Welcome to the KuppingerCole Analyst Chat. I'm your host. My name is Matthias. I am not. I am lead advisor and senior analyst with KuppingerCole analysts. My guest today is Annie Bailey. She is an analyst working on emerging technologies out of Stuttgart. Hi, Annie. Good to see you.
Hi, Matthias and great to be back.
Great again to have you, as usual. We are talking about a highly interesting topic and we're talking about an area where you've just done some more extensive research on. We want to talk about business intelligence and business intelligence, and it's overlap towards artificial intelligence slash machine learning. What is business intelligence in general, before we add the AI part? What is it when we want to have a definition and to start from that?
So exactly it's, it's always good to know, you know, where it is that we're coming from before we figure out where we're going next. So if we think about business intelligence, this is in its essence, it's transforming data into an insight that a company could actually use to inform itself about its own operations. So it's really internal looking. It's, it's not analyzing trends, which are external outside of the company. And so business intelligence, we have to call it BI just for short BI tends to be kind of a finished product. You could think of, of data science and of analytics products. And so this is really taking, taking data and putting it into the form of a story which could be using visualizations. It could be using charts and graphs, graphs, or it could be presented in a interactive dashboard. That's on a certain topic. So for example, your, your sales for a particular region, and it could be all of the, the many different impacts and influences on sales for that region that you would have at hand in a dashboard ready for particularly decision-makers, but more and more commonly for all business users,
Right? So it's applying the individual business rules to existing data and getting informed decision making processes out of that. And when we are talking about that right now, why is this then a topic for you in this area of emerging technologies? Why are we today talking about it?
Exactly. So business intelligence is a tool, but in the context of the digital business, it can be really an enabler and, and help launch the business and give it an overview where it's perhaps over the past 10, 15 years lost an overview. And what I mean by that is that as businesses are becoming more and more digital, as they're anywhere in this process of, of going through a digital transformation, they're typically adding application systems, their environments are becoming more complicated, oftentimes more hybrid environments between cloud and on-premise. And with all of this added complexity, they can lose an overview because all of that data becomes siloed. This is a story which many have heard before, but it's, it comes down to a similar problem where as the volume of data within organizations is increasing the problems of silos it's being exacerbated, because some, you know, the state is being replicated across the organization, may not completely agree. There may be inconsistencies between these, and of course this data may not be shared across departments. And so really having an idea of what's happening across the organization can be very, very difficult as this process is happening.
Right? So maybe that, that is also a kind of, of, of trap more, the more an organization is relying on technology, the more data it produces, the more data is available for decision-making internal and external information. And, and to, to manage this in a proper way is probably something which makes it more and more difficult for, for BI to apply properly without additional help. Is this the, the, the, the message that AI for BI or machine learning for BI comes with when it comes to a promise towards the businesses?
Yes and no. And so I think you're exactly right when you say that, you know, it's a, it's a cycle that as there's more data in the business, then they need to be able to manage this. And it's, it's always escalating, but this could either be a vicious cycle or a virtuous cycle. And it has much more to do with the, the organization's approach to their own digital transformation than it does with actually using AI or machine learning to support that. And so what they mean by that is that if an organization is tending to collect technologies and tools, without any clear strategy of how that is going to concretely advance their business goals and stay with the changing environment and perhaps critically look at their own processes and business model and question, if that is going to match the changing times, then they're not going to get very far.
If they're using BI as a, as a solution to, to address silos in their business, that's not going to be enough, a BI solution. Isn't going to fix the problem of having very independent unconnected, siloed departments. And so if, if you're an organization which is using BI as your digital transformation, this is where it becomes a vicious cycle where perhaps you're not going to have enough control governance capabilities over your data in order to use that efficiently. However, on the other side, we could, we could think of this as a virtuous cycle. You know, if there really is a, a critical look at how the organization is able, able to meet the changing demands of today for their particular industry and business. And if they're able to adapt the right tools, being very choosy at which tools they bring in to help them achieve those goals, then perhaps BI can support in that process and propel them in their digital efforts forward.
Right. And you did research on that area of BI and looked at, I guess, so at exactly these use cases and produced this report that is entitled next generation BI platforms. And next generation of course, sounds like emerging technologies, but what is really next generation about these next generation of BI platform?
So you were on the right track earlier when you were mentioning artificial intelligence machine learning. So that's, that's in there somewhere and we're going to get to there in a moment, but what we've termed next generation is the, the shift from making data exploration really, and an expert topic where you, you have to be highly trained. You have to be a data scientist to really be managing the tools and understanding the, the modeling and the calculation steps that need to be done here. It's the shift from that to making data exploration available to everyone and enabling data-driven insights for everyone in the organization. And so in order to get there, you do have to abstract away the parts that would require an expert, your data scientist, your data analyst, and to do that often artificial intelligence or machine learning is used. So kind of hidden in inside the workings of this, for the goal of making it available to everybody, AI ML, this is being used to automate some of the common data preparations or the modeling of the calculations that we need to get there.
Right? So we are drawing a line between the, the data analysts, which is more a feature set, a set of capabilities of a person and the actual business expertise. So we were all commenting the, the, the expert from, from the business perspective, with the additional data analysts skills, and that is delivered via M L at home. Am I getting that right?
