad decisions faster.
Key takeaways
Contact data decay is not new, but AI is compounding it. People changing jobs creates multiple versions of the same contact in a CRM, and AI layered on top of that data inherits the inaccuracy rather than correcting it.
AI is an enabler, not a strategy builder. Organizations that treat AI prompts as a substitute for defined processes and clean data consistently struggle to operationalize what AI produces.
AI-generated insights are often fragmented across disconnected platforms, which recreates the same silos that marketing, sales, and customer success already struggle to bridge.
Handoffs between marketing, sales, and customer success are where data gets lost and revenue leaks out. AI is useful for auditing where those handoffs break down, not for fixing the handoff itself.
A new marketing operations hire’s first 90 days should start with a full systems and process audit, not a playbook. What worked at one organization will not automatically work at the next.
CONTACT DATA
Contact data decay is older than AI, and AI inherits it
Marketing automation promised the right message to the right person at the right time, and the original roadblock, as Danielle describes it, was never the message. It was the data. B2B contacts change jobs often enough that most organizations end up with multiple versions of the same person in their CRM, each tied to a different email address from a different employer. Most data cleansing processes still key off email address, so instead of merging those records, the system treats three or four job changes as three or four different people.
How contact data decay compounds
One person
One contact
→
Changes jobs
New email + employer
→
CRM
Contact 1
Contact 2
Contact 3
Contact 2
Contact 3
→
AI
Inherits the inaccuracy
That is the condition AI is now being layered onto. According to Danielle, AI does not fix the underlying data problem; it sits on top of it and compounds it, because most teams are not documenting the context behind their systems, such as why a workflow was built or who it was built to reach, which means AI has even less to work with than the raw records suggest.
“Now it’s compounded because now it’s AI. So now AI is sitting on top of these very now antiquated data centers, and there could be great outcomes from them, and you have to add additional layers of context to it, which a lot of the times people are not documenting.” — Danielle Brestra
This is the practical argument for treating data decay as foundational infrastructure work before layering AI on top of a CRM, and for building an ongoing CRM data hygiene process rather than a one-time cleanup.
AI & REVOPS
AI is an enabler, not a strategy
The host raised a pattern that is showing up across most GTM organizations right now: teams are buying AI tools and starting to use them before they understand what problem the tool is actually solving. Danielle’s framing is direct. Organizations that succeed with AI are the ones focusing on fundamentals first, clean data, clean processes, defined customer journeys, and alignment across teams, because AI cannot repair what is already broken in the system underneath it.
AI works best as an enabler
Foundation
Clean data
Clean processes
Defined customer journeys
Team alignment
→
AI
Accelerates
Audits
Documents
Surfaces insight
→
Better execution
Faster work
Better visibility
More informed decisions
Danielle’s description of where AI genuinely helps is narrower and more specific than “strategy”: AI is strong at acceleration, not creation. She points to martech stack auditing and process documentation as the two use cases that have changed the most in her own work. What used to take hours of manually digging through a marketing automation platform, and documentation that often never existed in the first place because it was never written down, can now be produced in a fraction of the time, with a human still reviewing the output before anyone acts on it.
That distinction matters for how teams scope their first AI projects: a martech stack audit is a reasonable place to apply AI for speed. Building a GTM strategy from AI-generated suggestions, without first fixing the data and process underneath it, is not.
AI & DATA
Fragmented AI insight recreates the silos it was supposed to fix
A second compounding problem Danielle raised is fragmentation. As more GTM teams add AI capabilities tool by tool rather than as part of a connected system, the output stays disconnected: AI-generated insight lives in one platform, customer data lives in another, and the actual execution happens in a third. That is the same structural problem GTM teams have had with disconnected point solutions for years, just with an AI layer added on top instead of removed.
The handoff problem compounds this further. When a lead moves from marketing to sales, gets qualified into an opportunity, closes, and transitions to customer success, every one of those transitions is a process, and every process creates an opportunity for data to get lost, delayed, or misrepresented. Danielle connects this directly to revenue impact.
Where AI is genuinely useful here is diagnostic, not corrective: auditing the handoff points to see where the leak is actually happening, then connecting that insight back to its impact on revenue and customer experience. The fix itself, closing the gap with the right notifications and automations between teams, is still an operations and process decision, not something AI builds on its own. This is the same argument behind why sales teams ignore marketing leads: the breakdown is rarely a tooling problem, it is a process and ownership problem that shows up at the handoff.
MARKETING OPERATIONS
The first 90 days: audit before you touch the playbook
Asked how she would approach the first 90 days in a new marketing operations role, Danielle’s answer started with a full systems inventory, not a plan of action. Before changing anything, she wants to understand the complete ecosystem: which technologies are in use, who administers each one, and whether those systems are actually integrated with each other. In her experience, that single question, are these systems connected, often reveals the root cause of problems a team has been struggling to explain.
From there, the audit expands to people and documentation: interviewing stakeholders across roles, reviewing existing documentation where it exists, and getting into the systems directly to check dashboards, velocity trends, and whether performance is holding steady or slowing down. The output of that audit is a map of quick wins versus long-tail work, where quick wins might be an integration fix or a shortened form, and long-tail work requires real change management because it touches a process someone built intentionally and believes in.
The mistake she sees most often is the opposite approach: arriving with a standard playbook and assuming it applies regardless of how the organization actually operates.
This same audit-first logic applies to assessing a CRM’s current state before any AI or automation initiative. A CRM database management review and a clear read on how marketing operations is already driving business results give a new leader, or an organization considering AI adoption, the baseline needed to know which problems are quick fixes and which require sustained change management.
Action items
01
Audit CRM contact data for duplicate records created by job changes before introducing AI-driven personalization or scoring on top of it.
02
Document the “why” behind existing workflows and automations; undocumented context is what limits AI from giving useful, accurate output.
03
Scope AI projects around acceleration tasks you can verify, such as martech audits and process documentation drafts, rather than strategy generation.
04
Map the handoff points between marketing, sales, and customer success, and use AI to help identify where data loss or delay is occurring, not to fix the handoff itself.
05
When starting a new marketing operations role or engagement, begin with a full systems, admin, and integration audit before applying any standard playbook.
FAQs
Does AI fix bad CRM data automatically?
No. According to Danielle Brestra, AI is only as effective as the data and processes behind it. It can surface patterns and generate insight faster, but it inherits the same duplicate and inaccurate contact records that already existed, particularly those created by job changes in B2B data.
Where does AI actually save the most time in marketing operations?
Auditing existing martech stacks and drafting process documentation. Danielle estimates AI can cut the time spent on these two tasks by 60 to 70 percent, though the output still needs human review before anyone acts on it.
What should a new marketing operations hire do first?
Audit the full systems ecosystem and who administers it, interview stakeholders across departments, and review dashboards and performance trends before proposing changes. Applying a standard playbook without this audit is one of the most common ways new hires fail.
Why do handoffs between marketing, sales, and customer success cause revenue loss?
Each transition between teams is a separate process, and each one creates an opportunity for data loss, delay, or miscommunication. Left unaddressed, these gaps show up as revenue leak and a degraded customer experience, independent of how good the tools on either side of the handoff are.
FINAL THOUGHT
Great GTM is not built on tools alone. It is built on the people, process, and data that connect them, and AI only strengthens that connection if those fundamentals are already in place. If your team is further along on tools than on the data and process underneath them, that gap is worth fixing before the next AI initiative, not after.
Talk to Digital DI Consultants about a marketing operations and CRM data audit
