With all the talk around generative AI, GTM magazine wanted to bring together two people who are deeply engaged in the sales technology market to get their perspective on this game changing technology. In this interview we will talk with Seth Marrs, who covers sales technology at Forrester and Rocco Savage, the Head of Growth at Regie.ai, a generative AI company focused on fully personalized sales outreach.
We spoke to them about the emergence of this new technology and how it will affect sales organizations now and in the future.
GTM Magazine: What do you think of all the talk around generative AI?
Rocco Savage: Whether people love it or hate it, Generative AI cannot be ignored at this point. There is no way to put the genie inside the bottle, so to speak. Just like how every new technology looks scary at first (imagine creating fire for the first time), it will over time show its boundaries and people will understand its limits and how the new world will operate with this new piece of technology at its disposal.
Seth Marrs: It has been amazing how this technology has captured the world’s imagination. Before the ChatGPT announcement, there wasn’t a lot of talk about this from customers or vendors. Since the announcement, I rarely have a call around technology where ChatGPT isn’t discussed.
GTM Magazine: How would you define generative AI?
Seth Marrs: I initially defined it as a tool that could generate responses at the same level as Open.ai’s ChatGPT because that was the technology creating all the buzz. I was concerned that sales technology vendors would start inferring that they had the same level of capabilities, making it harder for buyers to understand what they were getting. As I’ve learned more about the technology, I have loosened my definition to any tool that can create useful responses that help users generate responses.
Rocco Savage: Machine Learning and AI traditional use cases have been around recognizing content. For example, we had speech recognition, natural language processing, etc. Now we have a lot of action on neural networks that can generate human-like content – text, images and videos. That is basically the current paradigm of generative AI.
GTM Magazine: What is the difference between ChatGPT generative AI compared to the current Sales Generative AI products on the market?
Rocco Savage: Chat GPT is good for general purpose uses like a recipe or general blog. With GPT, it’s looking at the whole internet so it can take just as long to write a prompt as it does to write the email. Regie.ai leverages both public and the private datasets that an organization has within this cloud to create custom emails that are unique to their story-telling. This allows the AI to generate content that mirrors the organization’s messaging, allowing teams to scale their content creation needs.
Seth Marrs: The technology behind ChatGPT is revolutionary, so not something every company can do. No company in sales tech can replicate this technology. Sales technology companies using generative AI are integrating with GPT type technologies or using less powerful large language models that are trained for specific use cases.
GTM Magazine: How will this technology improve the lives of sellers?
Seth Marrs: In the short term, the best use case is content creation. Many people only see it as a way to send even more emails, but the real value is in the ability to create a quality emails. That use case is well established and provides a solution to a real problem for sellers. In the long run, Generative AI will be used in sales technology products to generate accurate insights that improve seller outcomes. Next best action has been fairly remedial to this point, but this technology will enable companies to generate insights that are truly valuable to sellers.
Rocco Savage: I agree, this is not a spam engine. These tools improve the quality of communication between the buyer and seller. Most companies are 5% customized with 95% generic messaging. Using generative AI tools will get 20% by the end of this year. In the future, you will no longer have templates. Everything will be customized and rephrased with AI once a week to keep it fresh.
GTM Magazine: Based on the last question, email efficiency seems to be the biggest use case for this technology. How does it make the seller more efficient?
Rocco Savage: It’s less about efficiency and more about productivity gains. Generative AI that is trained to produce custom content that allows sellers to close the gap on their meeting and pipeline goals. The easiest way to solve this is by giving sellers a better way to create truly personalized emails at scale.
Seth Marrs: I believe the biggest use case is in helping sellers efficiently create content for their buyers. One challenge in doing this is getting started. It is much easier to edit than it is to write. Generative AI gives sellers a first draft that they can edit and customize quickly. That saves time and increases quality.
GTM Magazine: How long do you think it will take for this technology to become mainstream?
