AI Implementation in Sales & Product Teams | Alexis d'Eudeville (Lemlist)

Most companies talk about AI.

Very few actually implement it in production.

In this episode of Business AI Explained, Vladimir de Ziegler sits down with Alexis d'Eudeville, AI Product Manager at Lemlist, to discuss how AI is being used inside real companies.

Alexis shares practical lessons from building AI products, launching startups, and working at Google.

They explore how AI is transforming product management, sales automation, and go-to-market strategy and why the most important factor is still keeping humans in the loop.

The conversation covers:

• How AI is implemented inside modern sales teams
• The role of generative AI in product management
• Why data quality matters for AI adoption
• How companies can move from experimentation to AI in production
• Practical examples of AI improving marketing and customer success

If you're a founder, operator, or builder trying to understand how AI is actually used inside businesses, this episode breaks down the real strategies behind AI implementation.

Key Topics:

AI implementation in business
AI in product management
Sales automation and RevOps
Go-to-market strategy
AI in production systems

Chapters:

00:00 Introduction – Alexis d'Eudeville & AI at Lemlist
01:45 AI in Business: Why Implementation Matters
05:12 How Companies Are Using AI in Production
08:30 Generative AI for Content and Data Analysis
12:45 Human-in-the-Loop AI and Ethical Considerations
18:20 The Future of AI Tools in Business Workflows
22:15 Real AI Examples in Marketing and Customer Success
27:40 AI Adoption Challenges for Startups and SMEs
31:05 How Companies Can Successfully Implement AI
35:50 Key Takeaways: AI Impact on Business Growth
38:15 Final Thoughts on the Future of AI

Episode length: ~40 minutes

👤 ABOUT THE GUEST

Alexis d’Eudeville

AI Product Manager at Lemlist

→ LinkedIn: https://www.linkedin.com/in/alexis-d-eudeville-348bb858/

→ Company: https://www.lemlist.com/

🔗 WORK WITH VLAD

If you’re implementing AI in your operations and want hands-on help building real workflows:

👉 https://www.elementsagents.com/

🔔 SUBSCRIBE

→ Linkedin: https://www.linkedin.com/in/vladeziegler/

→ AI with Vlad: https://www.youtube.com/@aiwithvlad

Watch the full video version on YouTube: https://www.youtube.com/@aiwithvlad

New episodes every Tuesday.

Transcript

SPEAKER_00 0:00

In today's episode, I'm joined by Alexis Dudville. Alexis is an AI product manager at LEMLIST. LEMLIST is a tool that allows you to automate prospecting across multiple channels. In today's episode, there are three topics that I would like to unpack with Alexis. The first one is to understand a little bit better how he's using AI in his job. So as a product manager. Today the narrative is that it got so easy to build, but to build something that people want is actually even harder. So I want to understand how Alexei is figuring out how to use AI in its process to build features that people actually are willing to pay for. The second topic that I would like to unpack is around AI and sales automation. If you're listening to this podcast, it's very lucky that you've received some call DMs already on LinkedIn and email. I am a LEMless customer myself, so it's quite interesting to also go over how to best use these tools, how can we actually stand out with all that noise? And then finally, because Alexis is spending a lot of time working with go-to-market teams who are trying to implement automation in their go-to-markets. I'd like to get us his post on what are the most successful go-to-markets nowadays. Are people using tools like LEM list? And what are the ideal team structures for like successful go-to-markets in all types of companies? Let's get started. Amazing. So thanks so much, Alexis, for being here and joining the podcast. I'm super excited to have you here because I've been following your journey as a PM and also your take on sales on LinkedIn and other and other platforms. And the whole point of this podcast is to cover how to build and how to sell. So I'm very excited to have you here to cover all of that. So yeah, thanks. Thanks for being here again.

SPEAKER_01 1:30

Thanks for having me, Vlad. Happy to be to be here and to share with the community.

SPEAKER_00 1:35

Just to set the stage a little bit, I think 2026, we've all been overwhelmed and we've all been drowning in like new tools. People are getting crazy about cloud code, open claw, uh, and there's so much going on. So first of all, like what's your what's your overall feeling today with uh all these AI releases that we've come across, the new models? And uh yeah, what's what's your thoughts on that?

