Name: Spotfire Quarterly Update - October 2026
Uploaded: 2026-10-08T16:16:29.961Z
Duration: 37 min 18 s
Description: Spotfire Quarterly Update - October 2026

Transcript for "Spotfire Quarterly Update - October 2026": Hello, everyone, and welcome to the second edition of our Spotfire quarterly update looking at what's next in Spotfire. We're thrilled to have you with us today. I'm Jean-Philippe Richard-Charman, and I'll be your host for this session. Let me just start with what we'll be exploring today. Generating analysis has never been easier. Trusting it is another matter. Scrapped lots, rejected batches, and mispriced bids are the cost of an answer nobody can verify. Industrial teams need AI augmented answers. They can check against their real data, built on methods they can reproduce on the governance their organizations already runs. That's why Spotfire AI is our focus today. It brings AI augmented investigation into the visual industrial analytics platform engineers already use, working inside the analysis rather than beside it. Engineers and scientists can go from fragmented information to defensible decisions with domain expertise at the center of everyone. Now before we dive in and before I present our main presenter for today, I'd like to cover a few quick housekeeping items so you get the most out of today's session. Now the webinar itself will last up to around forty five minutes with a ten to fifteen minute q and a segment that'll be held at the very end. If you have any questions during the presentation, please use the q and a panel, which is located on the right side of your screen, and we'll address as many questions as we can during the q and a segment at the end. We've also made a couple of links available, linked to today's webinar, which you can access in the dot section of our webinar platform. So please feel free to access these. That's located, just in between the chat and q and a, icons on the right hand side of the webinar platform. Now after today's session, a recording of today's webinar will be made available on demand, and we'll email you the link shortly after the event. Now with that, I'd like to dive right in, and I'm excited to introduce our main presenter today, Niklas Amberntsson, our director of product management here at Spotfire. And with that, I'll hand it over to you, Niklas. Thanks, Jean-Philippe. So today, we are announcing Spotfire fifteen one, which is planned to be available in the end of the month. As usual, Spotfire fifteen one has many new capabilities, but in this presentation, we will focus mainly on how Spotfire makes AI an integrated part of the user experience through Ask Spotfire, the new AI assistant. Of course, the mandatory disclaimer slide. Most of what we will talk about today will be generally available in next few weeks, but plan could always change. So to start with, the problem that Spotfire was built to solve. Across semiconductors, energy, chemicals, and pharma, there are operational environments that generates enormous amounts of data continuously. And when there are problems, when there are issues, you need to be fast, whether it's about finding the root cause to a manufacturing problem or figuring out the right bid for an oil and gas lease. Almost always, data is required from different types of source systems. The signals that explain what's actually happening are buried inside the complexity of disparate data sources. Finding the signals to follow is rarely a dashboard question because it requires exploration, making new combinations of data, checking spatial data against equipment history, and applying statistical reasoning. It's also not linear. You explore. You refine your exploration. You test. You go deeper. You follow a lead that turns out to be nothing. You go back and take another path. In other words, understanding what data means requires investigation. Premodeled applications or dashboards, they are good at the problems that someone anticipated in advance. What we mean with investigation, it's the other kind, the questions that nobody thought to write a report for. In complex industries, this happens all the time. That's the category of problems that Spotfire was built for and the path that we continue to follow. Our roadmap evolved, and this is because recently it has became a lot easier to analyze data with AI. Load up your data to your favorite AI and ask it what the problem is, and it will analyze the data. And it will come up with several ideas and conclusions. It's fast, and it's easy to generate analysis. But validating the conclusions and trusting them to the level that you dare to take a high stakes decision did not get easier or faster. Industrial decisions are high stakes. The consequences of making the wrong decision, that can mean that you have to scrap lots of semiconductor wafers, can mean incorrect valuation of a lease of land, or something else, that's not good for your business. So the problem has moved from producing analysis to trusting the analysis and its conclusion. The conclusion that we have drawn from this is maybe the opposite of what people can expect. So the better that AI gets, the more valuable Spotfire becomes. AI needs context, and Spotfire is where that context gets assembled. And it's where the answer becomes checkable because a claim about an excursion can't be trusted until an engineer can see it and validate it with real data. Spotfire calculations are validated methods that give the same answer every time. So when you use Spotfire AI, the result is an artifact that survives, something that you can review, that you can reuse, that you can govern, and it has lineage. In Spotfire, we're using AI