How AI Helps Sales Teams Keep CRM Notes Updated

Sales teams lose critical deal context in the moments between meetings and CRM updates. In this episode, Eric Housh and Zack Terry break down how AI and voice tools help capture that information in real time and improve data accuracy.

They explore the “windshield gap,” the impact of CRM friction on adoption, and how a simple Agentforce solution can streamline data entry, improve reporting, and support better forecasting.

Prefer to watch? Find the full episode on our website: https://loom.ly/eJrCeH0

Transcript

SPEAKER_01 0:09

AI in Action is brought to you by Fast Slow Motion. Our team helps growing businesses put AI to work with practical, scalable solutions. To learn more about how we can help you implement AI in your business, visit FastSlowMotion.comslash AI. Hello, and welcome back to AI in Action. I'm Eric Hush, joined as always by my good friend Zach Terry, director of AI at Fast Slow Motion. Zach, great to have you today, man. Always good to be here, Eric. Zach, let's let's start off with a scene that every sales leader recognizes. Rep walks out of a high-stakes meeting. They got everything up here in their head, they're fired up, but but they've got another appointment they're rushing off to, right? They get in the car, they look at their notepad, they look at their laptop, they look at the ignition, they look at their watch, the watch and the ignition usually wins.

SPEAKER_02 0:59

You know, it's it's always, I think every sales rep's favorite thing is to log information in Salesforce as soon as they're done with the meeting, right? They're just like anticipating it. They can't wait to do it. Yeah. No, everyone hates this, right? Like nobody wants to have to spend the time to do that. And you're right, it's all kind of in your head, right? Like you finish the meeting and you got to remember what you talked about and how was the deal going. And you gotta open up your phone and you gotta log those notes, right? Well, I think what typically happens is most of the time those notes just don't get logged, right? Or you're logging them way later on. So, you know, we were talking to some of our AI assistants when we were preparing for this meeting, and it called it the windshield gap, which I I kind of like that term. It's like, you know, you get right back into the car and well, you got to go into your next meeting, and all of a sudden, all that good context that you have is probably going out the window, right? As soon as you drive off, it's getting left behind. So it's this dead space where all these valuable details of a deal just kind of go to die, right? And so for today, this particular industrial distributor that we worked with, the tool that we implemented for them was all about closing that gap because it was costing them a fortune and lost context.

SPEAKER_01 2:10

Yeah, they're not just losing sales notes, right? They're losing their ability to report back to the manufacturers they represent. It's it's a lot more than just stuff uh that's useful for closing the deal. It's it goes beyond that to other parts of the business. So today that's where we want to put the microscope. We want to talk about the tools, like when once you get in there, once you get that that uh windshield time, how can you turn that into a high-powered data entry engine? So uh this this customer, this particular customer, they they had a solid foundation on Salesforce, right? The system, uh I think they were nine years on Salesforce by the time that that we met them. Uh, but over time, as CRM will tend to do, uh it had become a little bit burdensome. So for the person on the ground, the salespeople involved with interacting with the CRM, what did that look like?

SPEAKER_02 3:01

Yeah, it's all about that friction, right? When you when you talk about coming out of a meeting, you know, if you're outside sales and you're having to log this information into the CRM, that is just it's friction in your day, right? And so the the more burdensome that that is, the more it interrupts your flow and your process, the less likely you are to log that. And so what ends up happening is over time, and yes, you're exactly right, nine years, we'll just call it a decade on Salesforce. They've accumulated so much visual noise. It's it's tech debt. We talk about this a lot just in CRM implementations. The longer you're on it, you're implementing new features, and certain other features kind of get left behind, but people still use it. So you're getting cluttered page layouts, you're getting legacy fields that have been there, but no one's really populating it. Maybe they don't even know what the purpose of those fields are anymore. You've got architectural debt, you've got things like old process builders that you need to upgrade to flow. Maybe you have duplicate contacts, maybe maybe you even have a duplicate deal that you've been working with the exact same account because you just lost track of the one you already created. And all of that, when you talk about rolling that up to the executive level and reporting on this data, well, that that leads to a lack of trust in that data. And so, really, what we were trying to think through here is how do we reduce that friction? How do we keep people from getting to the end of the week and saying, oh, I better go log all of my meetings because I didn't do it throughout the week, right? How can we actually create a solution that makes it much easier for reps to do that without them having to feel like they're wasting time sitting in the car going from one meeting to the next or feeling like they're behind by the time they get to the end of the week and rushing to log all that information so they get credit?

