By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
Revenue Operations

AI-Powered Revenue Generation: Putting the Buzz of Generative AI to Practice

swirled squiggle accent

We've all heard so much buzz 🐝 about generative AI, but many of us have not taken the time or had success trying to integrate this new tool into our day-to-day workflows. Conga has a long track record of deploying AI and machine learning into its solutions and technologies. That's why we were excited to host Angela McKay, VP of Customer Service; Chris Thompson, VP of Global GTM Operations; and Paola DiPalma, VP of Technical Support; and hear how Conga has managed to incorporate generative AI into various aspects of their go-to-market motions! 

The group shared the company's policies around generative AI, what's worked (and what hasn't), and lessons learned about processes needed to ensure customers aren't negatively impacted as changes are rolled out.

Where Has Conga Found Generative AI Helpful?

Conga has prioritized introducing generative AI to increase productivity but has had a very realistic (and humanizing) view of its limitations.

VP of Global GTM Operations, Chris Thompson, said, "We're not going to replace people with generative AI. That conversation is not on the table. We're talking about how we make people's jobs easier? How can they be more productive? How can you move faster? How can you make more profitable business decisions and drive value ultimately to your customers? That's what we're talking about."

The areas Conga's team shared they've had success integrating generative AI include:

  • Prospect Research
  • Account Planning
  • Customer Support Research
  • Forecasting Account Risk
  • Summarizing customer call data
  • Reducing repetitive case handling

All the panelists emphasized the need for human oversight and stated that much of what they have rolled out to date has been internal-facing. Generative AI is being used to reduce research time but not communicate to customers without human oversight (yet!). While much emphasis has been placed in the market around using generative AI to create content, it's important to realize that although the tool is constantly improving, it's not at a point where humans can disengage completely from making communications sound human.

What struck us is that Conga is using generative AI across its use cases in a few novel ways:

  1. Reducing repetitive tasks like prospect and customer research
  2. Aggregating multiple data points into a score or flag
  3. Analyzing call data for sales coaching opportunities or customers at risk

Angela McKay, VP of Customer Success, said, "I think most of us know you cannot replace humans with AI, and managing a customer relationship takes human interaction. However, our teams spend a lot of time researching their customers before getting on calls: understanding usage, their support cases, and what's happening in the news with your customers. 

"Leveraging AI to speed up the time the teams spend doing that research really helps." 

Angela continued, "There's also a way we can use generative AI to enhance relationships. We're leveraging a tool to look at our customer health. Many CS teams do have customer health scores that pull different data points in. We're leveraging a tool that helps us talk about and visualize the various components predicting potential churn or degrading NPS rates. It allows us to focus on the areas that that particular customer is struggling with versus having a holistic health score. We can reduce the time to forecast a risk within an account – or potential opportunity, for that matter.

"We have been leveraging a tool helping us a lot with our coaching and identifying risk that may not have been picked up during calls. It has streamlined the time it takes for the team to track notes from the calls and summarize action items, but it's also allowing my management team to go back and do some quick picks of calls versus having to listen to every call. They can quickly leverage those recaps to identify what coaching is needed from an individual or a team perspective. And we're also pulling out some key risk themes from this AI."

Chris Thompson shared how Conga uses generative AI in sales, "We're currently taking transcripts from recorded calls and a conversation intelligence tool. We are putting that into the generative AI, which runs through our current sales methodology and sales process that is being written onto the Salesforce record directly. And then, using an internal product that Conga offers, it composes an email - literally with the click of a button - the email summary, the next steps, the action items, and any questions left to be answered. The prospecting team can effectively, within 15 to 20 minutes of a discovery call – with one of your prospects, a customer call for an upsell, or a cross-sell opportunity – look incredibly professional by sending a message based upon your own sales methodology, out to everyone who attended that call."

Paola DiPalma, VP of Technical Support, shared how her team is leveraging generative AI. "The way that my team uses it is a global search. When one of our engineers receives a case about an issue they didn't know how to solve, they will search for other similar cases to see how they were resolved. They will look at knowledge base articles, and they will look at documentation. These are three different data sources, so they had to do three different searches. Now, it's just one search, and it will bring information from all these three different repositories and a summary or an analysis of the data retrieved with some recommended steps. It has been a significant time saving for our engineers. 

"My leadership team and I use it as well as more of a reporting tool. For example, if we want to know how many cases a particular customer had submitted in a particular timeframe or what was the C-SAT of that customer. Before, we had to go to Salesforce and create a report to get that data. Now we can ask a simple question and receive it as a quick answer."

The Importance of Setting Guidelines & Goals for AI

The Conga team emphasized the importance of understanding how AI tools interact and protect your data. For example, the public version of ChatGPT ingests your data and may expose it elsewhere if someone asks a question. Your data is being "learned," so it becomes public if you provide it with privately held information.