so that you can get an answer grounded in your real data with a path straight into the analysis behind it. We are enabling AI to use Spotfire to help engineers and scientists work faster and being able to concentrate on the high value decisions rather than the tool mechanics. With that as a background, Spotfire is the visual industrial analytics platform for engineers, scientists, and domain experts working in complex operational systems. So it brings visual analytics, AI, advanced analytics, and industry specific workflows into one environment. So AI in Spotfire is for engineers and scientists that investigate these complex operations. You can ask, use conversation as the entry point to any analytical task. Spotfire AI can assist by doing what used to require significant tool expertise or just lots of manual work. You can use agents and skills included in Spotfire that carry out real analytical work, or you can build agents and skills yourselves to solve your particular, business specific problems. You govern AI. You choose the AI models to integrate with Spotfire. You have detailed access control for your user groups. AI in Spotfire operates inside your analysis, not besides it. AI in Spotfire lets you drive Spotfire with natural language. It can build the layout. It can set up the drill downs, apply transformations, and put the analytical workspace in your hands. So one principle governs all of this. We don't build black box AI. Everything that AI creates in Spotfire lands as normal Spotfire content, visible, editable, still there when AI is done. You can expect it, and you can change it. So AI in Spotfire does what the user can do but can make it easier and faster for the user to get it done. AI in Spotfire has tools, skills, and agents that provide context and capabilities to AI. So when we say industrial analytics AI, we mean a few concrete things. So the structure is real. We know that the wafer has geometry. It has zones and exposure fields. A batch has phases. An oil well has a trajectory and is located on a lease. These relationships are known, so drill down and correlation are possible. The analytical methods are validated and deterministic, and they are hardened by decades of use because you are making decisions that scrap lots or price speeds. The result in Spotfire appears in the form your discipline already reads, be it a wafer map or a well log or what other typical visualization you would use. But, of course, we are not trying to anticipate every possible workflow and ship an agent or skill for every possible workflow. That's something that we couldn't do, and it would never be finished right. So we are also providing building blocks that you and AI can orchestrate together, find the data, enrich the data, reshape it, compute, render, and drill into the data, in addition to that, the ability for you or partners to create and share your own AI skills and agents. In general in Spotfire and particularly in AI, we try to be open to have interfaces so that you can integrate Spotfire into your environment and ecosystem. So we want you to be able to use your own agents, your MCP servers, your own models together with Spotfire. You connect Spotfire to the models that you trust. You deploy where you want to deploy on premise or in any cloud. Now, of course, everything that I described so far has to run-in an environment potentially with thousands of users, tens of thousands of analytical assets, and jobs that needs to be done because otherwise the business stops. So we keep investing in the enterprise side of analytics, standard container images, web based administration that keeps getting extended, scheduling and automation with conditional flow control, better workload distribution, notifications to tell the right person when something didn't run. We have more capabilities in the web browser in every release, and an extensibility framework where extensions can be centrally trusted and governed. But now let's look in more detail at what's coming in Spotfire AI in 15.1. Ask Spotfire is a natural language AI assistant that lets you use Spotfire features from the chat. It's the main AI interaction point in Spotfire and it provides a natural language interface to capabilities such as explain visualization, ask about data, explicitly invoking an agent to do something. But it also helps you create purpose built analytical workspaces for your data. You can drive the creation of your analysis by natural language, but you can also use the interactive user experience in combination with using AI. And even if it is AI, it's transparent so you can trust it. Whatever you create from Ask is like you would have created it yourself. You can inspect it. You can work with it, change it, and you can trust it. So here's an example of a specific command you can make. We are looking at wafer data here, and we want to divide the wafer into three radial zones of equal area and eight angular ones. Then we will ask Spotfire to update the wafer map on the screen here to show the zones by color in order to verify, that we got what we want for. So, I'm writing and asking Spotfire to start the creation of the wafer zones. And the... Ask Spotfire, the AI assistant, is processing the request and found an agent that can help by adding these zones. So a number of calculated columns were added, and then we visualize these, zones that were just created on the waveform app. So we can see what it looks like and make sure that this is what we thought we were... What we thought we wanted. Ask can also run calculations and look for patterns. So in this case, I'm asking to check if there are spatial or temporal patterns for the columns m one