There is a readiness gap in Marketing. Every Marketer wants to be AI- driven, but they are hindered by many different things, the biggest being the state of their CRM data.
Validity surveyed 500 marketers worldwide to understand how they are managing the gap between customer relationship management (CRM) data quality and the AI initiatives that data feeds. One of the biggest takeaways is that 62% have suffered direct revenue loss due to poor CRM data quality.
Agentic AI relies on high-quality data. There is no way around that. However, many organizations move forward with AI initiatives, even when data isn’t ready. CRM data feeds campaigns, drives forecasts, and improves customer relationships. This data links together contact information with engagement history, campaigns, and revenue. It also ties together marketing activities with sales and support.
When asked how much of their CRM data is complete and accurate, only 25.6% said it was 76-100% complete and 45.9% said it was 51-75% complete.
Almost half those surveyed admitted their company struggles with CRM data quality issues. These issues can take hours each week to resolve (39% spend 2-5 hours a week, and 24% spend 6-10 hours a week). Why does it take so much time to clean CRM data? Unclear ownership and lack of collaboration between the different departments that feed CRMs are the primary reasons.
Adobe’s 2026 AI and Digital Trends report also acknowledges there is a readiness gap and data is a top focus:
This readiness gap persists even as organizations acknowledge the problem. When asked about their priorities for AI investments, only 32% named data quality, unification, and governance as a top focus, and just 20% prioritized increasing the value and understanding of data. This is despite the fact that 52% admit that their current data unification and structure limits advancement of AI initiatives, and a full 75% cite data integration and quality as the top challenge for implementing agentic AI solutions.
Bad data hits the bottom line
CRM data quality directly impacts revenue. 44% of C-Suite and SVP/VPs say their companies have lost revenue due to data quality. Missed forecasts, stalled renewals, dropped deals, and campaigns directed at the wrong people are top revenue-loss issues, and executive leadership sees them at a level others don’t.
As Courtney Grab, Senior Director of Customer & Lifecycle Marketing, Validity, says:
Revenue loss from bad data rarely shows up as a single dramatic failure—it shows up as a hundred small ones: a forecast that’s off by a few points, a segment that should’ve converted and didn’t, a renewal that slipped through because the record was stale. By the time leadership notices, the damage is already baked into the number they’re presenting to the board.
Sixty-eight percent of survey respondents said they had a revenue pipeline or performance number challenged or walked back because the underlying data was wrong. From the Validity report:
Bad data can make those concerns a reality—through outdated subscriber preferences, mismatched personalization, and messages sent to the wrong segment at the wrong time. Every instance is a small breach of trust with a subscriber who wants to feel seen, not guessed at. Multiply that across thousands of records and the ‘small’ breach becomes a pattern.
In the State of Marketing from Salesforce, marketers get around bad or siloed data by sending generic campaigns, which 84% of respondents in that survey do.
One of the most concerning things about using bad CRM data in marketing initiatives is when marketers let AI do everything with no human review or approval. In this study, 44.7% do just that (i.e., sending campaigns, scoring leads, and reallocating budgets).
Bad data not only creates poor digital experiences, but also compliance gaps and privacy risks, including stale consent records, missed opt-outs, and duplicate profiles. If no one is reviewing what the AI is writing or sending, these issues only compound.
Board pressure beats bad data
Despite knowing their CRM data isn’t ready for agentic AI, most companies push forward because the board and other leadership pressure them. The report notes that the board only sees dashboards and reports; they don’t typically see the underlying data, so they can’t see the potential problems that can arise when the data isn’t good. Only 56.7% of marketers feel “somewhat” prepared to use their data for AI initiatives and in AI tools.
Knowing what data-ready looks like and actually being data-ready aren’t the same thing. Just because an organization figures out what data-ready means doesn’t mean they are doing the work to ensure CRM data actually meets those requirements.
Laura Christensen, Senior Director of Professional Services & Customer Success at Validity, makes a good point:
Leadership often signs off on an AI initiative assuming the data underneath it is solid, when in reality, duplicate records, stale fields, and gaps that nobody’s flagged are the norm, not the exception. It’s not that leaders are being careless; they’re just several steps removed from the data itself. Their confidence is coming from the dashboard, not the database.
In Salesforce’s State of Agentic AI in the Enterprise, Shibani Ahuja, SVP, Data & AI Strategy, Salesforce says:
Every boardroom is asking whether it’s moving fast enough. Two years into the agentic shift, the answer from the data is that the advantage was never in starting first; it’s in starting deliberately. The organizations getting real returns got specific about a shortlist of things before conditions were perfect: the data they made trustworthy for the job, the point where a person stays in the loop, and the guardrails they built before they needed them.
My take
Of course, diginomica also did solid research on this topic, finding that data quality and change management are the primary drivers of AI value. Reports like these clearly show that data quality is a problem for agentic AI; however, companies push forward and are living to see the damage that results.
And it’s not only bad CRM data that causes issues; it’s bad organizational knowledge, such as support, product, and help content. This type of content also causes many issues for agentic AI, with incorrect product and support information reaching customers through agentic applications like AI assistants.
If the board is pushing hard, leadership must take the time to educate them on the challenges the company faces and what’s needed to overcome them. The diginomica findings point out what has to happen to ensure success:
Technology has never moved this fast. Proof of concepts that failed nine months ago now work perfectly, making it difficult for CIOs to maintain credibility with business stakeholders. Boards that have consumed too much vendor marketing expect transformative results, while the reality on the ground is far more nuanced. CIOs need to become skilled at tempering enthusiasm with reality while still evangelizing genuine opportunities.
How do you do that? Have ongoing conversations about the importance of data quality and how your company’s data needs proper governance. Not only CIOs, but workers lower on the ladder who are closer to the data than anyone. Even before real pilots run, prototypes and deep testing must identify where the data might fail. In the rush to get something out the door and beat the competition, this work doesn’t get the time it needs.
