Jiminny Prompt Writing For Revenue Teams

Welcome to your webinar everyone. Um, thanks very much for joining and today we are going to talk about prompt writing for revenue teams specifically. So to introduce myself, I am Lauren. Um, I'm an enterprise customer success manager at Jiminy and I've been here for about a year and a half. Um, and I'll let Petco introduce himself. — Thank you so much Lauren. Obviously my name is Petco, customer success leader at Jiminy. been with Jiminy for like last three years and very excited of course to show you what's up today. Thank you. — Awesome. So going through what we're going to cover today, we are going to be covering for the next 30 minutes how to maximize your insights from your conversation data using ask Jiminy anything. We'll also be covering prompt writing best practices, how to go about writing prompts, how to get started if you've got no idea where to start, you will have at the end of this webinar. And we're then going to dive into some real life examples of how you can use Ask Jiminy at scale. So, you might be familiar with ask Jiminy anything and you most likely will have used this on a call level to dive into a call to either work out how well you've performed or if you're a manager to see how well one of your team members have has performed on that call. You might also have used Ask Jiminy anything on a deal level to understand what's happening across that deal, whether there's any risks or blockers. Now, we've just expanded Ask Jiminy and we've launched Ask Jiminy Panorama. And with this, you can look at multiple calls at once. So, you're able to get insights at scale. And at the moment, we can look at up to 100 calls at once to look at trend data and insights across those calls. So whether it's you trying to understand what's happening across the marketplace within how your competitors are positioned against your product or you're trying to understand as a sales uh manager how your team's performing against discovery calls in Q3 specifically and how they can improve into Q4. So to do this, the first thing that you need to do is understand how to segment your data to make sure that the calls that you're looking at are relevant. Um, and I'm going to hand over to Petco to talk a bit more about that. Thank you, Lauren. Now, let's look into how we turn pile of unstructured conversations into something we can analyze. For today's presentation, we actually picked six but very critical uh search criterias we commonly using across our team and our clients as well. So starting from team so for team if we analyze all code togethers without filtered out on first the specific team we can see the data blurred across and we technically can get insights that are not relevant to what we're looking for. the meeting type specifically. Um I would definitely take a minute here to um mention about the recently released uh feature. So from now Jiminy is able to automatically define the meeting types across your conversations. That's important as uh it's really we are going to touch upon that in about five minutes but it's important when you would like to analyze for example your discovery calls and they're not going to be the same as your QBRs or negotiation calls. So that's why meeting types is so it's becoming so powerful feature to use in search criteria the deal stage at co and current stage I would very briefly put uh like a highlight here to as they are very common at the same time there is a slight difference between both when you record your conversation Jimny is capturing the metadata of course analyzing and transcribing your conversation but when it comes to deal stage at co Jimn is capturing what was the current deal stage when this call has technically occurred where we are mentioning the current stage we can do analysis backwards. So for example what's the current stage of deal stage of the current opportunity and what are all the related meetings to that specific opportunity. So deal amount we obviously we are not going to uh take the same approaches to 5,000 deals or half a million dollar deals for example and that's why it's a very powerful search criteria to analyze strategically your [clears throat] opportunities. The code duration is very powerful for SDRs and BDRs teams as we know uh it's quite often to hear that they have a KPI called effective calls above 3 minutes for example for different companies in different teams then that can be different but for example if you

filter out all of your conversations around above 3 minutes so that might very effectively filter out the voicemail conversations that are not containing very great insight. So this is very much it about the uh the actual filters. Once we know how to filter out all the data, it's really important to take into consideration the second piece of the puzzle. So first on the front end we have the data filtered. The second one is how to craft a great best pract according to the best practices great prompt. The prompt is having five different criterias. So this is a framework that you most probably have seen already. However, it's quite powerful and starting with the role and mode. So it's really important to say to to the LLM to the AI how you would like the AI to behave because there that would give you different output different insights. For example, if you set it to behave like a sales strategist or for example coach or revenue operations leader. So that make a difference when it comes to defining your goal. It's very important to be super specific about it. For example, not just like analyze my course or something generic like that. What you can do, you can set for example, identify the top three coaching opportunities across my discovery questions. We're going to get into that one in the live demo. Return format. The LLM at the back end is quite powerful to return the output as bullet points. For example, table format and something very interesting like heat maps. We are going to show you later today. Guard rails and warnings is technically where we set the rules such as if we call does not mention something do not guess about something. So you can make it and fine-tune it very easy. And the last bit is the context dump which is all the information we give the model to work with. So for example for panorama specifically it's not just the transcripts but all the related metadata to the activity. So you can for example mention in under the context on that you would like to specifically focus on the transcripts or just the metadata or the amount of the deals that these activities are related to. So you have full flexibility to craft a prompt in order to analyze everything that you need. So thank you Lauren. — Thanks Petco. So now Petco's talked a bit about how to segment your data and how to get started with building a prompt. Hope you feel a little bit more confident with that. You might be thinking is it is this relevant to my role? Um the answer is yes. This Jiminy Panorama is relevant across all the types of um revenue personas that we work with. So if you're a sales rep for instance, you're an AE, you might want to understand um how can I improve my discovery skills alongside and in comparison to a top performer. So looking at their calls and looking at mine, how can I get to that stage and then essentially improve my closing rate? For instance, it might be that you're a sales manager and you've got a team of uh BDRs or AES who are using a specific sales methodology. You've implemented spin recently, for instance, and you want to understand who is really owning that sales methodology and following it through all of their conversations with prospects up until close. You can use it to look at that. Or it might be that you're a CEO of a company and you want to understand how your product is positioned in the market against some of your competitors so that you can use that to help direct your go-to market strategy alongside your product team, your marketing team, your customer success team, sales team, etc. So there's a huge amount of ways that you can use this. and to dive into how this looks in Jiminy. We're just going to show you a few different ways um a few different examples of this. So, I'm coming in here as a sales rep. So, I'm pretending I'm a sales rep. Now, I've made sure that all my filters are set right. So, I've only got my calls showing if I was Nick. Um come down here. You can see that the time period I've set is Q3. So, I'm going to look at all my Q3 calls and I'm going to filter that a little bit further to go get that relevance in terms of what call types I'm looking at. I'm only [snorts] going to look at my discovery calls. And Petco touched very briefly on this, but we have now launched our auto detect activity type, which is absolutely brilliant at accurately tagging the call type automatically

