AI Readiness for Sales Teams

Hello, everyone. Good afternoon. This is the fourth episode of our AI Readiness series, and today we will be talking about sales. In this twelve part series, we are covering AI Readiness for every department. The first two sessions were about what it means to be AI ready in general, and then we talked about the different agentic systems in the Microsoft applications, Microsoft stack, and then we started off with marketing last week. Today, I will be talking about sales. Next week, we'll be covering customer service, field service, the week after. And that will conclude the CRM related episodes, and then we move into the more ERP back office, systems such as supply chain, finance, manufacturing, and then HR. And towards the end, we'll cover IT. And as a capstone episode of q one end of end of March, we will cover IT governance. So hi, everyone. I am Balaj Horvath. I'm the founder of Visual Labs. I have been working with Microsoft products all my professional life in its thirteenth year, and I've been involved in large scale implementations in the UK, in the US. Since twenty twenty, we've been building up Visual Labs, which has become a prime partner based out of Hungary, half our revenue coming from the UK and the other half from from Hungary. And today, I will show you what it's like to be AI ready for the area of sales, how you can create custom proposals while still sticking with your standard core offerings. And what's really important in today's world is that you can provide custom quotations and and custom offerings whilst still sticking with your with what you can do best. So last week, I showed you how to identify your ideal customers based on an ideal customer profile, custom GPT. Then when you get a new lead, there would be an automation that automatically runs and does a lead scoring against your criteria. It's a clever prompt and an agentic workflow in the background, and that will enable you to to select the leads that you you can best serve and you can best serve in a profitable manner. So, this week, we'll talk about how you handle quoting and and what it means to be ready for AI in terms of in terms of sales. And there's really two four aspects to this, and I'm happy to happy to share my screen, but it's it's actually really basic. There's when we talk about readiness for sales, I I like to think about four things. So one is your process readiness. And so you you can understand how you do sales, how things are run, what are the basic steps that you want to take forward. After that, you wanna make sure that your you have the right systems in place where you can track the process, track the the tasks, and and where things are, how things are moving along on that on that process. And and and after the process and system, you'll have data to make sure that you can log the necessary data pieces. And once you've got the data pieces logged, you will have the people side of it, obviously. So once you've got the people side covered, you can then really make sure that it gets into the people start using AI in a way that you intended it, and they can start actually improving things. So what I'll show you today is how Microsoft and Microsoft Dynamics supports the system, the process, the data side of it, and how you can use Copilot to help train your people on your sales process. And as I've been I've been growing Visual Labs for the past six years, my role has been always changing. I started off as a consultant back in back in Hungary, US and the UK. When I founded Visual Labs, I was very much a consultant working with with friends and colleagues. As we grew, I had to start picking up new skills such as sales. I am now the the sole salesperson for for Visual Labs. And and that meant I had to had to learn a lot of new things about about how sales works, how how real life, how that happens in real life. Although I've implemented CRM and sales systems, plenty of times, it's very different doing large, b to b sales. So that's that's what I picked up. And and I was thinking, how can I how can I improve myself, and how can I give myself better exposure? Because, you know, these large b to b accounts don't come along every single day. So so what I did is is I started standardizing our sales process. I was working with a sales adviser who helped us bring in best practices. But what's important is once I had the standardized process, I was able to build, actually two agents that would help me practice the standard sales, discovery sessions. So I used, that to create a, a role play agent that would pick take on the role of a c o COO, CFO, or CTO. And that would, that would allow me to, to actually turn on voice mode or or just chat back and forth. I would have another agent, an evaluator agent, who would score me against the different criteria that I predefined and help me improve how I can run better sales discovery sessions. This is the people aspect that I would actually like to show you today, how that custom GPT works. And I think that's it's it's really important that we as we as we grow our talent and and and we improve our our company through AI, we make it natural and we make it part of our standard operating procedures that AI is really becoming part of the vein. That will be the first demo, and I'll actually use that demo to show you how you can create a CRM opportunity in a very, very