SPEAKER_01 4:37

Yeah, by the end of the week, right? The way that things went down may not be the way they remember them. So you've got this uh sort of a game of telephone there with the data and you can't really trust it. So so to fix this, we thought beyond the keyboard, right? We didn't say, let's go in and clean up all this architectural mess, let's let's change this. We said let's move this interaction with your sales notes from the keyboard to voice using agent force. So walk me through the mechanics of that. Exactly.

SPEAKER_02 5:08

Well, one thing I'll mention before we kind of get into this solution is this is rapidly changing, right? Salesforce is releasing new capabilities all the time. They released a voice capability a few months ago that works with uh call telephony integration type of tools where uh you're actually dialing out of your CRM and it's capturing that information. They've already kind of released something like that, and they are on track to release, I believe, this month, and this is April 2026, a native voice capability that is more akin to the voice mode on Chat GPT. And so we've probably all been in a situation where maybe we're in the car and we want to, you know, turn on ChatGPT, but we can't look at our phone, right? Most of us are in hands-free states, we can't touch our phone. But if we have a voice assistant that can work with us, like Siri, we can have those conversations back and forth. So that's the capability that Salesforce is bringing. Now, in the interim, we didn't have this sort of native capability. Now, there's a lot of tools out there that can do that, but we wanted to build something standardized on Salesforce that was quick to implement to bring ROI in value really quickly. And so we had a little bit of a compromise here, which is instead of us having that back and forth voice interaction, we can use some of the standard dictation features that almost all of us have available on our mobile devices, on our laptops to capture our voice. And ultimately all it's doing is it's just taking the words that we speak and transcribing it into text. And so what we were able to do was build a Salesforce employee agent. Employee agent just means that it is accessible to internal Salesforce users. So you have a Salesforce license, you log in, and you can see the Agent Force icon open up a chat and begin having that conversation. Well, usually that means hands-on keyboard, right? Like you're typing something back and forth. But what were to happen if you were to use dictation to capture that and then send it? It's not anything revolutionary in that regard because it's it's the device that you have that has the capability to capture the voice. And then you're just sending it as if you had typed it, but it's cutting out a lot of the um time it would take to actually sit there and type it. And if you're on a mobile device, you know, obviously it's it's a lot harder to sit there and type on a mobile device, right? You got typos, you got things that are going to get lost, even if you have the luxury of time to sit there and record those notes. It's just a little bit easier if you have something that you can talk to. So this solution was built with the eventual evolution of the native Salesforce Force capability coming to fruition. But in the meantime, what their employees are able to do is not only access this employee agent on their computer, but it works natively on the Salesforce mobile app as well. So you can open up an Agent Force agent from the Salesforce mobile app. And then instead of sitting there and typing, you just enable dictation, you describe what it is you're trying to do. Maybe you're just having a stream of consciousness about the meeting that you just had and you're asking it to log a note in the CRM related to that deal or related to that account or the person that you were meeting with at that company. And then it's able to take all that context and you just press the send button, and then it works just like a regular standard Agent Force agent would work. It evaluates that text, it determines and reasons in terms of what it should do. And we created just a couple of actions for that agent to be able to handle those requests. And so, specifically for this customer, they wanted the ability to create and update opportunities. So the whole deal uh cycle that they would have when they're actually tracking pipeline. And then they wanted the ability to log tasks and activities as well. So this agent has both of those capabilities. Depending on the context of what you provide, it is able to understand if you're trying to update an opportunity. Maybe you're updating some fields or you're creating a net new opportunity because you have a new deal that materialized during that meeting, or you're just simply logging those activities. It's able to essentially interpret what you're asking and then go and call the correct action or series of actions to then go in and create those records or update those records. So, what's really great about that is from a technical perspective, not terribly difficult to implement. It's really no different than designing a standard agent that you would actually interact with using a keyboard. The difference is you're just using a different way to capture the information that you're gonna send using your on-device dictation. And again, long term, this will morph into a voice native capability. And so it will be even smoother. It'll have the ability to actually talk back to you so that you're not having to read it or look at it. The utility right now is beneficial probably as soon as you get into the car, but obviously you're not gonna wanna be interacting with it as you're driving to the next destination, right? So there's a little bit of a utility gap there. But as these features start to get more powerful, you can imagine getting in the car and then having that conversation just like you were speaking to someone on the phone, hands-free, of course, maybe using your car's Bluetooth, right? And imagine that agent being able to talk back to you and being able to take those actions for you. That's kind of where it's going. That's the vision. That's where we want to be able to take this when those capabilities are finally there, natively on platform. Obviously, you could you could bolt together other solutions today with other technologies and platforms. But if you want all that to be on platform in Salesforce, that's kind of where your central business is. That capability is is very near in the future. It's coming very soon. And so I think this will add to the utility that it already has. It's already speeding up that process for this customer. But then when that sort of capability comes out, it's going to be even more natural and enable them to um, I think be even more productive over time.