Exactly. And so that's one aspect. There are what I would consider three main ways that AI and machine learning is being integrated into BI platforms at the moment. So that's the first way is, is really using this for data preparation and, and kind of automating and, and reducing the complexity of some of the data preparation steps. Then we have it for focusing on generating insights themselves. So being able to recommend the most appropriate chart type for the sort of data that you're wanting to represent. So is this going to be better represented on a map or a bar chart? So preparing those sorts of insights or the third way is to throw in natural language interactions for, for, and between the user and the platform. And so what this means is a more natural query experience where you can simply type in a question or a partial question, you know, sales by region, and this type of query would then bring up an automated, an automatically generated chart, which matches your request. And so this allows for really ad hoc, querying of really exploring the data from the perspective of I business user or a decision maker, not necessarily from the perspective of a data analyst
On one hand, a bit bridging the skills gap, because you cannot buy enough data analysts. And it's some kind of democratising of these platforms, because it was just dig more difficult to use them properly. So usually you are looking also at trends are looking into the future, what to expect in these platforms in the future, or what is already starting to begin, what trends did you identify during the preparation of this report?
Yeah. So one of the trends we see is that BI is one of many finished products along the data chain or the data value chain. And what we mean by this is that of course data sets can be prepared for BI for use, for visualizations to make decisions about organizational processes and strategy decisions. However, data sets can be prepared for many other goals for many other finished products. And so we have to now consider how these data sets, which are prepared for BI could also be mobilized for perhaps data privacy programs, ensuring that the, the data access rights requests of individuals could be fulfilled through properly prepared data sets and well organized data sets. And that privacy would be maintained if those data sets are then used for BI or perhaps for building an AI or machine learning model for enterprise use using a different tool or working through Jupiter notebooks or any of the other machine learning development environments. So although BI is in itself, developing and expanding a really foundational piece and for organizations is going to be the data preparation and the maintenance of data sets for multiple different uses.
Okay. If I compare this, what you've just described with some other areas, for example, cybersecurity, there are also, we have AI machine learning coming in, and what they are promising here is that they, that they get more, more agile, more responsive, more to quicker tangible results when it comes to using the, the, the insights that they gain into security incidents to reacting to those. If we transfer that over to BI and machine learning, could that mean also that the results could be more immediately used for proper, faster, more agile business decision? Decision-making,
That's a perfect connection to draw here. One of the trends, and we see it in a, of the vendors at the moment, but this will certainly increase across more of the vendors in the space. And the next years is being able to connect action with the insight. And so in the, in the BI context, it's very common for a visualization or attach board to be embedded where the business user is already active. So within the CRM application that the organization is using where the business user goes to make a decision, the visualization is already there. And so this creates a synergy between using the insight to influence the decision. It's also becoming common to flip this, where if the user goes to an analytics dashboard, a BA BI dashboard to get insight, there could be, you know, perhaps a threshold which isn't met yet, which requires their action. They would then be able to take that action directly in the dashboard setting. And so that would require robotic process automation flow, where that decision-maker could then hit a button to respond directly to that, that data point, that insight, which isn't quite meeting the thresholds. So this will, in the, in the future, we're really expecting that this will become more nuanced, more natural to, to take action when you are consuming insights.
So it still needs the human factor to, to be the controller of what is actually going on when it comes to these decisions being prepared about augmented, or they augment the actual decision making of the actual yeah. The decision making of the in, in the business. So it's really supporting a more properly informed decision-making, you've been just working on this research. And as I understand this report is currently in the process of being finalized and published. When can we expect that to be out? Yeah.
So you can expect it soon, either late August, early September, actually it should be right around the same time as the EIC, the European identity and cloud conference. So if you are interested in this topic or topics like it, I'll be speaking on AI in general, the future of AI, whatever that means at the EIC. So I'd be happy to talk about this topic or, or others with you there.
Great. And there are many options to, to participate being either really fully digital remote, like our KC life event. And you can even join us onsite in unique in the middle of September. And who's interested as well. Can, can listen to me talking about hybrid cloud identity and access management, so different topic. That's interesting as well. So thank you very much any for joining me today for sharing your insights on that. Very interesting use of machine learning, really with a business connection that this often has this, this, this, this notion of being very technical very far off, but this is really supporting our today digital or today's digital businesses. So that was really interesting. Any final additions from your side that you want to mention? Anything that struck you when doing this analysis?
Yeah, I think the, the big takeaway here is that data governance is, is only going to become more important. And so, although we were talking about business intelligence today, the quality of the insights are all going to depend on the quality of your, your data preparation and, and the data governance of the organization in general. So it's, it's going to be hard to get away from the topic of data in the future. So that's something to hold on to
Great summary and also great thought, because this is something that we see across all businesses, if it meets machine learning and AI, but also the way how you treat data in general and how you manage it and make sure that you're controlling it, not only using it. Thank you very much again, Annie, for being my guest today.
Thanks for having me. And I look forward to speaking here again.
I'm looking forward to that as well. Bye bye. Bye