SPEAKER_01 1:58

Yeah, I mean, when I look at my job as a PM uh in a company like uh like Lemlist that has that has radically changed in the past two months, even though I have been uh using uh tools to vibe code and to really change the way I work um uh for for quite some time. But uh I really see that as an organization we are things are changing, and this is I would say linked to two things. First, some models have really uh reached a level of of intelligence that is good enough. Uh that's I would say only a part of it. The other one is really that we're seeing more and more harnesses, systems around these models that are making it really useful and cloud code. I would say cloud code, part of it uh if you look at it is is the the model, obviously. But the the biggest part is the is the the harness and the system around uh around the models that are really making it extremely useful. And in practice, uh what does it mean to build a SAS like from now on with these new agentic systems that are doing stuff for us? It's huge questions, uh huge question, and and uh it's also completely changing the way people are working, and especially PMs, I would say. Um so like it's it's amazing on a daily basis what's going on. Um there is a lot of unknown, of course, but a lot of stuff to to cover for sure.

SPEAKER_00 3:38

Coming back to to the role of like an AIPM, uh I remember like 10 years ago, and unfortunately my my background was in finance, and uh we used to say MA and tech, you know, there were like two buzzwords joined together, and we're like all super excited about it without really knowing what it meant. And you know, AIPM is like okay, we're product manager is a is a very interesting and attractive role, AI as well, and it's joined together. Uh so why is there so much hype and why is this role so interesting? Probably like drawing some of the best talents, and how do you think it differs from a traditional PM role?

SPEAKER_01 4:16

Actually, when you when you think about it, I would say that an AI PM doesn't really make any sense because when you think about it from a product manager perspective, um uh like AI is just another technical solution to solve a problem. And if you think about it, all the PMs should have in their toolbox to solve problems AI. Now that's in theory. Um, in practice, uh this is a brand new technology that requires a lot of uh understanding how it works, the types of of uh uh solu of problems that it can solve. I would say this is why in practice you have some AI PM roles that emerge of people who really uh understand how this technology works and can really drive a roadmap of uh of uh solving problems with Genai. Not all problems are are relevant to solve with uh with Genai, uh, but uh in practice the capabilities of this technology are really enabling the software industry to solve a lot of new problems and to solve existing problems differently. That's really the case uh like when you look at uh a software like LEM list, for example, and maybe we'll talk about that later, but um but uh but uh yeah, so and and I would say, and this is uh kind of weird, but uh I often um feel that that normally when you are a PM you are often told not to look at the the solution but to look at the problem. And I feel that is still the case with uh with uh with uh AI currently, but it's so revolution revolutionary in the capacities uh that it's um like important to wear the generic glasses when you look at problems and to really understand deeply what you can do with this technology and how you are successful, you can be successful with this technology. Uh and this second part is um is very important because um when you look at it, the quality of the AI features that you that you create, they don't really rely on the quality of the code that you are actually uh shipping in production. It's really relying on the quality of the output that your AI system is producing. And this is absolutely uh like this is really different from crafting traditional features because these are non-deterministic uh systems uh where you need to bring the right context, so you have a lot of data data questions, um some evals, and so on. So, in order to make sure that that you are getting uh shipping the value to the to the the client. So, all in all, this makes like this makes a role which is quite different um from I would say a traditional feature in the types of feature that are built. Um but like my my view is that uh at some point all the PMs that will remain um will be in some way AI PMs. Um the same way you don't really have mobile PMs or cloud PMs anymore. Um uh and and because this will be part of the the toolbox of uh any product good product manager, I guess.

SPEAKER_00 7:53

Yeah. Um I thank you. It's great because you're touching on all the key dimensions of of how to build with AI today, like evals, context, uh prototyping to also know what's uh what you can actually do with it. Uh so this brings me to maybe uh my follow-up question. How do you actually use AI today to come up with features that you can test out to see whether you know they are behaving like you want them to behave, even though there's a stochastic aspect to it? But can you maybe walk me through an example? I know you have some nice little uh um code.