automatically um against the type of call it is. So, I've got 59 discovery calls that I can look at. And remember, you can use Ask Jiminy anything panorama for up to 100 calls. So, 59 calls is a really nice amount to get um a good um good idea about how I'm doing across my discovery calls. And I've got a ready to use one I made earlier prompt that I've built. And you might recognize the layout of this prompt. Petco talked about it earlier and we've got the roll and mode, the goal, the return format, guard rails and warnings, context dump, live input and work plan. And that was the exact structure that we are using as our best practice structure. And very quickly, even though it's looking at 59 calls in chat GPT style, we're going to get some really nice insights straight from all of those calls. So Jiminy AI isn't just looking at one call. It's looking across all of those calls and giving us a really nice idea of how I'm performing as a rep across all of these. And once it's finished, just go up to the top and we can see how much information we've got. So it starts with a really nice clear summary assessment. So how I'm doing in general, it's then going to go into a little bit more detail and break it down into areas of discovery. So very quickly I can pinpoint here that I'm really strong on my value linking. So my value linking out of five is 4. 3. So I know that I can probably tick that off the list and think right that's I've kind of nailed that. Um what I need to focus on maybe for next quarter is pain deepening. So digging a little bit further into that pain. What are the pain points for the customer or prospect? Also my qualification looking at medpic or med medicmed medic. So aligning on understanding who the economic buyer is, um any budget constraints. So very clearly I can see there's two areas that I can work on for my next quarter to improve this. It then goes into my strengths and my opportunities for improvement. So this is really nice for self coaching. If you're somebody who wants to do really well, which I assume most sales people do, most customer success people do, you can come in here and really understand where you can improve and what you're really good at. Go down a little bit further and we've got recommended coaching activities, um, actions, and then a nice valuation at the end. So, really nice, nice and clear. I'm now going to pass over to Petco, who's going to talk about this from a sales manager's perspective. — Appreciate it, Lauren. So as a sales manager um we can say we can set ourself a goal for example two goals actually for example drive revenue and of course set my team for success. Now what we are taking in terms of search criterias today is we are analyzing our UK sales team uh conversations but this time we've picked the current stage of the opportunities. So now we filter out 69 activities, 69 conversations that are related to successfully closed deals. Now going into uh panorama, we have prepared for you a quite interesting prompt actually that would analyze the last quarter successfully uh closed deals and that will show us right now a behavior um successful behavior patterns across these conversations. So what I'm going to do with that data, I'm going to take it with me for the next quarter and try technically to apply all of these best practices that Jim Mini has analyzed and um of course extract from these conversations in order to like replicate that success. You will see this prompt starting with a heat map of winning behaviors by rep. Now each rep is we we've set five different criterias. Discovery quality, value framing, objection handling, deal momentum and closing effectiveness. So you can fully tailor this heat map criterias of course according to your specific needs. And now very quickly I can see where exactly is the lacking of performance across all of my team and how we can uh how we can drive performance. Now if Lauren can you please just scroll down a little bit to the bottom. The other piece is that all the data that's been analyzed panorama is going to give you the quarterly plan how to implement these best practices through the next quarter. So starting with master CRM and dialink integration qualification this is specifically for giminy in toalize the AI customization as a core differentiator and etc. So it gives you a quarterly plan that you can follow if you would like. Of course, it's not a must. However, it's a great starting

