00:00:07 Introduction and Akshey Gupta’s background at Microsoft Dynamics.
00:01:53 History of ERP and how it emerged in the market.
00:03:30 The impact of cloud technology on ERP and modularity.
00:04:22 Akshey’s perspective on the current state of the ERP market and Microsoft’s focus.
00:07:00 Challenges in simplicity and extracting data from ERP systems.
00:08:00 The complexity of ERP systems and potential improvements.
00:09:17 Importance of a common language in the ERP system for user adoption.
00:10:57 The concept of a common data model and its benefits.
00:13:46 Long-term vision for ERP: simplification, natural user interfaces, and AI integration.
00:15:49 Future trends in the industry, focusing on AI and improving user experience.
00:17:48 Importance of snappiness and performance in software
00:19:28 Comparing user experience in B2C vs B2B applications
00:20:29 Using machine learning to improve on-time delivery performance
00:22:00 The challenges of AI in supply chain management.
00:23:11 The importance of understanding and managing probabilistic forecasting.
00:24:59 The challenge of accepting uncertainty in the supply chain.
00:26:37 Embracing digital transformation in manufacturing and distribution industries.
00:27:44 The future of ERP and supply chain practitioners.
In a discussion with Kieran Chandler, Joannes Vermorel and Akshey Gupta explore the evolution of ERP systems, emphasizing the need for simplification and modularization. They discuss the challenges in incorporating AI and probabilistic forecasting in supply chain management, noting that failure is acceptable if benefits outweigh errors. Both experts agree on the importance of integrating advanced technologies and probabilistic approaches, with Gupta highlighting rapid digital transformation and the need for companies to embrace new technologies to stay competitive. They envision a future of simpler, faster ERP systems with natural user interfaces and improved user experiences.
In this discussion, Kieran Chandler, the host, is joined by Joannes Vermorel, the founder of Lokad, and Akshey Gupta, who is responsible for Microsoft Dynamics ERP and supply chain solution sales in Europe, the Middle East, and Africa. Akshey has over 20 years of experience in ERP, Big Data, and AI solutions and has helped develop Microsoft Dynamics ERP solutions.
Akshey begins by sharing his background and experience in ERP, Big Data, and AI solutions. He started working in a customer organization, where he gained experience in ERP and planning and optimization. Afterward, he joined Microsoft and helped develop Microsoft Dynamics ERP solutions. Akshey later worked in the Big Data and AI organization within Microsoft, focusing on leveraging new technologies and simplifying insights for Dynamics ERP customers. Currently, his role is focused on technical sales of Microsoft Dynamics products, particularly for the manufacturing and distribution sectors in the EMEA market.
Joannes then discusses the development and history of ERPs. ERP systems emerged and became successful, mainly due to the limited capabilities of the internet during the late 70s and early 80s. The internet was not yet accessible or reliable for retail and manufacturing companies, and networking was incredibly difficult at the time. This situation led to the rise of ERP vendors that provided all-in-one solutions to manage various aspects of a company, such as financials, human resources, sales, procurement, and sourcing.
Companies like the German software corporation SAP and Microsoft became major players in the ERP market. SAP emerged as the leader for super large companies, while Microsoft became a strong contender, particularly in the mid-market segment where SAP was not as dominant. The core concept of ERP systems in the 80s was the all-in-one perspective with a monolithic architecture. However, the advent of cloud technology and the ease of interconnecting distributed software have led to a shift in this perspective.
In the new paradigm, ERP systems are now composed of various interconnected pieces that work together through the internet and cloud providers. This shift has significantly impacted the traditional all-in-one approach of ERP systems and has opened up new possibilities for businesses to optimize their operations. The discussion highlights the evolution of ERP systems over time and the potential they hold for the future in light of technological advancements.
The conversation revolves around the complexities and simplification efforts within the ERP landscape.
Gupta explains that Microsoft has primarily focused on the mid and upper-mid market segments when it comes to ERP solutions. He emphasizes the importance of simpler solutions that can be tailored to suit the specific needs of organizations, regardless of their size. Gupta states that Microsoft Dynamics offers modular solutions for sales, customer service, project service, finance, and supply chain management, with the aim of providing a productive platform underneath to help organizations move faster.