to m 10 and asks come up... Comes up here with some, findings. So AI used the Spotfire data engine here, so that's what happened, to check details for the columns m one and m two as requested, and it found that m two has an edge drop drop off for specific lots. M one also has a pattern where the center zones are significantly lower than the outer zones. But for now, we will ask Spotfire to investigate m two in more detail. So what I'm asking is to visualize the patterns for m two over wafer zones and lots on a new page using a box plot and a drill down wafer map. So the AI assistant is working and reorganizes the visualizations here. And we can see in the box plot that there are a number of outliers for a few selected lots. So if we mark those, we can also visually tie that, yes, for sure, these, these outliers are located on the edge of the wafer. So it's a confirmation of what, the AI assistance was coming up with. Another example here from the energy side of things, the Spotfire well log is popular in energy, and AI can work with this and, for example, create drill downs to specific parameters or create new well logs if you like to do that. So we are marking some data here and asking to create drill downs to rock hardness and drilling efficiency and also to weight on bit and rate of penetration. And we do get, when the AI assistant has finished the reorganization of the page here, we do get these drill downs, from the well log. So we can use the well log as we would normally do to mark data and see how the details visualizations, change. So a more complex example. So now we are going to try to ask Spotfire about showing us the yield trend by wafer with a drill down to wafer map trellis by wafer. And we will ask it to mark the eight wafers with the lowest yield. So note that there is no yield column in the data. So AI will need to figure out how to calculate the yield based on the bin data. So we see here we ask Spotfire, we close the table so we get a clean page here. Spotfire calculates the yield, visualizes yield over sort date in the line chart to the left, which I now moved up on top of the drill down and then I ask mark the eight wafers with lowest yield. And Spotfire figures that out and marks in the line chart and visualizes the details in the wafer map below. But I can still use regular interactive analytics to mark different sets of wafers as I need to. Apart from analytical tasks like this, Ask can also be used to search and navigate in the analysis, explain visualizations or pages, update visualization configurations such as changing column p2 to p3 in all visuals where it's present or changing the color or most of these common visualization tasks. So... But speaking about automation, skills are the next level of this. So skills in general in AI and in Spotfire as well are a set of instructions that help AI complete tasks and workflows. So with skills, you can very easily create your own custom instructions for analytical tasks, whether you want to make it easy to apply a set of data transformations, create pages of interactive analytical visualizations that include drill downs, or why not both of them. Skills can contain alternative paths such that if the data looks like this, then you do that or else you continue with another thing. So skills is a way of capturing your knowledge, your best practices in actionable assets in Spotfire. So how to create a skill? Just ask Spotfire. In this case, we will create a skill by, just asking it to do it by... With the description that's on the slide here. So it's a four step description that goes through using the wafer data prep agent to add five radial and eight angular zones to the data. Then I ask to calculate the yield per zone combination per wafer. And if the yield of any of the zone combinations is more than one standard deviation lower than the wafer average, then list the zone and wafer. And then point four here is then to offer to visualize the findings from step three, if any, in a wafer map that is trellised by wafer, and also some, specifics to have the zone combinations color yellow if it was listed in step three or blue, otherwise. So the skill runs and then comes up with a number of findings. So you see it says here that there were 167 zone combinations on these 25 wafers that exhibited yields that were more than one standard deviation lower than the respective wafer average. And it gives some more information. And it makes, a number... It lists a number of wafers and some combinations here. So if we agree to the visualization, we'll get this. So it... We can then visually confirm that the yield is, in most cases, it's lower in the edges. That's where the vast majority of the, let's say, where where the low performing zones are. But there are also some, like you see in, wafer four here, where we have some spots in the middle that actually yield a lower yield than the other, and the same here on number 13. So skills are really custom analytical workflows that you can build inside an analysis only, like I did. I just asked the Spotfire AI assistant to build the skill for me based on my description. And you can keep it just in your analysis, or you can save the skill to the library and share it with other users just like if they were mods, which in fact they are. You can also print the skill and review it. And if you want to, you can actually, take... Copy this, put it into your text editor, modify it if you want to, and then submit it back to Cloud Software Group and ask it to update the skill mod as you had, just created. In my experience, it may sometimes take a few iterations to get the skill right. Remember that I I only wrote, like, four