Insight
Using AI in RevOps Today, Part 2
February 15, 2024
Our previous article kicked off a discussion into how businesses could dip their 2. into turbulent AI waters, so to speak, by incorporating a few relatively simple tools into their workstreams.
Today we continue that discussion with a look at other aspects of RevOps that can be streamlined with AI integrations, such as data consolidation.
Consolidation of Survey Results or Feedback
Let’s say you’ve automated some aspects of your data collection, integrating Zapier to upload call summaries and analyses instantly. Your sales reps are ready to go out and hit their targets, but it would be helpful if they had more insight into what their customers truly wanted. Your company has a large repository of data from numerous surveys, but no one has had time to parse through hundreds of pages of raw feedback to get comprehensive insights.
While there are a number of LLMs that can tackle this (e.g., CoPilot, Perplexity, etc.), we’ve found that Claude gives the best results when doing this kind of analysis. For one, Claude has an extremely large context/input window, able to accept around 500 pages of text. And in general, Claude has outperformed when asked for transcript summaries, next steps, and action items from lengthy documents or transcripts.
This prompt generated the exact summary our client needed: actionable insights for the product development team in an easily digestible format, which could then cascade down to the sales team to use in shaping their pitches.
Speaking of those pitches, how good do you think your messaging is in presentations? Is that messaging aligned across your organization and is it addressing potential customers’ challenges? How do you know what’s working and what isn’t?
Again, an integration of AI capabilities—in this case ChatGPT—offers truly impactful opportunities for taking your presentations to the next level.
Improving the Messaging In Your Presentations
For something so ubiquitous, sales decks can be the contentious third rail of a team’s messaging. You start out with a master deck that no one is 100% happy with (sales reps wanting their personal stamp on everything? Say it isn’t so!), but it captures the key points that your company wants to get across. Then one person decides to tweak it a bit (“oh, I really like this slide from our old deck better”), then another (“I’m just going to add some flashing graphics to this page, I’m sure the clients will like it”), and the next thing you know, total chaos. And no one has any idea what actually is or isn’t working.
To find out in a logical and methodical manner, we turn to our friend ChatGPT, with its new vision capabilities. To get started, take a handful of presentations that your sales team is actively using, and ask ChatGPT to do the following:
Compare these multiple presentations to find what’s common across them and what’s different (i.e., does your best salesperson have something others are missing?)
Compare a salesperson’s presentation with an official corporate presentation to see if key elements are missing (do they have year old architecture drawings or lists of features that are no longer supported?)
Ask ChatGPT which of the presentations has the most compelling content and the best visuals (this can take multiple prompts and requires giving it very specific guidance – but the results when done right can be amazing)
Ask ChatGPT for specific ways to improve the presentation and messaging (“Please analyze each slide in this deck and make one recommendation per page to make the content crisper or more appealing”)
ChatGPT can analyze each slide in a presentation to provide answers to these types of detailed questions. It’s not quite as straightforward as some other tools; you do have to do some experimenting to figure out which prompts will land. Compare that though to what your sales team is currently doing, i.e., throwing random slides at the wall to see what sticks. Having a presentation that your customers respond to vs. one that has them suddenly remembering a call they have to make can mean the difference between landing the sale or not.
Conclusion
There’s a lot to figure out these days when it comes to incorporating AI into your company’s workflows, but the rapid innovation in AI shows no sign of letting up any time soon. We’ve started you off with some easy tools to integrate, just to see the power of AI in everyday processes, but if you want to talk AI on a deeper level, feel free to give us a call.
One recent client had hundreds of pages of raw feedback in the form of text across three major categories. We used the following prompt to capture information that could be used by the product development and sales teams:
I'm uploading product feedback from some end user testing results. There are three sections in here. Can you give me a detailed summary of the key issues and findings from each section and give me a list of the most commonly found keywords in each section as well. After summarizing each of the three sections provide a final overview summary and key issues found that cut across the document.