Gupta agrees that simplicity is crucial for user adoption and the development of a common language among professionals in the industry. He suggests that the complexity of ERP systems has grown due to factors such as company growth, mergers, and acquisitions. To address this issue, Microsoft is working on the idea of a common data model, which underlies the entire suite of Dynamics applications and integrates with other systems like Office 365. This common data model aims to streamline data access and reduce the need for users to navigate through thousands of tables.
Vermorel concurs with the importance of clarifying concepts and having a well-defined semantic structure for ERP systems. He believes that a clean and clear semantic representation can help other players like Lokad to rely on the provided data for their operations. Both Gupta and Vermorel seem to agree that while the current state of ERP solutions has room for improvement, their simplification and modularization efforts are essential for better user adoption and more efficient operations.
Gupta highlights the development of the common data model, aiming to include as many entities as possible while building industry-specific reference models. The long-term goal is simplifying ERP systems and incorporating natural user interfaces. He envisions a future where ERP can be interacted with through speech or mixed reality devices.
Gupta reflects on the challenges of incorporating AI into supply chain management, particularly regarding the acceptance of KPIs tied to probabilities. He suggests that more education and adjustments in the industry are needed before AI can be widely adopted in supply chain management.
Vermorel points out that despite the complexities, probabilistic forecasting is a crucial aspect of supply chain optimization, as it accounts for irreducible uncertainties in the market. He emphasizes that failure is acceptable in supply chain management, as long as the benefits outweigh the errors. He also acknowledges the challenges in changing people’s perspectives, especially regarding zero-defect mentalities, which may not be realistic for supply chain management.
Gupta highlights that digital transformation is happening rapidly and that companies need to embrace new technologies to avoid falling behind. He sees a bright future for ERP practitioners and supply chain practitioners, with growing interest in probabilistic optimization and cloud ERP solutions. Both Vermorel and Gupta emphasize the importance of integrating advanced technologies and probabilistic approaches in the field of supply chain management.
Kieran Chandler: Today on Lokad TV, we’re delighted to be joined by Akshey Gupta, who’s responsible for Microsoft Dynamics and supply chain Technical Sales in Europe, with a particular focus on manufacturing and distribution. Today, we’re going to discuss with him the rise of the ERPs and understand how they have more potential in the future. So Akshey, many thanks for joining us today.
Akshey Gupta: Thank you for inviting me here. To start, I can just tell you a little bit more about my background and my role at Microsoft. I’ve been focused on ERP, Big Data, and AI solutions for 20 years. I started working in a customer organization, where I cut my teeth into the ERP and especially the planning and optimization area. After that, I was hooked onto it. I joined Microsoft and helped develop Microsoft Dynamics ERP solutions. For a while, I worked with the Big Data and AI organization inside Microsoft, looking at how we can leverage new technologies and bring those insights to make it simpler for the Dynamics ERP customers. Right now, my role is really focused on technical sales of Microsoft Dynamics products, focused on manufacturing and distribution in the EMEA market.
Kieran Chandler: And as always, we’re joined by Joannes. Today, we’re going to be talking a bit about the development of ERPs, and they started from really humble beginnings. Perhaps you could just tell us a little bit more about the history.
Joannes Vermorel: What’s interesting about ERP is that it emerged and made the fortune of famous companies like SAP and a few others at a time when the internet technically existed but was not practically accessible for retail or manufacturing companies. In the late 70s and early 80s, if you wanted a system to run your company, you needed something that would cover all your needs in one package because networking in the early 80s was incredibly difficult. This situation gave a massive boost to vendors that offered all-in-one solutions to manage everything in your company – financials, human resources, sales, procurement, sourcing, etc. Many large players emerged from that, with SAP being the leader for super large companies and Microsoft being a strong contender in this market, particularly in the mid-market segments where SAP is not that strong. Now, I think we are facing a brave new world, as the core of the ERP initially was this all-in-one perspective with a monolith. With the cloud and the ease of having distributed software connected together through the internet or cloud providers, it brings a wave of change and novelty to this world where many solutions have been around for decades and are not necessarily considered super innovative, considering the pace of change that you observe in software in general.