SPEAKER_01 8:30

Yeah, I think like behind your question, there is the this this question of uh you know how do you use AI in the the daily job of a of a product manager and uh how like this is transforming the way um you know product manager are I mean really in the past two months the job has been two to three months, the the job has been like at LEMLIS profoundly changing. Um going from a traditional PM you know flow from discovery to QA and going uh through specification design uh um uh and and so on, to uh a most of my day spent on cloud code, um uh chatting with uh with uh the the agent uh on several use cases. So I would say like the one that has been really changing my life the most, um and is is also I think one of the most basic and accessible to any PM is uh the the one I call chat with codebase. Uh so now at LEMList, all the the PMs have a dev setup, um exactly the same as if we were uh developers in the in the tech team. Um and we all have access to the entire codebase. And the very first thing that we do that I've been doing a lot is to ask, simply ask questions to the codebase to really deeply understand how you know 10 years of software building and legacy have been you know uh compounding and on specific features, really understand what uh you know the code is doing, not doing, what are the edge cases. Um, and this has been you know huge in terms of uh helping myself as a product manager. Then if I look at uh you know writing specification, for example, uh right now I have a setup uh that I can actually uh share with uh with you, uh, which is totally different in terms of how I write specs. Uh what I do basically after doing some discovery, and uh uh this can be done partly with AI as well, of course. Uh and once I have a clear idea of what I want to build with the with the team, uh what I do is that I open a CLAP uh recorder, meeting recorder, and I talk about what I want to build uh in long details, uh, and I cover all the edge cases and all everything that I want people to be able to do with the feature, not to do with the feature, and so on. And then what I what I do is that I uh take the transcript and I have a uh specific setup with Cloud Code where I connected the the Notion MCP. Uh and what I do basically is I create uh I take a spec template that we have in in Notion and I pass the link and I say, okay, craft this uh uh craft this uh this specification, like generate the specification. And this is really an amazing process because the AI is taking all the transcript, is working on it, thinking about it, uh asking me questions about uh to to clarify some some elements, showing me edge cases that I haven't been thinking about, and at the end of the process, I just uh hit enter and I have a brand new spec that is ready to go uh to for the developers. So this is in a world where you still write specs on Notion, and we can talk about uh uh what's coming in the future, but like this is uh I would say a pretty clear example of uh how AI has been changing my well, uh my my day as a PM. But then uh I would say stuff like Cloud Code or Cursor have have really been opening new areas where I, as a PM not able to code, um uh can bring a lot of value now. Um and I would say the the there are several areas, but the the first area is really about being able to build your own tools basically uh to achieve uh some of your goals. And uh one of the key uh areas um that I find brings a lot of value is uh the ability to test your own integrations and test uh APIs and test tools that you are thinking about integrating in your in your feature. Uh normally you would need a dev to try and uh the stuff and do a prototype. And now what I do basically, and I I think I can show you I can show you this. I go uh I create a new uh folder on my Mac and I basically uh start prompting Cloud to uh do what I want to do. Uh so what I encourage everyone to do is to plug some some plugins and some MCP um that are really relevant and helpful, especially the one that is called Context7, uh, which is actually a project uh that has been uh collecting all the documentation of all the major uh tech projects and turned it into a LLM ready documentation. Uh and so what you do is plugging this to Cloud Code, and as soon as uh you ask something about a technology or a tool, then the the system has access to the to the documentation. And so what I did is that I created this folder, and I have um here in this uh in this case uh built this uh test uh UniPyle and Mpersend app that I've uh relaunched here for for the conversation, which is here. And as you can see, uh this is a very simple app. My goal was to try and test um the capabilities of a of a dev tool to connect uh an app with uh CRM uh in this case HubSpot, uh, and to understand uh the cap what the this technology was capable to do and uh not capable to do, and get my own intuition on uh should we go with this solution or not. And so what I did basically, I had uh my uh I created my own uh you know dummy uh um uh HubSpot account with a few accounts, a few contacts, a few activities, and I basically uh asked Claude Cod to uh to um craft this uh this uh this app uh for me. And in just a few hours I was able to basically pull a list of uh you know sales rep, uh automatically uh uh pull the the accounts associated with the the rep and for all uh you know the accounts getting the contact and the activities uh that I wanted to have so basically covering most of the elements I wanted to have in my use case. And this I was able to do by myself uh in just a few uh you know hours uh and and massive acceleration to what is done like compared to not using AI for sure. Um so that's I would say an example of uh you know crafting your own tool. And I'm not even talking about what I'm starting to do, which is shipping you know actual code in production. Uh and this is fairly new because initially the process was um mostly uh you know as a PM you can vibe code and prototype on the code base, but only on a specific branch on uh on uh GitHub that was never meant to be merge at is. Um but as uh things go and and and as an organization we feel more and more confident. Uh I am I have actually started to ask for uh you know to send peers and assign peers to to some devs. Uh obviously they do not take my my PRs as if it were a PR from um you know another uh another developer because I'm not and I will never be a developer. But um like some of my code is is currently uh merged in in production, so that's a pretty huge change in terms of uh you know what was done before.