Seth Marrs: In most cases when I’ve seen technology come out, the hardest part is building mainstream awareness of capabilities and explaining why users should buy it. Generative AI has been the opposite. ChatGPT has made Generative AI mainstream immediately and now companies are racing to capitalize on the technology and get it built into their products. I think email generation will be mainstream this year, but the more powerful use cases like advanced next best action capabilities are a year or more out.
Rocco Savage: It already is in many ways. ChatGPT has been the trojan horse to accelerating the awareness, application, and accessibility of this technology. Every company right now is thinking through what their Generative AI strategy is. The closest I can think of are moments like (1) the internet, when every company had to think about what their online strategy was (2) social, when every company had to think about what their Facebook / Twitter strategy was and (3) mobile, when every company had to think about what their mobile strategy was. Timelines are much more compressed now given the speed this technology is developing, which is forcing every single company currently to think through their Generative AI strategy.
GTM Magazine: How accurate is this technology? How much does a user need to adjust emails created for them?
Rocco Savage: In the context of Regie.ai, we like to say it gets you about 80% of the way there, and then human involvement is required to take it across the finish line and press send. But that is by design. The human-computer partnership is incredibly important to us and it’s important the human is involved in the creation process. I think people often think AI will replace humans, but I believe this is far from true, and probably stems from a fear of change. What it will let us do is allow humans to work on problems they could not before. Generative AI gets you to an editable 1st draft in seconds, rather than hours.
Seth Marrs: Agree and think it’s important to state that it’s not a great idea to use tools like this to do all the work for you. These tools are trained on what worked in the past. Buyers’ preferences are constantly changing and requiring engagement to adapt to their current situation. Sellers need to focus on these things and adapt.
GTM Magazine: A lot of these tools have questions at the beginning to help structure responses. Can you talk through why that is necessary now?
Seth Marrs: Many people see the ability of ChatGPT to write music, build shopping list, pass tests and think getting answers is as simple as asking for them. These systems require specific prompts to get the right answers. This is even more important when trying to get an answer to complex industry specific questions. Adding structured responses at the beginning guides the system to the right response without risking the potential errors associated with entering the wrong prompt.
Rocco Savage: The quality of your output with AI is directly tied to the quality of the input. While humans get comfortable with this dance, as we like to call it, you need a structured application to sit on top of the tech in order to best guide the end content it produces.
GTM Magazine: Could bias in the responses and a seller not paying attention lead to unintentionally discriminatory responses?
Rocco Savage: Absolutely. As it relates to Regie.ai, we’ve included the ability to scan your email for words or phrases that perpetuate implicit biases and strong gender preferences and offer points of education to fix it. That not only helps you course correct an email, but gives you a tool so you can train yourself, and your front-line teams, on being more cognizant and aware of the inclusivity of your word choices. The biggest issues will revolve around softer compliance than information security compliance. For example, now that users can generate a lot of content quickly, it makes it harder for companies to enforce brand consistency and accuracy.
Seth Marrs: I think this is a challenge that all companies in this space will need to monitor. Just one discriminatory email getting publicized can lead to a wave of public criticism that could affect the future of these companies. This will become an even bigger challenge when responses get correlated with positive outcomes using the demographic data of buyers.
GTM Magazine: What is the next frontier for this technology?
Seth Marrs: From a sales technology perspective, I think this is just the beginning. We will see more and more capabilities released that solve a wider range of use cases until users can create their own solutions. Sales technology providers are already evaluating this technology and coming up with new ways to integrate it with their current capabilities and create more value for their customers. Also, once this technology is connected to more relevant datasets, it will open up new insights that will finally allow companies to generate insights that are valuable for sellers.
Rocco Savage: Two immediate changes will happen: Mathematical and logical reasoning will improve dramatically. For example, we will soon be able to solve mathematical and logical problems similar to those in SAT and other standardized tests. Without teaching all the rules of Math, AI will do Math similar to how it can write sentences without us teaching it grammar. Second, the combination of different media types simultaneously will be mind-blowing. For example, combining text, images and even movies. Instead of creating it piecemeal, we will have AI able to generate them simultaneously.