Generative AI is having a bit of a moment. The release of ChatGPT by OpenAI late in 2022 started a ton of conversations around the latest advancements in artificial intelligence. Some of these have been incredibly funny to read—others are a little more ominous.

One question people have been asking lately is whether ChatGPT and other chatbots with the same ability to write human-like language will someday take over from people in offices and classrooms. Will AI eventually replace the PreSales team? 

We don’t think so. This technology is nowhere close to replacing all the great things that a PreSales team member brings to a deal. However, it is on the way to helping create key leverage and improving PreSales team effectiveness. 

In fact, Vivun customers are already seeing tens of millions of dollars in revenue impact from these innovations. Our platform uses natural language processing to assist PreSales teams by helping automate mundane tasks and point the way to better, data-driven decision making, with the same core technology that underpins ChatGPT and other large language models (LLMs). 

Wait, what the heck is ChatGPT?

ChatGPT is a chatbot designed to interact with users in a conversational way, and is built on GPT-3, which belongs to a family of natural language processing models commonly referred to as generative pretrained transformers. These large language models (LLMs) are trained on massive datasets to generate text that sounds as if a human wrote it. 

Language models themselves are a kind of probability distribution that aims to determine how likely a given sequence of words will occur in a sentence, and they’re used in computational linguistics and artificial intelligence to help computer systems properly understand human language. 

Something that’s easy to overlook with large language models is that they’re not able to think on their own; a ton of pre-training is needed for them to work. For example, users have given ChatGPT prompts asking it to write code in a specific programming language, such as Python—and the reason the chatbot is able to do this is because OpenAI included code written in Python as part of the model’s training dataset. 

Without that domain-specific knowledge added beforehand, the answers you’d get on a prompt like “write me a Python program that does X” would be completely nonsensical. 

We’d argue that the most significant part of ChatGPT is actually in the “T”, or Transformer. Until relatively recently, large language models were designed to process input data sequentially—meaning that sentences could only be processed one word at a time. Transformers are a type of machine learning model that use a mechanism known as self-attention to process entire input sequences at once. 

In simplified terms, it’s as if your computer was capable of understanding the meaning of sentences as a cohesive whole in the same way that human beings do, rather than having to process the meaning of each word one at a time before determining what you said. Language models built on Transformer architecture are much better at learning to read and write in ways that are believably human. 

Harnessing the power of LLMs and ChatGPT for the PreSales Team

A natural language processing model built on Transformers to understand human language is actually a perfect fit for helping PreSales teams bridge the gap between customers and R&D. Product managers and engineering teams are typically drowning in mountains of feedback—from support tickets, customer advisory boards, and of course—Sales and PreSales teams who tell them that their deal is at risk if a required feature isn’t shipped in time. 

But most people can’t sift through the noise efficiently enough — they get buried in duplicate asks from multiple sources even when they’re already working on the requested feature, or customers call the same feature by 10 different names, making it hard to spot meaningful trends in the data.

What if AI could help them do this faster? 

Vivun maximizes your R&D investments by combining PreSales expertise with AI

Our first product, Hero by Vivun®, was designed for PreSales teams to scale their efforts effectively and bring product insights backed by revenue numbers to R&D so they can make the right roadmap decisions. It’s already using some of the same technologies that underpin ChatGPT. 

We don’t just act as a place for the PreSales team to take notes on what missing features are a deal-breaker for the prospect—this information is analyzed via natural language processing built on s-BERT (sentence Bi-Directional Encoder Representation from Transformers), a model which serves as a foundation for the structures in ChatGPT and other LLMs to helps computer systems more fully comprehend human language. 

Product Gap Intelligence in Hero is able to analyze the descriptions of requested features, and make recommendations to our users on how different product gaps should be categorized together, even when they’re not worded exactly the same. This helps our users see if an incoming feature request is something the R&D team is already aware of, or signs of an emerging trend in the marketplace that warrants future investment: 

Best of all, our customers have been able to impact tens of millions of dollars in revenue by highlighting, categorizing, and addressing product gaps in this way, with PreSales teams bringing the voice of the customer to Product and Engineering. 

Just in the past year, we can see that we’ve impacted over $80M of business in addressed product gaps. Vivun is a fantastic investment. It helps us go get deals, keep happy customers, and bring in our renewals.

Henry Sowell, VP Solutions Engineering Operations, Cloudera

Companies like Cloudera, Recorded Future, Branch, and Puppet all track product feedback from the field using Hero, and leverage AI-powered assistance to cut through the noise, connect feature requests to revenue impact, and influence R&D to help build what matters most to their customers. 

We’re all incredibly thrilled to see what advancements OpenAI and others will continue bringing to the field. It’s even more exciting to know that AI is more real than you might imagine—and it’s already being applied to solve some of your toughest challenges at work. 

Interested in learning more? Sign up here for a demo of Hero by Vivun.