By Dominik Facher, Chief Product Officer at Zoominfo. Zoominfo were finalists in the ‘AI Implementation of the Year‘ category at The 2024 A.I. Awards. 

The race to adopt generative AI is astounding.

We’re seeing top-down investment across every industry, and amazing strides and use cases — from medical advancements to fraud detection — that are truly setting the stage for exponential progress. In fact, AI-related positions have seen a 200% increase in just two years, and there are no signs of slowing down.

I’ve been building AI-related products for more than a decade, so to be clear, the craze we’re seeing is not new. But GenAI has sparked a renewed interest and excitement and has helped make AI more accessible.

What I’ve noticed though, is that many people are focusing on the last mile of what GenAI can do, especially as it relates to using it in your business. But the most important thing to remember is that generative AI can’t operate well — or at all, really — without quality data that is comprehensive and accurate. Here are some lessons I’m making sure my team is keeping in mind when it comes to generative AI.

Lesson 1: GenAI is an Additive

GenAI is a tool, and needs to be thought of as such. No one should be relying on ChatGPT or Jasper.AI to construct a flawless output because you will be disappointed, or worse, be caught with egg on your face in front of your customers and prospects. So if you think of GenAI as an end-all-be-all, or that you can replace the bulk of your staff with AI products, think again.

Despite improvements to ChatGPT’s programming and new model, the software is said to have an 88% accuracy rate, but this varies greatly based on the type of questions you ask. Think about asking the software to tell you about a company that you’re prospecting. Would you trust it? Should you trust it?

From my many years of operating in this space, I’ve learned that 70-80% data accuracy does not do the trick for go-to-market teams. A chatbot can likely tell you the 10 largest companies in Atlanta, for example. However, probing further than that without supplemental information and you’ll likely get stuck with inaccurate data or straight up junk. But if you use the program as a tool, rather than a source of truth, it can be incredibly effective. Generative AI works well when fed the right prompts.

“Can you write the email for me?” and “Can you contextualize the email for me?” or “Can you make it personalized to my customers for me?” are all great questions to ask the program, but you need to ensure you’re providing the best data.

This brings me to my next point.

Lesson 2: You Still Need to Feed AI Quality Data

You can’t get quality outputs without the infrastructural element of data. What do I mean by this? It needs to be up-to-date and give you a full view of the customer that you’re reaching out to; filling in the gaps that will come if you rely solely on a chat tool.

Whether you’re using Microsoft CoPilot, Salesforce’s Einstein, or any other generative AI product, it has to grab data that exists somewhere in your system in order to generate accurate responses. The proprietary data you feed to AI is what helps it identify things, including who you should reach out to, why and when you should do that, and what you should say to them. For example, you can find out what technologies and partners they currently have, whether they just raised funding, how big they are, and so on.

All of this information — which can be leveraged from the systems you already have — is important, but tends to be forgotten since the product can write a great message. If the data in your CRM, marketing automation system, and your data warehouses aren’t accurate and complete, your messaging will fall flat.

This is how my team thinks about AI, and it was the guiding force in designing ZoomInfo Copilot, our own AI-powered solution. With unified, quality data at the foundation and an embedded and vetted AI solution, we’re able to get closer to replicating our customer’s best sellers, their plays, their approach, and their success formulas. Relieving salespeople of time-consuming administrative tasks and instead surfacing for them actionable insights and assistance with crafting personalized messages and follow up steps allows them to build relationships with prospects and close deals faster.

Lesson 3: Don’t Underestimate the Time Required by AI Use

While AI-powered tech promises improved efficiency for employees, workers surveyed by the Organization for Economic Co-operation said that the intensity of their work has increased after their workplaces adopted AI. That’s because there’s still a lot of editing and refinement that needs to be done, regardless of how complex the tech is.

For example, customer expectations and buying processes are evolving as well. Although AI will help level the playing field, buyers will still look to sellers as trusted advisors to help navigate the purchasing process. Don’t underestimate the amount of effort it will take to train the technology, or the amount of time editing will take, especially if you want quality work.

Lesson 4: AI Responsibility is a Must

As businesses increasingly adopt AI, the need for responsible AI practices becomes all the more critical. Ethical considerations, transparency, and accountability should be at the forefront of any AI strategy. It’s not just about what AI can do, but how it does it.

To ensure responsible AI use, businesses should look into obtaining certifications and adhering to governance frameworks. For example, the AI Governance program at ZoomInfo emphasizes the importance of establishing clear guidelines, maintaining transparency, and ensuring that AI systems are aligned with ethical standards. Certifications such as those offered by the AI Ethics Lab and other governing bodies can help companies build trust with customers and stakeholders by demonstrating their commitment to responsible AI practices.

Responsible AI isn’t just about compliance — it’s about fostering innovation that benefits everyone, from customers to employees to society at large.

Lesson 5: There’s Enormous Economic Value – When Properly Executed

The economic value of AI in business is undeniable. Companies leveraging AI are seeing increased revenue streams, improved customer retention, and enhanced operational efficiency. AI-driven automation is reducing manual tasks, allowing employees to focus on higher-value work. For example, AI can identify new market opportunities by analyzing vast datasets, or personalize marketing campaigns to boost customer engagement.

AMcKinsey report estimates that AI could add $13 trillion to the global economy by 2030. Businesses that strategically implement AI are poised to capitalize on this growth by optimizing processes, improving decision-making, and scaling operations. Whether it’s through cost savings, revenue growth, or customer loyalty, AI is driving tangible financial benefits that can make a significant impact on a company’s bottom line.

Generative AI undoubtedly holds tremendous untapped opportunities to add value to your strategy. However, it is crucial to approach the tool with a clear understanding of its limitations and requirements. Prioritizing data quality, investing the necessary effort to train and refine the technology, and treating it as a tool rather than a source of truth will allow your business to unlock the true potential of Generative AI and leverage its capabilities effectively in your operations.