basic, manner. Those of you who have seen Dynamics, that might be old news. But what's important here is when you actually get to the proposal, you can create standardized but still very custom proposals yet again with a custom agent. And what are the data points that are required to capture along the journey? In the past month now, the way I've been creating proposals is I run these discovery sessions that I just talked about in a fairly structured manner. I was able to obtain the necessary information through the clients and and come up with with solution proposals and and ideas and identify the pain points that that are worth solving. And I would actually have those those meetings recorded, obviously, with the consent of the client. Many times, these are just run through Teams. And once these Teams meetings are transcribed, I would just feed the transcript to this agent, through my custom pro custom, proposal generator, and it would actually come up with a Word document that is pretty much ready to send. I would only have to tweak it. And that has drastically reduced my time of four to six hours to create a proper proposal down to less than half an hour, four to five minutes. And that includes administration, CRM stuff, writing up the email, and all that. So I would like to walk you through that journey today and show you how how how I use the different custom components. And I think what's important is there's there's still plenty of room to to actually automate and integrate all these tools. So what I'm showing today is the different puzzle pieces that are currently stitched together. And this is the right level of automation for me that works for me good enough. But if you wanted to roll this out at scale, there's certainly opportunities for for further automations and and and the more coupled coupled system. So with that, let me actually start with that with that qualifier sorry, that discovery training And I will just share my share my screen and jump right into it. Okay. So we've got If I wanted to share my entire screen, let me do that here. It's always a bit challenging starting starting sharing screens. Here we go. So here I am with with my Copilot, good friend here, and you can see here we've got all my agents that I've created, and I will have my discovery agent discovery simulator prep. That's an agent that I've that I've built up. And I just wanted to show you where can I edit this this guy here? Won't be able to edit it, but here you go. So so if I just put in, let's get started, and I do that, the discovery simulator will just start ramping up, and it tells me, I can select an operation scenario, field service scenario, field service scenario, or project confusion professional services scenario. And I can just give it as do number number three. Let me do another layout just to make sure that you can see this right. There we go. So I am now speaking to the CFO of an engineering consultancy, and the role play starts now. Go ahead. So I can just type in and I can actually start using voice mode as well if I wanted to, but you wouldn't be able to hear the responses. So what I'll just do is, hey. Please, can you tell me your challenges are. Right. So let's let's get started. And then and then as it responds, stepped into the CFO role challenges. There's so many things. They don't have a single source of truth for the project. Billing is always late forecasting is guesswork. It's probably too high level. I'd like to understand how you define success. What are your main KPIs? Obviously, I'm just want to want to make this fairly quick. You know, I could go into acknowledging their pain, so on and so forth, and that would actually give me a better score. But for now, let's build this out. I've got the KPIs, billing cycle time, utilization rate. Look at our dashboard. Right. So to get a dashboard, you will actually need a good source of truth. Have you got all your customers in a single system? Let me ask a few questions. I can just go through go through this. Not really. CRM is in hotspot. Excel is right. So you have three separate systems, every that be using their own. How about we consolidate and you'll get everyone working up the same data. Right? So all this is done on a custom agent that I preconfigured. And, actually, this is a fairly single sim simple prompt. But as you can tell that it's actually simulating a pretty realistic CFO conversation, CFO scenario. It is a big project, big change management. Yes. Sure. But we can start small with one that and then gradually roll it out. Obviously, now I'm in sort of the objections piece and and cater to requirements as we go. How does that sound? K. As as I would be going through the different sales frameworks, you would be able to see what works best. Let's do well, we're starting small. So we want to make sure you've got the right foundations. So no rework is needed. You'll also want to reap the benefits right way. Understandably so. And what do you let's let's get her involved. What do you think is the best place to get started? Alright. So I'm going through this. And as I'm as I'm thinking as I'm going through, I will round off the conversation. Just wanna make sure that I have enough data to actually start creating a enough material. I'd like to use this conversation to actually generate my proposal, which I will show you how how we are doing that proposal. If