point in the in in your day like to start with this the every beginning of the quarter. You can refine it and you can make sure that you are of course aligned to your previous like experience, your team's experience in the past quarter. So this is a very good use case for managers. Lauren is going to tell us a little bit more about the executive level representatives of course. Uh so thank you so much Lauren for now. Thanks Petco. So yeah as Petco said um so we've kind of covered the reps and we've covered the sales managers. So um we're now going to look at leaders. So, say if I was a CEO or I'm a sales director, a revenue leader who wants to come in here and I want to get really quick, sharp, concise feedback across multiple calls. I don't want to be I don't have time to be digging into specific call types, looking at each of them on a daily basis. I want to come on here maybe on a weekly, quarterly, monthly basis and have a look at trends across multiple calls. What we're going to be focusing on here as a use case is our close lost customers. So, as a CEO, I might want to come in here and understand looking at all of our customers we had conversations with in the last half of this year who then churned. Why? Why did they churn? Were there risk indicators that we might have missed? are the is are these something we can look at and use as an indication to mitigate risk in the next half of next year for instance. So if I come up here to ask Jiminy anything again we've got a nice example of a prompt that we've built. So we're looking at 21 calls now. So I'll just start this again so you can see it working. and it's here CEO disco strategic insights. So we're going to look mainly at the discovery calls but strategic insights is what we're going to be focusing on now. I want to know as a CEO what should the team be doing next year to stop this happening again and to essentially reduce churn across our client base. So looking across those 21 calls, we've got different elements that it's broken down into. So just give it a minute to load. And we've got five different elements that we've broken it down into is to keep it nice and sharp. So the main thing I'm going to be focused on is this end two elements at the top. If I just go up here, we've also got a nice deal summary. So a summary as always just a very clear outline of um the kind of key themes key indicators. Go down a little bit further we've got strategic insights. We've also got revenue impact assessment. The key elements I'm going to focus on is the cross team implication and the recommended CEO actions. So cross team implications here I'm going to get really specific um directions for each team. So I can then go to my head of marketing and say look this is what we need to do as a company to mitigate churn in the next half of the year. We need to prioritize our CRM integration. We need to u make sure that we're populating the right data etc. I could even ask Jiminy to dig into this area a little bit more if I wanted to. Um I can then go to my head of marketing or my marketing director and say this is what we need to do with the marketing um strategy for next year. And then below here we've got recommended CEO actions. So in terms of product focus, um contraction and pricing transparency issues, develop and communicate a clear data portability strategy, refine go to market messaging. So, it's all very clear and it gives me specific examples of what I should be doing and what the team should be doing for next quarter without me having to dive in and try and find these uh details and these trends myself. So, we've given you a few examples of reps, sales managers, and leaders. Obviously, the options are endless. There are so many other um different ways to do this. So, please do speak to your CSM at Jiminy if you want to find out more and they'll be happy to give you um a trial of Panorama and show you how it would work. So, we've got time for some questions now, so we can move into our Q& A. Um, and just looking at the chat.

I've got a question here that I can start answering. I don't know if there were any more that uh people want to write into the chat with. Um, but somebody has asked how can you share these with other teams? And that's a really good question. So, you can actually share these. I can show you here. Once you've come in here and you've designed a prompt and say you're you love it, you think it's working really well and actually you might be looking at this one. You might be a sales rep and you've written this one that's great around discovery calls and you want to share it with your team, your teammates. Um even though you're competitive, you want everyone to be doing well. You would come into your manage prompts at the top here. So I just click this bit and as long as you've written the prompt yourself, you have the ability to share this with other teammates. So for instance, I can come down here and I can actually select specific users to share this with or I could just find my team, click that, apply it, and save the prompt and it would be immediately shared with them. So, as you probably know, Jiminy is very much we encourage sharing. Um, so that's definitely something we would want you to do. Are there any other questions? There's another one about how to how can you write the best prompt for different teams, Petco? So, do you want to talk a little bit more about prompt writing best practices for different teams? Maybe if you're I don't know a CS person versus a salesperson and how you can differentiate. — Yeah, no problems, Lauren. So, yes, it's before you start crafting your prompt, it's really important to like write down in a very simple like um with a very simple words what specifically you are trying to achieve. how you see the output, how you see Jiminy gives you the output when analyzing the data set. Just focus on these two pieces. How to filter it out your uh your unstructured data, how to make it from unstructured to structured data. Of course, um your customer success managers can definitely help you out with that if you need some help. So, no worries. And of course from there what you would like what's important for you what's critical for you to understand from this specific data set because there are companies with like really thousands of conversations and if you go to analyze like the whole bunch of like thousand 5,000 conversations there is a really great chance to have like first the LLM of course to hallucinate and of course not giving give you the gen generic data but at the same time also to have a very like blurred insights around all of that big data set. So just to summarize it, think about two things. What specific data you would like to analyze, how to filter it down, how to narrow exactly that data set, and of course what you would like to get from that data. And then you can of course start writing your prompt according to best practice framework. — Thanks, Petco. Any other questions from anybody? If you do have any, remember you can speak to your CSM um and they'll be happy to take you through panorama. I think yeah. — Yeah, leave it there. Anything else, Petco? — No, I think we uh we are ready to go. Um, if you have any questions, please feel free to reach us out. That's fine. — Yeah. Thanks very much everyone for listening and yeah, hope to speak to you soon. Have a good rest of your day.