sentences to describe this skill that that did this analysis. So that is, I mean, a fairly short description considering everything that was actually done, here. So that means that sometimes you you have to iterate it, and take a few, turns in order to get it to do what you want because the precision in natural language is not always as, as good as as one would think. So sometimes you you learn a bit for how you actually should express yourself when you're creating skills. We see another example here from manufacturing. So in this case, we are looking at, manufacturing data for a set of lots that have been processed. We are seeing the normalized mean on the x axis and the normalized standard deviation on the y axis. By looking at the normalized mean and standard deviation of quality measures, it's possible to assess if a process is capable of consistently producing outputs within the specification limits. So, of course, we also need to have specification limits here. So what we are going to do is to ask Spotfire to create a visualization of the, process capability zones. So a CPK zones visualization is a special type of visualization that you can create using a scatterplot in Spotfire, but it takes, configuring some reference layers and so on. And we... What we have done here is that we actually created a skill that does this for us. So I just need to type visualize CPK zones, and AskSpotfire will identify that there is a skill that can help with this. So it reads the skill, starts working with the skill, and started by adding the mean line, has added some zones and getting the coloring right to the zone. So in the scale, I was actually specific about the color that it should use for the different zones as well. And it also suggests some follow-up steps here, which we can choose to follow if we want to. But, as I mentioned earlier, we consider it very important to have open interfaces to enable Spotfire AI to be integrated with our customer systems, whether these are AI services, agents, ERP systems, maintenance planning systems or document knowledge bases, or issue tracking systems, or something else. In Spotfire 15.1, we added now secure and governed MCP connections that lets you connect to external systems and use these from AskSpotfire, from agents, and from skills. So MCP stands for model context protocol, and it's a more or less standardized protocol that is used by many AI tools to provide context to AI models. So MCP connections become available for asks and for agents on skill, as I mentioned. And the connections provide MCP tools that can be used to request services in the MCP server system. As an example, in this case, the Databricks MDS execute SQL MCP server allows executing SQL queries. So it's a number of methods that we can use. First, I use this for, looking what catalogs there were, then finding tables, describing tables, and then I used a query to figure out which is the most common equipment found in the data. We are also continuously building new industry native visualizations and actions and making them available on Spotfire Exchange. And we are now starting to build AI agents and making them available through the agent registry in the add ons browser. As a matter of fact, we already made some agents available there. So we have the equipment commonality analysis. We have suggest well log views, data preparation for wafer data, and the data recommender agent available. But this is an area where we're continuously adding new agents and making them available. And as I mentioned earlier, while we are providing industry native agents for semiconductor manufacturing and oil and gas, we know that some of you wants to build your own. So as we usually do, we have opened up for customers and partners to extend the AI capabilities of Spotfire by add... By adding custom AI agents that integrate seamlessly inside the Spotfire user experience. AI agents are developed similarly like mods and may use the LLM and data in the Spotfire document, can use calculations and data functions, can use data from information links and other library assets. So this means that you can build your own insight agents that incorporate your company specific rules and knowledge and turn them into a recommendation for a user. And of course, AI provides tremendous opportunities for the enterprise, but it requires careful governance and security. So Spotfire AI works with the AI models that you trust and lets the administrator configure access to AI models for user groups. So this means that you can control who is allowed to use AI, which AI models each user group is allowed to use. So administrators control access to MCP connections in the same way by defining connections and assigning access to users and groups. Spotfire AI provides the governance, monitoring, and security framework required for enterprise deployment of AI applications. So that's a quick tour of AI in Spotfire 15.1. We are really eager to get this out now, and in your hands and getting your feedback on it. But I also want to take the opportunity to show a few things we are planning after 15.1. First one being AI integration with compute and data platforms. So in Spotfire 15.1, you can use AI that runs inside Spotfire. It works with the Spotfire data engine within memory data. But we also, plan to make AI available when you're having data in Databricks, Snowflake, Cloudera, compute platforms where the data is stored. So running analytics where the data is stored, of course, is... Enables much, using much larger data sets than fit into memory when you're doing analytics on them. We also plans, we also plan to add Spotfire AI before you have either even loaded data