Kieran Chandler: Today, we’re going to be looking a bit towards the future of ERPs. Akshey, what’s your current take on the state of the market? Would you say it’s super innovative, like Joannes mentioned?
Akshey Gupta: I think there is a massive change, as Joannes mentioned. Everyone realizes that to make ERP really productive and useful for most organizations, it has to be simple and modular. We like to think about this from an organization’s perspective, considering the span of control.
Kieran Chandler: Joannes and Akshey, I’d like to discuss the state of the ERP market, especially for large organizations. Microsoft has been leading in the mid and upper-mid market segments. How do you see the current landscape?
Akshey Gupta: I believe that in the ERP market, there aren’t really any super large organizations. Instead, they are a combination of large organizations. Microsoft has been focusing on the mid and upper-mid market segments, which form the basis of larger organizations. Large organizations are composed of various processes and responsibilities, making it difficult to have a single system across the entire organization. That’s why we focus on developing simpler, modular solutions that work for parts of the organization, whether they have a billion-dollar revenue or five billion dollars. We’ve moved away from monolithic systems and now offer modular solutions like Microsoft Dynamics, which includes sales, customer service, project service, finance, and supply chain solutions. The idea is to keep things simple and use our platform to compose and adapt these solutions quickly. Our competitors are also moving towards similar modular solutions.
Kieran Chandler: Joannes, how do you think current ERP solutions are performing, particularly in terms of their simplicity and modularity?
Joannes Vermorel: Simplicity is a worthy goal, but the reality is still challenging. At Lokad, our supply chain scientists often face situations where we start probing the data contained in the ERP systems and find thousands of tables and columns. There can be multiple ways to represent sales data within the same ERP. While end-users value simplicity in terms of screen design and ease of use, we at Lokad focus on the data and find that the situation is far from simple. There is enormous potential for improvement across the industry.
Kieran Chandler: Akshey, what are your thoughts on the complexities that ERP systems have to tackle?
Akshey Gupta: It’s true that ERP systems have grown complex over time. Simplification, especially from a UI perspective and ease of use, is crucial for adoption. An ERP system serves as a control mechanism for various parts of an organization, and it also creates a common language among them. However, it is important to address the complexities and continue working towards simpler, more efficient solutions.
Kieran Chandler: Joannes, you mentioned earlier the importance of a common language when it comes to software applications. Can you elaborate on that?
Joannes Vermorel: Yes, definitely. The absence of code processes being managed by one application means that you have multiple applications doing that, and people speaking their own languages. In the 60s and 70s, an invoice was not called an invoice everywhere in the world, and sales orders were not sales orders everywhere. Now, we have common terms because people have adopted a common language. So, I think that’s a very important purpose that they serve. Simplicity is important from a user perspective, otherwise, we won’t have adoption. We can deal with complexity, but nobody’s going to use it in the end. We want our users to be able to interact with the application and use it. So, I think that’s really key about the simplicity.
Kieran Chandler: And what about the large number of tables and fields in software applications? How do you view that?
Akshey Gupta: Well, I actually see that as a good thing. It wouldn’t happen if people didn’t like the software they were using. We have seen people using Microsoft Dynamics software for the last 20 years, and they move to newer versions because they like what they have. As companies grow, mergers and acquisitions happen, and the business expands, it needs to be represented in some way. When it goes beyond the span of control of a few people or new people join the organization, they try to represent things differently. This is a natural cycle of development that shouldn’t be stopped. However, what we are doing to keep it manageable is the idea of a common data model. This common data model underlies the whole set of Dynamics applications, and we also use it to integrate other systems like Office 365 and other technologies. We even welcome contributions from other vendors if they’re willing to contribute to this common data model. So, the idea is to avoid the need to go through thousands of tables to look for sales data. Would you agree with that?
