Conversations with Ops professionals across the industry make one thing clear: AI is no longer a fringe topic in RevOps. Every team is either piloting it, planning to, or fielding executive pressure to show results. But adoption is moving more slowly than the hype suggests, and the reasons are consistent across organizations. Three roadblocks keep coming up:
- InfoSec mandate — For most medium to large enterprises, the information security team has blocked the use of commercial AI services due to security, compliance, and IP concerns. These organizations are building internal AI models, but they are slow coming.
- Usability and feedback loop — Current GenAI is better at open-loop tasks such as summarizing or writing drafts. It is more difficult to get AI to produce precise and repeatable results. To help you understand the challenge, here is a good article on why AI can't spell the word "strawberry". Prompt engineering often takes as much effort as traditional methods to get precise and repeatable output. The feedback loop for improving AI output is also different from traditional software configurations, and the impact of that feedback is not immediate and precise.
- AI doesn't exist in a vacuum — No one is ripping out existing solutions wholesale and replacing them with AI versions. Rather, the need is to inject AI into existing processes and technology stacks for improvement, which introduces integration, change management, and data quality complexities just like every other technology introduction.
The third roadblock is the one most teams underestimate. Data quality is one of the biggest prerequisites for AI success. AI is possibly the most data-driven technology humans have ever invented, so the classic garbage-in, garbage-out challenge applies to AI in spades. AI is likely the one technology that will expose your data quality issues the most — which may finally raise the awareness and urgency of dealing with your GTM data quality issues.
And the urgency is real. According to Gartner's 2025 Hype Cycle, 80% of AI projects never reach production. MIT Sloan research from the same year puts it more sharply: 95% of AI projects fail to deliver measurable ROI. The reason, consistently, is not the model. It's the environment the model runs in — the data, the governance, and the accountability structures that determine whether AI outputs are trustworthy enough to act on.
The scale of the underlying data quality problem makes this more urgent than most teams realize. In the 2025 State of RevOps Survey — conducted with RevOps Co-op and MarketingOps across more than 150 operations professionals — 99% of respondents reported struggling with technical data challenges. The specific breakdowns tell the full story: 80% have missing or incomplete data, 75% have duplicate records, 59% have non-standardized data, and 52% have inconsistent data across systems. These aren't edge cases — they describe the baseline condition of most B2B GTM databases. When you layer AI onto a database in this state, each of these problems becomes a source of compounding errors: bad prompts, false positives in scoring, misdirected personalization, and hallucinated outputs that are nearly impossible to trace back to their root cause. Fixing data quality before deploying AI isn't a preparation step. It is the prerequisite that determines whether your AI investment produces reliable outputs or quietly misleading ones.
Prompt quality = data quality
A prompt is the instruction you give to an AI model to execute the task you want it to perform. Just as you would expect from giving instructions to a junior employee, AI's ability to successfully complete the task heavily depends on the quality of the instruction — hence the emergence of the new profession, "prompt engineer." For most RevOps use cases, constructing AI prompts will involve embedding various GTM data into the prompt, such as account name, website, contact name, LinkedIn profile, and email address. If your GTM data is of poor quality, it would be hard for AI to perform the task successfully.
For example, if you're asking an AI agent to perform web crawl and research on a prospect, but the person name and LinkedIn URL you provide in the prompt do not match, and the company name and job title are from two jobs ago, you're hampering your AI agent right out of the gate. And unlike a human employee, AI has yet to master common sense and ask good clarifying questions.
Openprise's AI Orchestration capability addresses this directly through context orchestration — preparing high-quality, enriched data with the relevant context an AI model needs before a prompt is ever constructed. The prompt management layer then dynamically builds and versions system and user prompts, ensuring consistency and security across every AI task that runs.
Trust but verify your hallucinating AI
If you have read about or used AI at all, you are familiar with AI's tendency to hallucinate. If you're open to reading a hilarious yet extremely informative academic paper on this topic, ChatGPT is bullshit discusses why "hallucination" is the wrong word and why "bullshit" more accurately describes this AI shortcoming. Humor aside, hallucination is a major technical challenge we have to account for if we want to use AI to scale operations.