in Spotfire. So in fifteen point one, you actually need to have an analysis open, in order to be able to use AI. But this is something that we plan to change so that you actually will be able to start Spotfire and use AI as your starting point for finding data, for loading data, and so on. The intention here is also to enable you to ask questions about your data even before you have loaded any analysis in Spotfire. So that AI will actually figure out what data you have if it has information about the conveyor belt on line two, for example, about the maintenance of that. And then report that back, in the ask Spotfire panel and enable the user to choose to open that or to to investigate the data further. We're also working on an AI assisted custom agent builder. So this is intended to let users describe their needs in natural language, and the agent builder will then generate an agent connected to the analysis and ready to be used and validated. So the idea is that the user can iterate the description with the agent builder and store and share the custom agent through the library. So this enables subject matter experts to automate tasks that help their productivity without relying on IT or software developers to build specific agents for them. We plan to put the same experience in place for data done... Data function development. So, of course, we want to be able to have AI to generate data function, the implementation of the data function code, but we also want to make the development environment of data functions in Spotfire more modern, more easy to use by actually integrating with, modern development environments like Visual Studio Code, enabling you to put breakpoints to step through the code of the data function, to monitor variable values, and so on to make it easier and faster to create advanced custom analytics in Spotfire. Another side of being open is that we also want Spotfire to provide services to other tools, to other AI applications, for example. So we are planning to add a Spotfire MCP server that may help AI applications to find Spotfire analyses on the library that are relevant to the user's needs, query running analyses to retrieve specific answers, such as providing a decline curve or estimated reserves for a specific well, or identifying the most common defect pattern across a set of semiconductor wafer lots. So the intention here is to enable AI users to leverage Spotfire through their AI application and make it possible for them to open Spotfire from within the context of their AI application. So, in summary here, then Spotfire AI is industry native, integrated AI that understands the industry. It's transparent to enable you to trust when you make high stakes decision, and it's open to let you integrate it with your operations and AI systems and provides enterprise governance for deploying at scale. But what I, as I mentioned earlier, we we talked about AI primarily here, but there are also a number of other improvements in Spotfire. We have seen the AI parts, but there are also new statistical methods available that can also be recommended by AI. The wafer map has new statistic layers, exposure fields, and auto layout detection. It's also operational aware for the process team, OPC UA, which is a vendor neutral route into process data alongside the Aviva Pi connectivity that we already have, plus, data cleansing agent for time series data. For administrators, there are new capabilities in automation services for conditional workflows, better email and data, and custom time format, among other things. And there are even more things than that, so I'll just pull up this slide, for a few seconds. But I can also recommend you to go and look at the what's new in Spotfire page that we have on our website community and, on our ID portal page. And just before we go into q and a, I also wanted to say that we are eager to connect with you about AI and other subjects, so do not hesitate to reach out to me or to contact your contact at Spotfire or through the portal. That's actually a great way to contact us. With that, I'll hand it back to you, JP. Fantastic. Well, thank you very much, Niklas, for that insightful presentation, and thank you once again to everyone who's joined us today. Now before we get on to our q and a segment, just a few things that we wanted to share. So as mentioned at the start of the session, regarding on demand access, a recording of today's webinar will be made available soon, so please do keep an eye on your inbox for the link. Additionally, we will be updating our events page and our what's new hub on Spotfire.com with a link to the on demand version of the webinar as well. Additionally, if you're interested in learning more, please do feel free to visit our website at Spotfire.com or contact us directly. There are lots of ways to interact with us, whether it is via our socials, through our community. Additionally, our blog site has lots of great content where we share the latest on visual industrial analytics, dive into Spotfire Industry Pro in more detail. And last but not least, if there are any enhancements that you would like to see or have ideas that you'd like to share with us, please don't hesitate to visit our ideas portal. So thank you very much once again for joining the second edition of the Spotfire quarterly update. As mentioned earlier, you will see quite a few on demand assets come up, tomorrow as well as our new innovation blog, our new what's new page linked to today's session, and, of course, the on demand version of the webinar. So once again, thank you for joining us today, and we hope to see you at one of our future webinars very soon. Take care, and