SPEAKER_01 10:37

Zach, you know what this reminds me of? Uh I've I was in my undergraduate, I was pursuing a pre-med uh line of uh of study. And I worked with a lot of doctors, and this is how they operate, right? They come out of the room with a patient, they've got their trend their recording device, they they record their stream of consciousness and what they think about the patient. Now, back in those days, where I'm dating myself, that went off to some old lady that typed it all up. Uh, but now with the technology, we can make that a lot more seamless. The dictation device is in our hand, we carry it with us everywhere. Now, uh with this solution, we're plugging it right into the CRM so that we're able to not only record uh those really relevant, really recent notes, but also because it's in the system, we're able to check the work against the business rules in real time. Just kind of backing up for a minute, when you lower that barrier to entry, when you're not asking them to pull out a notebook or pull out an fat finger on their mobile device, what happens to data quality? I think it's two things.

SPEAKER_02 11:38

One is just the velocity of what you're able to capture, it's lowering the friction for what it takes to actually record quality notes. And so by lowering the friction, it makes it easier to get those notes. And so I think you get you get a little bit better adherence and you're not waiting until the end of the week to go back and try to remember what was discussed and plug it in. It just makes it easier to do that in real time. That's I think kind of the main benefit. And the benefit of that velocity is not just I can do it faster, I can do it easier, but because I can do it faster and because it's less friction, you're more likely to get more activities logged at the right time, which means the quality goes up as well because it's gonna be the recency effect. If it's on your mind, you're gonna have better results than if you try to do it at the end of the week. Because I think we've all done this, right? Like we've all done whether you're in sales or not, if you've ever tried to go back and remember something that you did during the week, if you wait three days, four days after the fact, it's not gonna be nearly as good as if you do it in the moment. So I think that's really what this is. It's lowering that barrier to entry right there at the point of entry so that you're able to do it more quickly and thus be more likely to do this when the information is immediately at hand. So practically, what that means for this solution is it's no typing, right? The whole idea here is you don't have to sit there in your car before you leave and type it up. You can just talk, let it record what you're doing, and then just send it when you're ready. And then the benefit is the agent has the ability to perform the action that you would otherwise be performing. And so instead of going in and navigating to the account and the opportunity inside of my app, and then from there having to find the right button to press to log an activity and then going through each field and typing it in. Well, you just say what you want. You can put it all into one stream of consciousness, and that AI is able to kind of parse out the key pieces of information that it needs to log, and it's doing that on your behalf. And so you can see how that could significantly lower the amount of time that it would take, as well as the, I think, perceived effort. I think there's a little bit of cognitive dissonance that goes on when you're having to navigate multiple screens on an app and type things into multiple fields. Well, when you create an agent that is prepared to listen for that data and then act on that data, it really takes a lot of that manual effort away. And now you're just focused on providing the information. And then it takes the information and turns it into something that is actionable, it's recorded, it lives in your system of record. And so instead of having to do that all by hand, it can just handle that for you. And so then that also leads to better forecasting accuracy. So we we talked about how if you're an executive and you you have a whole sales team that you're managing and you're relying on this data to understand how to forecast the next quarter or the next year, well, if that data is delayed because people are not entering it on time, then your ability to forecast is delayed. And so if we're able to encourage people to enter in this information and make it seamless and easy, and it doesn't cause a delay in their day-to-day, then I think that naturally leads to better reporting and better accuracy.

SPEAKER_01 14:50

Yeah. Faster, easier, more accurate. Uh, Zach, from a strategic perspective, for the CEOs, for the sales leaders out there listening to this, what's the takeaway lesson for them?