Kieran Chandler: So Joannes, can you tell us a bit about the importance of language in supply chain optimization and how it affects the way we represent data?
Joannes Vermorel: The best way to introduce a bit of simplification is to clarify the concept. The language is very important because one of the reasons why you have so many tables is that the concept is unclear. You end up with a proliferation of tables that attempt to represent pretty much the same thing, but you are not super clear. Actually, you end up with a lot of tables that are just values.
Akshey Gupta: And indeed, a great design is when you invest time in having a very clear concept of what an invoice is, for example. You really clarify what it is and what it isn’t. You clarify what a sales order is. Then you have clear entities, as you described. It’s like a higher level of abstraction on top of the underlying SQL representation. This is a commitment as a vendor to maintain this semantic very clean over time so that other players like Lokad can rely on the fact that the data will be provided with those very well-defined pieces of semantic. This is an invoice, even if underneath the fine, low-level representation kind of fluctuates to make the best of whatever is the latest feature of Microsoft SQL Server, for example.
Kieran Chandler: Let’s talk a bit about the future then, and what’s kind of the vision Microsoft is looking forward to over the next decade or so building upon that common data model kind of idea.
Akshey Gupta: So, from a common data model perspective, of course, we want to include as many entities as can be represented in this common data model. The short-term objectives are obviously to have industry reference models built on top of this common data model. A sales order is a sales order, but again, it’s a little bit different if you have a subscription service that’s doing subscription sales as opposed to product sales and so on. So, there are a lot of differences per industry. From that perspective, we’ll have our partners…
Kieran Chandler: Joannes and Akshey, can you talk about the future of AI in ERP and what trends you expect to see in the industry over the next decade?
Akshey Gupta: We’re already building a lot of industry-specific models on top, but I think, overall, in a long-term perspective, our focus is more on making ERP even simpler. Natural user interfaces are very important to us, and that’s where we want to go. For example, a salesperson out in the field should be able to just talk to a device to find out what happened to a customer order, and the device should be able to tell them in natural language. This is where the domains of what you do here at Lokad and what we do from an AI platform tooling perspective combine with business process automation. We’re focused on natural user interfaces, whether it’s mixed reality devices taking advantage of what’s inside the business process software or speech technology to talk to your ERP system. We’ll continue to see more of this, along with simplification and personalization. We’ll work on using AI and other technologies to really simplify the user experience. Right now, we see bloated tables and screens in current ERP systems. For instance, if you’re doing data entry and entering a sale, you have to tab through 15 fields to get to the place where you want to enter a value. With telemetry in cloud systems, we can see how users are interacting with the application and intelligently present only the necessary fields. This will simplify the user experience, whether they’re talking to the ERP or manually entering information.
Kieran Chandler: Joannes, what trends do you see staying put in the industry over the next decade or so?
Kieran Chandler: Humans, they don’t even perceive that, because if you don’t have some millisecond response time, once we compile all the world’s solutions that are made of many layers with partners and all those things that are interleaved, we end up with this three-second response time that is actually just the sum of many things that are individually quite fast but not just as fast. So, I think that’s going to be a big challenge, because actually going for millisecond or sub-millisecond latency is very difficult. I think it’s going to be something subtle, but I suspect it will make a difference, and that will really differentiate the super capable vendors from the ones that are just capable of doing, let’s say, crude create-read-update-delete screens that are just dumb. If you look at big successes like Instagram and WhatsApp, they are really people that have mastered super snappy things. We’re talking a lot about the user experience, and what we’re seeing is that the user experience a lot of time, when it goes to B2C kind of applications, applications that the consumer directly uses, they tend to be a little bit better than B2B kind of applications. So why do you think it is that they’ve been slightly left behind, and why are we not seeing these quite so nice applications to use in a B2B perspective?