SPEAKER_00 17:14

There are so many topics that you touched on. I I think just in general, uh what's very interesting is this underlying topic around um organ uh governance and or like the organization, basically, from what I understand is LEMNIS is really pushing everyone to adopt AI. So you have access to cloud code, you have access to the read uh MCP of GitHub so that you can review the code base, and you have the ability to also push code into production. So you know, unless you have those capabilities in place in the company, it would be impossible to move at the speed of LEMless.

SPEAKER_01 17:46

I think what you're uh talking about is is is really the center of the question. It's uh at the end of the day, it's it's not really a tool question and a tooling problem. It's much more of a people and organization uh questions. And and um like through that uh at the top level of uh LEMLIS at some point, they decided to uh go and uh to create the condition for the teams to be able to do this. It required a lot of work because it's changing a lot of stuff in the way people are working, and obviously, like once you're challenging people's habits and and process and everything, there there is some friction, but then if you have the right level of uh sponsorship, um then you create the the right condition. But without that, it's not really possible to to to to uh to move forward, and and you need to be in some way resilient because something very surprising that we had at LEM list is that um like at some point like they all had cursor and uh all the devs are had cursor and we're using it, I would say mostly for simple code completions and so on. And at some point, like the tech management decided to give uh uh cloud code licenses to all the devs, and what we saw was a drop in productivity. Like that was weird because you would expect people got too excited, so they were not working anymore. Actually, they were, but the thing was that they were spending less time to write the code, but way more time to own it and to re to read everything that was written by by DAI. And it's actually the problem is that it's actually changing the process and the workflow. Um and so if you are not changing the way you you work, uh it's really hard to really integrate a tool like Cloud Code in the in the tech process because you might find yourself really floated with a lot of code and and so on. And so you get you get kind of this uh first moments where uh the productivity decreases before uh starts getting up again. And we have really seen that um uh uh with uh with uh some of the devs uh that we were like depending on the adoption that they have, but at an organization we have really seen that that it was challenging actually to take this. So, really definitely a people problem, a governance problem.

SPEAKER_00 20:29

Interesting. Yeah, I think uh I've heard this a couple of times where basically it takes two weeks to dive into cloud code, play around with it, uh feel a bit overwhelmed by how powerful it is, and and you know it's not super accessible to especially non-technical people, like having a CLI and so on, it's maybe a bit intimidating. Yeah, but once you're past this uh mental barrier, you can really get a feel of how to get the information. With contact seven, you can pretty much integrate with anything. So it's um yeah, you basically have to cross that that.

SPEAKER_01 20:58

Yeah, you you you you have to to to jump that that first two weeks. And I would say in the specific case of uh you know product that has some legacy, existing codes, it's also there is also learning curve in you know really understanding how people are uh you know um how how the the the the agent is behaving in this code base. And actually now we have a a dev with who is now almost full-time working on the cloud code setup, uh so the entire file system um of skills, rules, uh subagents, and so on within Cloud Code to make sure that it's actually um uh helpful and and and really efficient in the LAMList technical environment, which is not a given when you have such a huge code base as we have and and and ten years and a bit less than ten years of of uh software building in in behind us.