you just start a project on the finance side, what's the smallest meaningful step you'd begin with? Alright. I'm getting some challenging questions back. Alright. Let's say just I want want to start with finance. That's the well, let's do billing. That's the core. That's the core, and it's tied to a measurable KPI days to bill. Cool. Let's do that. And it's a chain. Let's start daisy chaining up the pieces and automate systemize one by one. Move forward. Cool. So I should be good at handling typos. So I've got that and not supposed to like the idea, but starting small feels safe. Let's do that. So we are now talking about specifics. So let's let's let me just actually show you a little trick that not many very many people use about Copilot. So you see here's my whole conversation. So what I'd like to get to with this is I wanna put this into a Word document that I can actually use and and recycle. So if I just click this edit in pages, it will start adding things in here, and I can just start putting together the combo. I can do this as a or I can just let's see if I can just highlight everything, Because the the clever clever section is I can use this to yep. I can do that. That's also easy. Boom. Cool. We did that. That's a bit ugly. But the idea here is once I have this conversation, I can just copy and paste this into the evaluator, and it will actually give me a good idea of how things have been and how I performed during my conversation. I can actually use this as a genuine trainer and and coach sales coach for for sales staff. You know, it might not handle very complicated scenarios. But once you turn on voice mode, which I didn't do now, it actually becomes really, really a useful useful thing. So oh, man. It's a bit annoying that it copied that all those things over. So once I get those deleted, I can just get the full convo sorted. Cool. But what I'd like to do here is now in my CRM, which you can see there, I will create an opportunity for my my client here. And this is where I talked about getting the right data and getting the right process in place as and when it loads. There you go. I will now create a new opportunity. And this is where it's really important to have a proper sales process identified and created. So on my new opportunity, I've got what we call the business process flow, the qualified proposed contract and closed stage. So this is let's call it our project billing for CFO Jane would be my new contact. You know what? Let me just select the existing one. And cool. They are have an immediate time frame to buy, and I can just put in the bare minimum details. And what I want to get to is actually start creating a quote on the back of this. Okay. So here, I've got on the timeline, I can check that I just had an appointment, or I can just add it as a note. That will be a lot easier now. First meeting chat. And there we go. I can just copy and paste in here. And what I'll be getting right away is everybody will be able to follow-up on what or see what we've covered. So this is one of the things that Copilot can do automatically if you have a Teams meeting and Copilot sales would just automatically bring in the sales conversation, which is obviously not not what I have now. And as a next step so you are still gathering all the data in your CRM system, meeting transcripts, different quote quotations, contact, so on and so forth. And where it starts getting clever is is the actual agent. I'd like to show you how you can actually build an agent. This will be the most simplistic agent that you can build, and I will call that my proposal creator and use this agent to create proposals based on Teams meeting transcripts. So what I'll what I'll do here is we have a proposal prompt that I'll just copy and paste in. Word sales word proposal prompt. And this is the actual actual prompt that I'm using to create my own proposals as well. So you can see here, this prompt was used or generated with AI, and and it it's got it's in markdown, so the syntax is is markdown. But if I just give you a good idea of of how what it looks like, maybe in a better readable format, you'll be able to see. This is the prompt. It's it's quite a lengthy prompt here, but you can see that we we've got different proposal types for our ERP, agent tech, CRM, data, and BI projects. And what you'll see is we've got the output format as a where a document is defining the section headers in a in a pretty detailed way. And we have an executive summary, so everything is all prescribed. And we've got document structure, cover page, so on and so forth. I've got a table of contents. So this is this is what we'll be providing, the executive summary, define, the number of verbs, the structure, what it should look like. So I'm not actually asking AI to figure out the whole proposal for me. What I'm actually doing is I've created a really detailed template, a recurring piece, but there are custom elements in there. So understanding your situation. Right? That's the part that I want to want to be using that really speaks to the customer and is very much tailored. And whereas here we have proposed approach. This is how we're we're doing initiative A, B, C. These are the challenges. So how our initiatives would be approaching it. And and we've got the initiatives all broken down. So if you are proposing multiple initiatives, that would all be handled