Uncover Agent Value: Real-World Manufacturing Examples | Dreamforce 2025

Thank you for being our audience, for joining us at this session. I'm very, very excited that we're doing this. Also, thank you to you, Zomi, for taking the time and the preparation to share your success. For me, it's something very special that we're here because Brentac is someone in our industry to whom a lot of people look up, a lot of people look to you to see guidance where you're moving, what you're doing. So I'm very honored that you're sharing the story here and thank you for being part of this. Sami is our VP of global Commercial platforms. At Brentech, you have a leading position. You're connecting it and business. You're someone to reach out to. Also make sure to reach out to Zomi after the session. You're a thought leader in the industry, so it's wonderful to have you looking into what a manufacturing enterprise is in the agentic age. I think right now we're at a stage where a lot of us are saying, okay, I kind of see the value, but I don't know really where to go. And I also don't know what the target picture looks like. And it's probably early to say, like this is the agentiq manufacturing enterprise. But where we see the technology, we know that certain things are going to happen. One thing that's going to happen is everybody can be an expert. We know that we can take even very complex engineering knowledge and transform it in a way that we can make it available to frontline staff. We know that we can automate processes to an extent that we couldn't do before. And we've seen it, for example, today in the keynote, also in the context of Dell. And we know that both of this combined leads to a stage where employees can do things that they couldn't do before. And in particular in manufacturing. This leads to a target picture agent force for manufacturing where this goes far beyond customers interactions only. It goes into employees and operations. And I couldn't think of a more exciting story than the one of rentec, because you are a matchmaker in the market. You are connecting suppliers and customers. And I, without further ado, I kind of want to lead over to you, Zomi, and let you tell your story in this context. And why don't we start with Brentag as a company and your role?

Sure. Thanks. Thanks, Manx. So for the folks who doesn't know Brentech, so we are world's biggest chemical and ingredients distributor. So we operate in around 80 countries with around 18,000 employees, a few hundred thousand customers, we transact around 13 to 14 million transactions a year. So when it comes to chemical distribution, so this is a space where we manage highly regulated products, control product hazardous products and we have access to this unrivaled transactional data in our backbone, in our transactional ERP systems. And with a company that is 150 years old, 151 this year, there's a lot of history that we can tap into in terms of data and the supply chain related data points that we can really leverage as a commercial platform. And my role as a VP for commercial platform is really building that digital data connected platform that brings our customers, our suppliers onto a common platform and have that seamless experience with our internal customer facing Personas. To have the seamless experience. So that's much about brandtac. Yep.

And I think one special thing that you build is this commercial platform and I think now if you're not in chemicals or if you're not a chemicals distributor, you want to transfer this kind of to your world. But what is special here is that like BrandTech is connecting suppliers on the one side and customers on the other side on one unique platform and yeah, perhaps give us a bit of a spiel around that.

Sure, sure. Yeah, you called it out correctly, Max, earlier. Because we are really the matchmakers between the suppliers who are really looking for transparency and the reach on their products and the customers who are really looking into agility and speed and getting the products delivered. When we talk about the customers, we are talking about the customers in pharma, in food and nutrition, in energy. So the customer segment varies, the need varies. On one side we have a highly commoditized business which is very much transaction driven, high volume, throughput, it's all about cost and efficiency. On the other side we have the specialty business which is very niche, multi year project development. So it's a very specialized. Our challenge was to get all this into one common platform where we serve both the specialties and the commodities divisions together. And this is where really the salesforce come into the picture. Having that one nervous system that connects our customers, our suppliers and our internal Personas all together have that one seamless experience. And this is where really having that one common data model and one common platform concept came into play. And I think that's really what we are seeing in brand tag today. When we talk with our internal and external Personas that use the platform, they feel that seamless experience on Salesforce.

And part of your business model is of course like you said, that you have very different Types of transactions, you have like high volume running through actions and you have also these one off big projects. One thing where you started was this sales order automation and I think it's a fantastic example. You are quick. This is already established, it's running. How about. Yeah share some details about it, how you did it, why you did it and how you progressed.

So, so it was exciting. Dinner with our CEO and Marc Benioff during the last reinforce that kicked us off with this project. This is what we internally call as the order automation where we are essentially using AI agents to automate the incoming orders. We talked about the number of transactions in the volume that we manage, 15 million transaction ballpark a year. And these transactions predominantly consist of orders coming in in forms of PDFs, emails and things like that. So this is a very time intensive process. We're talking about 20 minutes ballpark to enter an order and we have customer service team members all across the globe, across 80 countries. And the first challenge we put in forward was can we leverage and deploy agents really to, to manage this complex task. And when we talk about how we deliver products it's not like Amazon that we ship a wrong product, that we can pull them back and resend the right product. Here we are talking about hazardous chemicals. So if you ship an industrial grade asset in place of a food grade asset, you know the complication that comes with it. So really this was, this was really was our proof point that with this agent that we can really leverage how to scale the order processing all across the globe. And I think the Transaction time from 20 minutes was cut down to 8 ballpark where these orders were really placed into the salesforce front end and it was just a matter of validating and pushing it through the backend systems. So if you think about it really this was something that we wanted to try out and we were able to get this rolled out in North America within the first few months and then right now we are taking it across the globe. I was seeing great ROIs, 3 to 4 million savings per year. And yeah this is a clear cut example of how we are able to scale fast across the globe. So AI agent doesn't need the morning coffee to call out the Polyethylene Glycol 400 chemical name. It doesn't mistype any product code but we actually caught it placing a lunch order once. All jokes apart but this is a very complex case that we wanted to establish and again with this we were able to claim success right out of the gate.