SPEAKER_00 21:57

Yeah. Um If we take uh just a quick step back on the role of a PM, if I recall correctly, as a PM, you're basically whenever you evaluate a feature, you look at three things. The first one is reach, like how many customers are you gonna reach, impact, you know, how will they benefit from the tool and how complex it is to build that feature. And from what you shared, basically complexity, you can look at the code base, you can like play around, you get a sense. Impact, uh, you get a feel because you can play around with those integrations and you see how far you can go and have a feeling about how useful the feature is gonna be. But when it comes to reach, do you use AI today to actually fetch testimonials, feedback? It feels like the last part that is not really scalable is gathering feedback from users to actually figure out what to build next. So, what's your take on that? And how can you actually uh um you know figure out faster what people want?

SPEAKER_01 22:47

Yeah, I mean, um AI is definitely extremely helpful there as well. Um, the same way I was uh you know, I explained earlier that uh I'm using a lot my my clap recording uh to really um uh explain a feature and uh to and pass the transcript to to Cloud Code, all my user conversations are recorded, and all these insights, um uh all the this raw customer materials are uh used to um uh to and and process to find some ideas. So coming from uh obviously the users and and the pains that they have with current features, features, uh the pains that they want to solve in their uh daily job and that we could for for which we could create new features. And then you also have, of course, uh looking at the market and the competitors and what they are doing. And like, for example, myself, I'm using Cloud Code a lot to actually go and fetch on a regular basis uh the product pages of some competitors that that I'm monitoring to look at their features and stuff that is changing and see how the market is evolving uh and the new features that are released, uh, which is helping me, of course, to to have a feeling of where the market is going and uh and and have some new ideas of uh things that we could um steal from competitors uh and bring to our uh to our uh to our users. So for sure it's it's bringing a lot of uh uh impact there as well um uh when it comes to the the PM role.

SPEAKER_00 24:32

Okay, nice. Um I I have lots of other questions around that, but maybe if we switch and go to the second section uh of this chat where I would really love to have your take, um, is around how people can use AI in their go-to-markets. Uh so once again, um for those who don't know, LEMLIST is a prospecting tool that allows you to automate like email sequences, LinkedIn DM, so it's omnichannel, also with voice, uh sorry, calls and so on. Um so maybe just first question for you. Um where do you feel like you know the massive the biggest changes are today? Uh you know, it's very high level, but and then we're gonna deep dive uh deeper. But what what you know what's your overall feel today and how it's changed in the last couple of months now with all these tools um available?

SPEAKER_01 25:19

Well, I mean, um if I take my my GI sunglasses and uh the glasses and and look at uh the the capacities of AI and and the and the tools, um I'm like what's really like the the the biggest change in my opinion is is the ability to get into unstructured data. So it's it's a bit uh uh it's a bit uh uh theoretical said like this, but it's really about being able to go deep into uh the content of a field and not being focused only on the on the field. If I take an example, um like the very first step of a GTM motion is is about finding leads. And uh we have, for example, a people and company database uh at LEMList where people can add filters like industry, say, oh okay, I want the software development industry, uh, and so and then you have a list, a very long list of companies there. And actually, so this was great in a pre-Genai world, but when you look at uh the capacities, uh the capabilities of Gen AI, you're like, no, this is not good anymore, good enough anymore. Um, I want to be able to deeply like semantically understand the description of the company that's um you know, of all the companies in this list, to be able to really uh apply my own thinking and reasoning to this process of uh you know finding the right uh um uh the right companies to talk to and contact or company research. Uh you can really go deep into uh really understanding uh uh who you're uh talking to and their pain points and craft the right value proposition. The entire industry and all the solutions that have been brought like in the past 10 years for um all pillars of the the GTM value chain are completely um you know in the process of being revamped right now.