there. And if we have discovery workshops, implementation and build, discovery assessment. These are sort of things that we are quoting. They are all described why we're doing those. So I've actually created this on the back of our central knowledge base that we are using to feed all of our AI content. And I just asked AI to give me a proposal prompt. These are our previous proposals. Can you just generate me a prompt that would give me this output? So once it did, I just copy and pasted the instruction in here. And I can actually provide external sites, only use specified sources, reference org charts, so I don't need actually any of that now, I want the capability to create documents for me. Right? So that would be a clever one to be able to create a Word document as in run. Cool. Now it's set up. And I want to save my agent. Is this too long? No. This is actually too long. So let me cut some of this off. The investment side, we won't need in this one. Nope. And this won't be a ninety day roadmap. So we are actually using different AIs and providers. And I'll be completely honest, for the proposals, are using Claude's Opus model, which is which is probably the most advanced model right now. It gives us the best results, but we're constantly testing CoPilot. And since Claude has arrived in CoPilot, it's it's really become a lot more stable and and more more reliable for for this. Cool. So let me yeah. Let me just delete some of this stuff so we make sure that we are sticking in here. There's the tones. So, obviously, it will come up with the stones. There are things that I want to use or not use. We don't like to use things like game changers and all that, so we put that in the in the prompt. And we've got a next steps section. AI review. This that's not what we're doing here. Cool. So we've got we're almost down. So I just wanna make sure that this will get some reasonable outcome of this. And now I can actually hit the create button. So that's all well and great. So let me do another item here where I can just pull in the conversation kickoff and get this convo all copied and pasted in here into into my agent. So once I've created my agent, right, this is this is really easy. This is how how quickly you can create an agent. You go away, and I wanna see my proposal creator agent right here. And please create a proposal for let's call this company NovoTech. That's become our usual go to demo name for CFO Jane based on this meeting transcript. And guess what? Because I'm in the context of m three six five, I can actually just use this conversation kickoff and introduction, and I can just hit. Boom. So it will bring in the meeting transcript, or I can just copy that over as well. So it's pretty clever. It's going away thinking about it now. And fingers crossed, it will work. Let's let's see what it does for us. So just going back into into what we were as it's working working away, this is the thing with real life demos. Right? It's always challenging. But what we saw here is the system side, the process, and the data. So once we have the right data captured and by this, I also in also mean oops. I also mean the the right data in as much of capturing the meeting transcripts and capturing your standard formula, it will create the right stuff. And you need the the CRM where you can put in the meetings. You you can track the process. And it's actually not looking too shabby here. So it just keeps on generating slowly but surely. What it does give us, NovoTech's finance and project teams are operating without a single source of truth for projects. That's exactly what we were covering. Right? So this is essentially a hard coded piece of text about Visual Labs, about me. I usually delete this section, to be honest. It's it's feels a bit boastful. Understand your situation. Right? So it's picking up directly from the the conversation with the client, and it's identifying business impact. This is quite often it makes up some of this stuff, so this is where I need to go in. And and if it's not being quantified what the business impact is, I will need to just make sure that it's not coming up with dummy numbers. And here's our proposed approach. So it brought in a pretty reasonable proposed approach because our defined standard methodology is inscribed in the prompt. So we've got a discovery session and the billing value stream assessment, billing chain pilot implementation. That's what we discussed in the script, and then we'll create that dashboard. So it came up with dummy durations. That's, again, something I would I would work up. And they've got initiative one, discovery and billing, expected outcomes, initiative two, the pipe shop pilot project, the wide deliverables, the steps, the activities. So you can see it's not it's not too shabby at all. So now where I want to get to as soon as it creates it is I want to put this into international email and and be able to send it directly from my CRM. So I can create a quote from here and and send it, or I can just keep it fairly simple for the demo demos' sake and click on an email. And what I'd what I'd like to do what I'd like to get to is start actually sending out emails directly from our CRM because it's got some of these clever things of of pulling in recipient activity will be followed. I can automatically schedule a follow-up reminder. I can set a reminder if if I do not