Akshey Gupta: I think historically, these things have come out of web technologies. All this advancement happened in the internet world, and the business world applications are still catching up to that. It’ll happen, like you rightly pointed out. From my experience working in the AI organization to look at how we can use those technologies for supply chain complexities, I realized that it was quite interesting and eye-opening. We were working with a customer and solving a complex problem for them where they wanted to improve their On Time In Full (OTIF) delivery performance. They said, “Look, why can’t we use some machine learning? I hear so much about it. Can you do something for me and improve my orders?” So, we did a lot of work and came up with this idea that their current OTIF is 93%, and it can go up to 95% based on the research we have done. There’s an 80% probability that they’ll get that OTIF in certain other parameters. The customer didn’t understand. They asked, “What do you mean 80% probability that I can get 95% OTIF? I already have 93%, so why can’t I get 95% all the time?”
Joannes Vermorel: I think it’s important to understand that there is always a probability involved, even with the current 93% OTIF. It’s not always certain. In the supply chain world, there’s no notion of probability, and I’m generalizing here. For that person, it was like going from something certain to something uncertain. I thought, okay, maybe it’s going to take a bit more education, and a lot of different things need to happen before we can get to the place where people will start accepting these KPIs which are tied to certain probabilities. They’re always tied to probabilities; nothing is certain, but we think of everything as certain in the supply chain world. We might need to wait with AI in that particular segment and solve problems where uncertainty is okay, like UI. For example, if you don’t get the perfect UI or the search terms you were mentioning earlier, it’s not going to break anything down. It’s not like you’re going to buy certain stocks depending on that probability. The cost of getting something wrong is much higher in the supply chain world, and the perception is that if you
Kieran Chandler: Scenarios like, would you say that for me, in that example, we’re getting customers that are getting very tied up in these buzz words like AI, machine learning, and things like that? Would you say that’s one of the real challenges over the next decade is people actually having an understanding of the mathematical optimizations that are happening in the background?
Joannes Vermorel: I would say it’s even more simple than that. I think you’re completely spot-on when you say there is a probability. It’s already there, they’re already measuring this probability as a percentage of success for certain processes like on-time delivery or just having no stock-outs with the service levels. It’s already there, but indeed, that’s one of the biggest challenges we face at Lokad. We are actually not even marketing things like AI. If you look at our website, when we say probabilistic forecasting, we are already struggling to just discuss with participants and bring the discussion to a state where we agree that the future is going to remain uncertain, no matter what, and that there is irreducible uncertainty.
It’s very funny because, for example, in fast fashion, it’s kind of obvious. If Lokad were able to predict which product is going to be a hit in the market next December, we would not be doing supply chain optimization; we would just be playing the stock exchange. So, when we are facing things that are erratic, irregular, and intermittent, as is frequently the case in supply chain, you have irreducible uncertainty, and thus you end up with this probabilistic reasoning. Indeed, it’s a challenge. Many supply chain areas are still unfortunately stuck with a zero stock-out perspective or zero defect. This can work in manufacturing, where you have complete physical control over the process and can have, as has been done in automotive, literally zero defect processes. But for supply chain, where you’re relying on people driving trucks, stuff happens.
Kieran Chandler: We must start wrapping things up now. So, the last word to our guest: with the advances we’re seeing and your experience in the manufacturing and distribution industries, would you say that people are ready to embrace change and things are happening quite quickly now?
Akshey Gupta: Absolutely, I think everybody realizes digital transformation is happening. It’s going to happen, and if they don’t take the lead, they’re going to fall behind. I see a lot of customers coming to us and looking for explanations about cloud technology and how it’s better to have a cloud ERP solution as opposed to the old-school on-premise solution. People are really looking at understanding that they need to embrace new technologies and embrace them in a big way. We have a focus on the mid-market and the upward market, and these organizations are procuring software on their own to do their own digital transformation, irrespective of what happens from the corporate side of things. So, I think there is a huge movement in this field, and the future is bright for ERP practitioners as well as supply chain practitioners. I personally would like to see a lot more happening on probabilistic optimization type of technology and bringing that closer to ERP.
Kieran Chandler: Brilliant, well thank you. That’s everything for this week. Thanks very much for tuning in, and we’ll see you again on the next episode. Bye for now.