SPEAKER_00 27:30

Yeah, so uh again for those who are not familiar with with signals, basically with signals, we are referring to whenever there's a job change, it's likely that that new employee is going to try to adopt new tools. So if you sell the tool, this is the kind of signal you're interested in, or if there are like hiring trends and so on. Uh so maybe just curious to hear your take on AI is very good at dealing with unstructured data. So, what are the new kind of signals that we can actually capture today with AI that actually people could could start using, you know, that are very new that we couldn't actually extract six months ago?

SPEAKER_01 28:03

Actually, uh when you when you think about it, um any question, and uh that's actually a feature that that we are currently uh uh releasing um uh and and working on uh uh in namelist right now, which is course, which is called custom signals, where you can basically ask uh in like virtually any question about a specific uh uh set of uh specific uh set set of company um and have a the system that will be running and doing on a regular basis the research uh about this question to see if uh you know the answer is uh triggering this custom signal. So it virtually you can you can uh yeah, any information uh you can leverage any information which is available uh on the web to do this. So like you had in the signal industry uh companies and offerings that have been built, like I would say only on uh on uh some signals, like uh, for example, uh the job offering. You have uh entire companies and product offerings that are that have been scrapping and that are specialized at scrapping all the job boards and and processing the data. Um in the near future, uh you might be able to set up your own system where you would want to monitor a specific uh set of uh you know tools or pages and ask questions to uh have an AI being able to capture the data, think about this new data that that is uh available and uh find an answer to the question which is at the center of your signal, basically. Uh so that's uh uh you know uh you could ask uh if uh is is this company uh you know uh uh has this company recently uh uh bought over another company, for example. Uh you don't really need to have uh access to a very specific uh database of uh you know uh that is reference listing all the MA transactions that occurred. You could just go uh and have an AI uh you know do some web search and finding some information. And once it's relevant, then this is a new signal that that is in your uh in your pipeline, basically.

SPEAKER_00 30:38

This really touches on my follow-up question around tool rationalization. I think you mentioned on LinkedIn uh that people will move away from you know having dozens of tools to having maybe one orchestrator tool, uh, either uh Cloud Code or OpenCloud, whatever. How do you think about um product development for LEMlist, you know, and how is it gonna fit into those new go-to-market stacks of of companies?

SPEAKER_01 31:05

Uh I mean that's a that's a that's a very good question. And when you if you take a step back and think about what a SaaS is, it's basically a UI on top of a database for people to be able to perform stuff themselves. Um but now with uh software that are able to do stuff, reason, take action, make propositions, uh then obviously uh the question is uh how do you go from a UI that is uh you know to help humans do stuff to a UI that is maybe helping humans, if you still have a human in the loop, reviewing things uh instead of doing it. So that's only partly answering your your question, but uh that there's a kind of a shift in the way you build features and the way you build your UIs and and and so on. Then you have this massive uh you know wave with the the cloud code and I would say the uh the version for everyone of cloud co-work um uh where you can bring your plugins, it's highly likely, and especially for for you know small one-person software um uh like software for for prosumers, for example, or for funders or or so on, uh, or very small teams, that these will be replaced. I mean, I don't have a crystal ball, but that that's kind of a feeling that they may be replaced by a very good agentic loop with uh a chat and connect it to the right capabilities. And so this is where you know um this is some stuff that we are really thinking uh a lot about um uh at LEMLIS, which is how can we provide the capabilities to an agency system and to any Agen T systems that uh may need uh you know um list of leads, capabilities to enrich uh you know, find emails and so on, uh send you know warm-up um uh uh domains, uh send send uh uh emails and follow-ups, uh trigger calls. So really in some way unbundling all this uh and making it accessible um in the right way, so through uh MCP servers that are well designed and so on. These are really new topics that that are uh daily, uh that we have on a daily basis at Lame list to become basically future-proof. Um it's likely that to get back to your question, it's likely that um you know uh GTM stack will probably shrink uh into uh an orchestrator uh that has capabilities. Um I'm even wondering whether or not it will it will have a UI, a dedicated UI, or it could be if you look at OpenClow right now that you can access in your uh Slack account or WhatsApp or uh Telegram and manage from there. You could definitely have that and have a system, and we see that with the MCP apps, uh the the evolution of uh of MCP to be able to feed UI, basically. You could have uh really the the brain, the core brain of your app, which is available through um uh you know WhatsApp or uh or Slack. And then when you need a UI, because in many cases a UI like a table to review lists or a button to click is more efficient and more effective compared to uh to typing in a chat, uh, then you can have your system like generate this UI on the fly uh and show it, and once it's used, you know, dump it and and and that's it. So um changing a lot of stuff. Um it's hard to know, like and obviously nobody knows how it's gonna end up, but um providing capabilities to agents seems to be uh uh an important uh move right now.