receive an in reply or if the email is not opened, or I usually put in a reminder anyways. And I would just put in a follow-up on NovaTech. So we've got that, and I can even have a template, so I don't need to write it all up. And or I can just write the email and then convert it into a template based on that. So, yeah, that's not what I wanted. Anyways and I can insert a signature, which is not configured in this environment, but I can I actually I'm actually using the exact same signature here? Once I've got the quote generated by my agent, hope it was. Yes, it was indeed. I can actually add to a page as you saw it here. You notice that it's pretty much in a ready to send format. I want to make sure that I delete that last piece. I can create, let's say, even a PDF. So I don't usually create a PDF directly from this. I always create a Word document, put our logo, put it in our template, make sure it's formatted with Visualab screen. But I can save it as PDF, and that's what I would get. Or I can just create a Word document, which allows me to format this right away. Unfortunately, what it did not do is didn't generate the Word document right away, even though I checked it. Here it is. It doesn't read too much like AI, but it looks like AI. This is where I've spent actually most of my time to ping it over, copy it over into something that looks like our format. It's got our logo. It's got our our summary, so on and so forth. But what I can do here is get get this, you know, just download it, download a copy. There we go. And after after all the formatting, I've got it saved. And I jump into my CRM, add attachments, Here we go. Please create a proposal. I just do that very quickly, and it does send the attachment. I wouldn't send it to my Anna, our colleague here, but I can write up the email, save and close, and actually turn the quotation into into proposed stage once I have the account. So that's that's the sort of the power of of an integrated systems. You could see that I was still running back and forth. But what really happened in the past less than half an hour is I simulated a training with a role play agent. I grabbed the copy of the role play agent. I actually created a proposal generator agent from scratch just using a pre created prompt, and I used the prompt and the role play transcript to create a proper proposal, which I then saved as a Word document. I could have edited, formatted, modified it, And I uploaded that into our CRM and sends that data out of the CRM, and I can just add it as a quote as well to get it get it signed off and and be moved on. So you can see that and that CRM is now tracking the that activity against me in the sales accelerator if it did, and that would allow me to track all the items that I'm that I need to be working on. So this allows us to to very quickly train the people, create proposals, make sure that those proposals are in line with with our standards, but also tailored and custom made for every customer of ours. Extracting our CRM so we can start reporting on it and ensure that we follow processes and data standards. So this is how AI can really impact sales without asking salespeople to do all sorts of clever AI stuff. If you think about it, we are making their lives a whole lot easier by not making them manually write up proposals or quotations. And there's as I said, there's still heaps of automation opportunities in this flow that you saw. So we could, in theory, as soon as I finished a call that was transcribed, I can get it recorded, transcribed automatically through AI, and run it through that prompt that I just used manually and have that first draft of proposal landed in a SharePoint folder for someone to review, pick up, modify, and send. All this backend can be automated. And this is where we actually say good AI automations and and good AI doesn't need an option because it just works and it just makes everyone's lives a lot easier. So this is this is what I wanted to show you today and just really, really hone in on the whole Readiness piece of in order to make AI work efficiently, what you will need is standardized processes, standardized quotations, systems, your ICP, so you know what you can offer. And once you've got those standardized, you can very easily build all the relevant automations, whether through AI or not, automations on top of those standards. And that will be your cherry on top, but that's where your real benefits will will really lie. So this is what I wanted to show you today. Next week, we'll talk about customer service. And if you think that this was useful for you, just do reach out to me. We are holding a full day AI readiness workshop for, for executives and decision makers, to really hone in on your specific problems where we identify what are the big value levers, what are the use cases that are worth automating through AI. And we can just make sure that in one day, you'll get a full blown AI roadmap that you can execute on ninety days. If you'd like, we can help you with the execution as well. So there isn't actually a q and a session at the end of these webinars. What we find is the questions are very specific, and I want to understand more of the details on the background. So if you have any questions, feel free to reach out and and we'll be happy to schedule a session for with you. That was it. All the best. Thank you for joining us.