SPEAKER_02 15:01

In this case, I'd say that the AI tooling that we're using is just a layer in the entire process. It's not the foundation. And so before we even got to the point where we had an AI that the employees could access, we spent the first month of this project auditing the org. It's nine years old, right? So we have to go in and clean up things like duplicate data, migrating some of the legacy tech that they had to more modern infrastructure and making sure that they had a solid foundation. And then once we had that foundation, then the AI was able to perform a little bit better. Um, now I think for this particular use case, it's not solely reliant on having a perfect data foundation. And this is a great example of how you can get started with AI without having everything absolutely perfect in your org. We just had to make sure that you had the right accounts in the system, that you understand what's required for a given opportunity or a deal flow. And the example that I'll use there is if I'm just saying, hey, like I need to put this amount into this opportunity and it's this customer, that's like a little bit of data, but it's not a lot of great data. And so for them, it was all about spending the time before we actually tried to implement this, understanding what the process is. Different types of opportunities have different requirements that have to be gathered. And so we have to make sure that the agent understands if it doesn't know what type of opportunity is being created, it needs to prompt the user to actually provide that so that it knows which pieces of information it needs to gather in order to create that opportunity. And so it's really having the result in mind before you build it, because I know the result that I want is an opportunity with good data that helps me forecast. Well, if I don't know what pieces of data that I need in order to help me forecast, I'm gonna get some variable results. And so in this case, we laid that foundation less in terms of the data in the system and more in terms of making sure that when we're capturing that data using AI, that it follows their process so that the end result is better reporting by the time they're finished logging those activities.

SPEAKER_01 17:08

So let's make it practical for those same leaders that are that where this is resonating. They know that their reps may be sitting in cars right now, avoiding updating the CRM. Where should those leaders start?

SPEAKER_02 17:19

Well, I would say go and talk to your employees and see what their pain points are, but you already know this, right? Like if you're in a sales organization, if you're doing any kind of outside sales, you already know that this is a pain point. But what I would do is just sit with your employees, maybe do a ride-along, or at least just interview them about how they're doing this throughout the day. When are they actually logging these notes? Look at if they're logging them when the meeting is over, or if it's taking them a few days to actually get it in the system. And that's not to go in and audit what they're doing because your salespeople, I mean, we all know this, everyone does sales a little bit differently, and you kind of have to respect their process. So it's not about making them adhere to some specific way that you're doing things, but it is making sure that you are capturing the data that you need in order to actually measure what's going on in the organization. And so with that in mind, I would say just interview your sales team, spend some time with them to figure out where the biggest gaps are. When are they typically logging this information? Are they logging it? If they're not, why aren't they logging it? Are they constrained by time? And thinking through some of those things can help you identify if putting something like an AI assistant in place is going to be helpful. Um, and so the other thing that I would say too is this is not really specific to AI. This is more just legacy CRMs having a system of record for a long time. Think through everything you're capturing and determine if you actually need to capture it. And so this is not just about capturing the right data, but also removing the barrier for data that isn't important, right? So if you go through and look at your opportunities or your deals, figure out if you have any fields that haven't been populated in a year or two years or whatever that looks like, right? And if you're not actually capturing information there, maybe you don't need that field at all. Or if it is something that you need, but it's not getting captured, why isn't it getting captured? Is there some kind of friction or pain point that's preventing that from happening? That can help you determine if you can streamline it with something like an AI tool or maybe it's just an automation. Or uh maybe your reps don't have the information that they need at that point in time, in which case that's an opportunity to reevaluate the process because if it's critical data, but the rep doesn't have it at the time that they need it. Maybe we need to reevaluate that process to make sure that uh they actually have the information that they need so that it can be reported on later.

SPEAKER_01 19:39

Yeah, bottom line, it's about making the tool serve the talent, not the other way around. Um, if you're hearing this, you want to see the blueprint for this kind of lower friction setup and how AI can help, check out fastlowmotion.com. We'll have the story there. And tune in over the next few weeks. We're gonna be telling real stories of AI applications in the real world, just like this one. If you tuned in to the last episode, you heard another one. Uh, we definitely want to hear your stories as well. So reach out to us if you've got some novel applications for AI in your business. Uh, Zach, final word.

SPEAKER_02 20:11

We said it before, we're just saying it a little bit differently here. I said the word friction, same thing as a pain point, right? So it's starting with that problem, identifying what the problem is, and then thinking through those solutions. So uh using this as an example, it's not just about getting more data, it's about getting good data, getting better data, and using tools to help you reduce that friction so that the data that you're capturing is timely and it's accurate.

SPEAKER_01 20:37

Well said, my friend. Thanks everyone for listening. We'll catch you guys next time.