SPEAKER_00 35:38

Yeah. Yeah. I like uh the the notes that no one knows, uh so we can just build and and have interviews with people paying for your subscription and you know hope for the best, kind of as long as they're paying for like a complex product, you're like, okay, we're on the safe path. Um maybe just a couple of just one last question. You've run interviews with those companies, and I know you just touched on like the stack. Uh do you already see some teams that are doing particularly well with like specific tools? And do you observe specific patterns in how the best sales teams today operate and how the you know LEMless potentially fits or you know uh similar tool in the whole stack?

SPEAKER_01 36:19

Yeah, I mean it's true that uh like there is a there is a there are some changes in the way uh uh in the way the best uh sales team uh uh operate and like it's hard to know whether or not they are more success more or less successful, but I think some are more innovative in in many ways. And um when you talk to GTM teams and to sales team, like some of them are obsessed with how do we how can we bring uh you know um AI into uh the day-to-day of uh our salespeople. That is especially the case for for companies obviously that uh that have a heavy sales-led uh motion. Um that's even more important, I would say, when the the sales cycle is is a bit longer. Um because this is where you know it's always hard to know like if you're making progress with your deals and so on. Um so like having this obsession of uh you know building the right systems to help reps with productivity and to help reps make better decisions is uh is some some some of the stuff that I'm seeing around me that that really uh uh I find the the most exciting. And especially the second one, which is how do we help uh reps with uh this core question of uh you know um uh taking the right next best action uh like to actually make progress with their uh with their deal. Because at the end of the day, when you if you if you simplify the equation of a of a sales rep, uh then you have a uh you know uh number of hours of work, uh you multiply that by a number of actions, uh, and then you get a revenue that hopefully needs to be over a quota. Um the number of hours is fairly known. Uh then the question is what are the right actions? Uh and I think like bringing context, uh, all the context that we can master and the reasoning capabilities well orchestrated can do some pretty awesome stuff um uh to help you know better prioritize and take the right action uh efficiently for sales rep.

SPEAKER_00 38:50

Amazing. Yeah, I think it's uh so basically a gentic workflows to prepare all the pretty much all the work to ensure that the sales reps can uh take the next best action.

SPEAKER_01 39:00

Absolutely.

SPEAKER_00 39:02

Nice, perfect. Uh thanks, Alexei uh before we wrap up, do you have anything else you would like to add? Where can people find you? Any final words?

SPEAKER_01 39:11

Um well thank you, uh Vlad, for for the chat. I think it was uh was pretty cool. Uh I'm you know, you can follow me on the on LinkedIn, get in touch on LinkedIn, happy to to share and uh reply to to some of the questions that uh may arise from links listening to that. So feel free to reach out uh and uh uh happy to always happy to chat about uh you know AI, future of uh of SaaS and uh how we can uh you know change our workflows with uh with AI for sure.

SPEAKER_00 39:40

Amazing. Thanks, thanks Alice, thanks Alexis. Uh yeah, please do follow him. Uh he has a good uh a bunch of great great hot takes uh on LinkedIn, so uh give him a follow. Thanks. Ciao. This was Business AI Explained, and I am Vlad, founder of Elements Agents. Thanks so much for listening. Whether you are on YouTube, Spotify, or Apple, don't forget to like and subscribe. And I will see you next Tuesday for a new episode. Ciao.