Super nice, thank you. So now the famous video would come, but I think we can't show it with sound right now. We can try it one last time. Nope, it doesn't work.

No.

Okay, so we're going to skip it and not. Thank you. But talk about a bit about the agentic manufacturing enterprise, because the second big thing that you did was product inquiries, right? And again to remind perhaps the audience, your business model fundamentally means that you have your own products, which you of course know well, but you also have to know well your customer, your suppliers, products in order to help your customers solve their issues. So perhaps share a little bit about the product inquiry agent, what you try to do and also the hurdles, challenges and success.

Sure, sure. So the product Inquiry actually we had a really cool demo that couldn't play because of the technical glitches. But what we launched was once we had established that confidence with the order automation agent, we took it to the next stage of really building that product inquiry. So this is where we launched the website based agent coming into play. So our customers can come in, check all the product details based on their requirements and then we were able to provide them the information and on the back end it was converted to lead and potential customer and then the SDR agents also kick in. So this was really a case where you had multiple agents coming into play. So you had the website agent, you have the product inquiry agent, the SDR agent or all of them coming into play. And this talks a lot. We can draw a lot of parallels with what we all heard this morning regarding the agentic enterprise solutions that we heard in the keynote this morning. So in the context of chemical industry or chemical distribution, also manufacturing in general, we could see the next big step is agents coming to augment across multiple functions, supply chain functions, sales functions. In our particular case it was predominantly on the commercial side or the sales side. So a good example that I can think out is where our order agent plays a inventory check and then the supply chain or the logistics optimization agent is able to ship their product to the right spot at the right time. So and another thing that we are also exploring is with our suppliers. So this is getting a bit outside within the enterprise. So if our, one of our biggest suppliers of chemicals is also big on Salesforce and they have their own trials and errors with agentic AI. So we are trying to see how these agents can talk to each other. So building their agents and our agents and how they can talk to each other, so that's also an extension that we see can have big impact as we scale the Agentic enterprise grids and.

Very tangibly on the order inquiry agent, you also transfer connections, inquiries that you take into leads, is that correct?

Yes. Yeah. So that was well captured in the demo video, but unfortunately we couldn't. So that leads then gets transferred over to the sales SDRs, sales development reps and that becomes a seamless experience for our commercial organization.

And so big picture, you have shown it a little bit already or talked a bit about. So agents talking to agents. Where do you see Brentag going? Because I understand your business model in a way. On the one hand, it's super important that you are the easiest company to do business with. Your customers rely on your quality, on the ease of the process, but you also need to deliver efficiency because ultimately it's in a market where it's very comparable. So you need to be as easy to do business with but also super efficient. Like what's the target picture for the agentiq?

So really now taking it to the supply chain, Supply chain is where we see a lot of strain in our organization. We are a very asset light organization and the transportation cost is really climbing up in the given context of everything going on with the tariffs. So we see that as the next horizon in terms of how we can build agents and take it to the next level.

Okay. And perhaps one point that we also I think should share because people often have an interest in like how to make it happen. You have built this unique data, you have this insight into the market that kind of nobody else has. You see what people demand, what prices they buy, you see what prices others require. How do you think about the value of the data? And also perhaps you can comment a bit on the data mesh that you built.

Yeah, so we have a complex ERP on the backend as most of you all might have. So we are talking about 20 plus ERP systems, homegrown, Oracle based, SAP base or a mix of everything. So our biggest challenge was getting all this data northbound and southbound into the Salesforce as a front end, that being front end facing our customers and suppliers. So what we introduced is a data mesh layer where we could get the data into that particular AWS based layer. And this is where the partnership between Salesforce and AWS came into play. And we were able to work together with these two vendors and really build that common data layer and which was already feeding to the salesforce. And with this concept of northbound southbound, we were able to get to the real time integrations also going back. And this is where we were able to leverage the concept of agents into placing orders back into the ERP systems.

Super. Thanks a lot. So I think overall, I hope we kind of portrayed the journey a little bit of Rentac, right from starting with sales order to product inquiries to the identity manufacturing enterprise that you see here. The cool thing is JB and the team, we're gonna have a roundtable on Thursday afternoon at 1:00pm, I think, where if you want to learn more and really discuss how you did it, we can go a bit deeper. We can connect. Eaton will also present jointly with Brentec and and so there is a unique opportunity to connect with the team also. Please, of course, feel free to stay in the room and connect to Sami directly or jb, who's here. The last question that we have for you is after we thank you quarterly for being part of this. If you like the session, please rate it. The coffee is on us. You're going to receive a $5 Starbucks voucher for the very, very long Starbucks queue down the block. But nonetheless, coffees on us. Thank you for being part of this. Thank you for being trailblazers interested in the topic and see all of you soon, probably tomorrow at the at the keynote. Thank you.

Thank you.