Comparing AI to a junior employee turns out to be a very useful framework for understanding what it takes to get the most out of it. Any manager knows you have to trust but verify, especially with junior employees, and this is equally applicable to AI. To verify AI's work, you will need high-quality data. If you don't verify your AI's output, your AI's hallucination can pollute your data with fiction and create more data quality issues than it can remediate — at record speed.
This is why hallucination management is one of the six operational pillars of Openprise's AI Orchestration framework. It automates the validation and remediation of AI outputs to catch errors before they propagate downstream into your CRM, your MAP, or your scoring models. CrowdStrike, working with Openprise, achieved over 98% accuracy in persona assignments using an AI-based LLM for segmentation — a result that required not just a capable model, but rigorous output validation built into the process.
Until AI can explain itself, you will have to
All the cutting-edge AIs we see today are black box technologies. AI cannot yet explain how it comes to a certain conclusion; as much as we would like to believe it, AI doesn't actually exercise logic and reasoning. Instead, it's pattern matching and generating humanlike responses. The paper cited above also explains that what people believe to be AI explanations are actually not logical explanations, but simulated responses of what a human explanation would sound like — facts be damned.
So, if you're going to use AI and AI can't show its work, then when challenged, you will have to explain AI's output on behalf of AI. If your data quality is good — and you therefore have confidence in AI's output — you at least have a fighting chance to explain why and how using the data you prompted AI with. If you have no confidence in your data's quality, then you shouldn't have any confidence in the AI's output, and certainly shouldn't try to explain or defend it. This is what Openprise means when it talks about building trust, not just pilots. The goal isn't to run an impressive demo. It's to deploy AI in a way that produces outputs you can stand behind.
Personalized content is irrelevant without precision targeting
One of generative AI's most powerful use cases is the creation of personalized content to power tailored buyer's journeys. However, personalization only works if it is delivered with precision to the target persona and buying group. For example, you can create highly personalized content targeted at a CIO, but end up using that content on a security engineer prospect because you don't have the right job title or cannot properly segment it to a job function and job level that can guide the targeting algorithm. All that investment in AI personalization is at best a wasted effort; at worst, it backfires and creates a terrible buyer experience.
Personalized content is only relevant if you can target precisely, and good-quality data is essential to that process. Without the ability to target precisely, AI personalization is not so much a precision GTM weapon, but a weapon of mass destruction. Denodo's experience makes this concrete: working with Openprise, they successfully classified over 90% of all leads and 94% of contacts — giving their AI-powered programs the reliable segmentation data required to reach the right buyers with the right messages consistently.
What data quality AI-readiness looks like in practice
The four scenarios above — bad prompts, hallucination verification, unexplainable outputs, and misdirected personalization — all trace back to the same root cause: GTM data that hasn't been cleaned, standardized, enriched, and maintained at the level AI requires. The teams that have invested in that data foundation before deploying AI are already seeing the results.
Equinix is a useful example of what this looks like at scale. Their Demand Center Operations team was facing constant manual effort managing lead data from disparate systems — exactly the kind of data fragmentation that breaks AI prompt quality and targeting precision. They worked with Openprise to automate data quality, enrichment, and segmentation so that leads arrive already enriched at the moment of creation, with accurate job level, function, and account association in place before any downstream automation runs. The results were significant:
- 130% improvement in lead-to-account match rates — the foundational match that AI-powered account-based plays depend on
- 5 weeks of work saved annually by automatically identifying and purging false leads
- 1,000 hours per year eliminated from manual reporting
As their team put it: "Marketo, Salesforce, and Openprise are our 3 most critical systems within our tech stack."
Nutanix faced a more severe version of the targeting precision problem. Their job title and function match rate in the APAC region was only 15% — meaning 85% of contacts could not be reliably segmented for targeted or AI-powered personalization. After automating job title classification and data quality processes globally with Openprise, the APAC match rate rose from 15% to 75%. That improvement directly enabled accurate persona-level targeting, and the downstream effects were equally significant: duplicate and inactive records were eliminated, reducing their CRM account count from 650,000 to 180,000, and removing the data contamination that causes AI scoring models to surface false positives. The team saved the equivalent of 15 FTEs in manual data work per week. "Openprise is a key pillar in our data quality and automation strategy."