Chris Thompson said, "Our general counsel, as well as our CISO people who are very pragmatic about new technology applications, have to be very concerned about data and privacy."

Due to these concerns, Conga has developed corporate policies about which tools can be used and how. They've also built their own data lakes and use first-party data to train generative AI instead of feeding in external sources for some of their analysis. They also use tools like Gainsight for customer health analysis and call tracking tools for phone call analytics.

We were very impressed by their disciplined approach to setting goals and measuring the impact of these tools once they are in place.

Paola DiPalma said, "We use AI for case management on the internal facing front. The goal is to eliminate or reduce those non-value-added repetitive tasks that our team members have to do so they can shift and focus on more relevant things. When we analyzed the time commitment around these activities, our team members spent about 30% of their time on these tasks. The AI opportunity reduces their time spent on these tasks to 15%. 

"As a technical support leader, I'm always looking at ways of scaling and being able to be more efficient and to do more with less. We've been doing case deflection for quite a while, but AI is taking it to the next level. It's helping our customers be more self-sufficient with the resources that we provide in a more personalized and curated way. Then, they can engage us when they have more complex issues that will need our engineers' assistance. The time commitment around these activities is around 70% of their time, and the time-saving opportunity that we see is to decrease the workload to around 30% of their time with the potential to gain more as customers become a little bit more familiar with and more confident with the resources we're providing them." 

Lessons Learned on Rolling Out Generative AI & Minimizing Risk

The Conga team had the following recommendations for our audience:

  • Don't build when a tool already exists that you can buy
  • Build a plan so you don't waste time on creating something that isn't a substantial win
  • Test everything on internal-facing use cases before you make anything public-facing
  • Don't unquestioningly trust AI-presented data as fact - test by comparing it to your manual until it can be trusted

Chris nicely summarized the most significant risk with generative AI. "Avoid the bright, shiny object-chasing squirrel syndrome. Everybody thinks they've got a perfect example of what generative AI can do and the value it can provide to the business. And if you chase every one of them, you'll never get anything done. We took a pragmatic approach internally of what we were trying to solve and then tried to apply the generative AI against it and not vice versa."

Chris Thompson shared some of the don'ts they learned along the way. "Don't copy and paste. This is specific to content generation. We've made this mistake. It comes off as very disingenuous. People can see it. It feels robotic. Personalize as much as you can. If you use generative AI in outbound communications, whether in emails, write-ups, documentation, or product sheets – try to keep the human voice in it. All these tools, including the ChatGPT 4.0, are not perfect. They're getting better every day, and they will over time. But make sure a human has the touch. Never take something straight out of it and send it immediately without some oversight to your prospects and customers.

"Everyone's pumped about generative AI, but it will not solve everything. It can solve some things well and struggle with others, but it is not the silver bullet." 

Our audience had some brilliant questions about minimizing risk when implementing AI, particularly when exposing customers to poorly crafted communications or incorrect data.

Paola DiPalma said, "We started with internal facing use cases to feel comfortable with the results. On the customer-facing use case, we're still in the process of testing it. With case deflection, we're being extra cautious to ensure the customers have the best experience from the first day. My recommendation is to do a lot of testing. Take your time with it and start with internal facing use cases you can validate before going externally.

"The approach I took was that I didn't release these tools directly to our engineers," Paola continued. "We first started with my leadership team and extensively used it before starting or sharing it with our engineers. We have a feedback mechanism in which you have the typical thumbs up or thumbs down on a box to provide specific comments about why you are giving the particular rating. So that has been helpful, and that also has been helping and improve the overall accuracy of the tool." 

Angela McKay agreed, "It's been hard for me to want to trust the data. But I've done a lot of double-checking to ensure I truly trust what I'm seeing before deploying it to the team. I've been very careful about throwing something new to the team that will just weigh them down or cause them more stress. We're asking a lot of our frontline teams as it is. So I've done a lot of that kind of validation and partnership with some of my other leaders and partners across the business before deploying it. And then my trust has grown."

Chris Thompson said, "My take is that you should have some skepticism, right? You should. However, at the same time, the overwhelming consensus is that it is here to stay. It is a productive, good thing when used correctly - solving the right problems. But to be a natural skeptic of something that is so. It shouldn't be a matter of being a skeptic. How do you mitigate and minimize those risks and do it intelligently to get value out of it? "

For more best practices from Conga on making the most out of generative AI in sales, check out their ungated guide by clicking here. To see Conga in action, request a demo by clicking here.

Looking for more great content? Check out our blog and join the community.

Related posts

Join the Co-op!