Zendesk illustrates what happens to ops teams when data quality is poor and manual cleaning is the norm — and what changes when it's automated. Manual data cleaning and standardization was delaying lead routing and keeping their ops team in a reactive posture, spending time fixing data rather than building AI-powered programs. After automating data cleansing, enrichment, normalization, and deduplication with Openprise, they achieved a 25% improvement in data cleansing efficiency, a 25%+ increase in marketing and sales team efficiency, and generated $500,000+ in productivity gains. That freed the ops team to focus on strategic work — including AI program development. "There's no way we could have done this without Openprise. It was like Openprise was a silver bullet."
AI is an executive's wormhole into bad data
Unfortunately, in many companies it's hard to make executives care about data quality and invest in it. One of the reasons is that most executives consume data in the form of reports and dashboards. They don't ever see the raw data and the sheer amount of human effort that goes into cleaning and preparing the data for these reports. A data analyst might have to spend 20 hours collecting, stitching, and scrubbing data every month just to generate one report, but none of that pain, effort, and cost are visible when an executive looks at a polished dashboard.
The vision being sold by AI vendors is that every executive will have a co-pilot at their fingertips, where they can type in questions like "What is the difference in ACV and NRR between our ICP and non-ICP customers?" and AI will serve up the answers instantaneously. For those of you data people who have managed to stop ROTFLYAO — AI is promising your executives a wormhole directly to your company's raw data. Out goes the human abstraction layer. Maybe this is a good thing, because AI may finally give executives an unadulterated view of how bad most companies' data quality actually is. Maybe this will finally make them care about and invest in their data quality and infrastructure.
That shift is overdue. As Openprise's own research shows, 74% of CEOs believe failing to deliver AI results puts their jobs at risk, and 64% of executives report "AI fatigue" within their organizations. The pressure to show real results from AI — not pilots, not demos, but production deployments with measurable ROI — is arriving fast. The executives who will navigate it successfully are the ones who treat data quality as the strategic infrastructure question it actually is, not a technical detail to be delegated downward.
How to get your GTM data AI-ready
If the four failure modes above describe where AI goes wrong, here is what the preparation looks like to get it right. Getting your data AI-ready is not a single project — it's a continuous operational process built on four foundational layers.
Cleanse and standardize first. AI prompts are only as reliable as the field values they pull from. Before any AI agent runs against your CRM or MAP, your contact and account records need normalized job titles, standardized state and country fields, consistent company name formats, and validated email addresses. Non-standardized data doesn't just produce bad AI outputs — it makes those outputs untraceable and unexplainable, which is exactly the problem described in the "until AI can explain itself" section above.
Deduplicate before you enrich or score. Duplicate records split engagement history, inflate lead counts, and cause AI scoring models to rank the same person twice with different scores. A contact with five touchpoints spread across three duplicate records looks like three low-engagement contacts to a scoring model. Deduplication needs to happen before enrichment, and enrichment needs to happen before any AI model runs.
Enrich to fill the gaps AI depends on. The personalization and segmentation use cases in this post require specific fields to be present and accurate: job title, job function, job level, company size, industry, and intent signals. A single enrichment vendor typically achieves match rates of around 30–50%. A multi-vendor enrichment waterfall — sequencing providers to maximize match rate and fill rate — is what gets contact and account data complete enough to be useful as AI input.
Automate continuously, not periodically. Data quality AI-readiness is not a one-time cleanup project. Contact data decays at roughly 30% per year. People change jobs, companies get acquired, email addresses go invalid. A database that was clean enough for AI in January may not be clean enough in July. The teams that maintain AI-ready data are the ones running automated cleansing, enrichment, and deduplication on a continuous schedule — not the ones doing annual data projects.
Openprise's AI Orchestration capability and AI-agent Factory are built on top of this data quality foundation — so that AI agents run against records that have already been cleansed, enriched, deduplicated, and normalized before any AI logic executes. That sequencing — data quality first, AI second — is what makes AI outputs trustworthy enough to act on, and defensible enough to explain to a leadership team that is now, more than ever, paying close attention.
Want to go deeper on moving AI from pilots to production? Read Openprise's executive imperative for AI or schedule a demo to see how Openprise handles the data quality foundation that makes AI reliable in practice.
.jpg)
