Seeq Industrial Analytics & AI Suite

As a Process Engineer, my main use case for Seeq is to treat the data from the PI ProcessBook, the signals, in order to do some diagnostic on the problems on the process.

For example, if I have a problem in the temperature and I want to know at a specific moment how the temperature has been reacting, I combine that with the positions and directions of the thermocouples and verify if the burners are working well at the same moment. This allows me to set certain conditions to find the moment that is problematic.

That's principally how I use Seeq, but I also use it to create notifications, such as email notifications, so I am able to know what is going on under specific conditions. This helps us react faster when problems arise.

I think Seeq's best features are that once you know how to use it, it can do pretty deep analysis. The data treatment helps you identify problems, which is very hard to do in traditional ways such as Excel. The things you build are usually easy to modify, and there are features such as notifications, dashboards, and tree maps that help you visualize your data in different forms.

As a Process Engineer, I find the deep data analysis to be the most useful feature because we often try to identify what the regular pattern is, what the problematic patterns are, and what is going on at the same time with which equipment is not running well. With Seeq, we can analyze multiple data at the same time and supervise those conditions.

Of course, when data is processed correctly, all three outcomes—saving time, reducing downtime, and improving efficiency—are achievable. We save time because many datasets are periodic, so we can apply the same conditions once and treat data in the same way for the future and the past. We reduce downtime when we can identify the real problems with equipment, which allows us to implement new measurements quickly. Ultimately, we improve efficiency as well.

I believe that with the traditional way of processing data, which is Excel, many people ignore the power of the data. Most people tend to look back only at the last two or three days of data and then try to take action. However, with Seeq, we can process data over much larger ranges, such as six months, one year, or two years, very quickly. This helps us understand trends and identify which parameters have the most impact on a process very fast and deeply.

I think the new feature that Seeq offers, which is their AI assistant, is pretty interesting, but I find that sometimes the results provided are not exactly what I'm asking for. I feel there is still a lot of manual adjustment needed with Seeq. In the future, if a platform such as this had AI integrated in such a way that I could just show what I want, even with a picture, and have everything ready, that would be even cooler. However, I haven't tried this feature in Seeq and I don't think they have it yet. That's something I would suggest Seeq improve.

I think it would be beneficial if our IT department could give more permissions and link more data to Seeq so we wouldn't have to rely on other platforms to manage information. It would be beneficial to treat all data on the same platform more smoothly. While Seeq is powerful with signal data, it is not very powerful with data that exists in Excel sheets.

I find Seeq useful, and there are many new features I haven't explored yet. I believe they are continuously building more features, so I am confident that one day the platform will be even more powerful in data analysis. However, the downsides are that it still requires a certain level of expertise to use the tools effectively, and it seems to be built for someone more devoted to the applications. I believe it can take some time for a regular process engineer to really get hands-on with it. There are a lot of features I haven't explored, and I feel I would benefit from guidance on those things, as it can be a bit hard to navigate alone.

I have been using Seeq for around two to three years.

I haven't tried any other platform, so I can't really make a comparison, but I rate Seeq at seven point five to eight. So, eight.

I haven't tried any other product, so I can't really make recommendations. However, if your organization decides to go with Seeq, my advice would be to try their AI assistant, as it is pretty interesting and could save you a lot of time.

I rate Seeq an eight out of ten overall.

My main use case for Seeq is incident investigations. Upon unexpected quality results, incident investigations can start to determine where the incident first began occurring. Seeq allows the retroactive examination of many different ingredient load cells as well as equipment settings to help isolate the root cause.

Seeq has positively impacted my organization, and I can share a specific outcome where we have found a new metric on an important piece of equipment to raise the flag when it is underperforming. This prevents large production issues from occurring when we're able to catch the underperformance through the Seeq alert.

The new metric was discovered from a production issue, and investigation using Seeq allowed us to identify this key metric and start monitoring it. It has improved our quality consistency since implementation.

The best features Seeq offers include email alerting, which is extremely helpful. We are able to monitor specific critical metrics in our process and alert the team when one falls out of bounds. This prevents larger issues during production.

The Seeq alerting integrates very well into our workflow, into shared messaging accounts as well as individuals' emails.

I believe Seeq can be improved because report rollouts on similar pieces of equipment can be challenging. If Seeq had an easier way to modify our tag structure, then duplicating reports onto similar equipment throughout our facility would be easier. Improvements around user interface or other functionalities would also be beneficial.

I have been using Seeq for four years.

Seeq is stable for me, and I have no issues with reliability.

Seeq's scalability is functional enough, but if it was easier to duplicate reports on similar tree structures, it would be even more scalable.

Customer support for Seeq has been great, and I have had positive interactions with their support team.

I have seen a return on investment from using Seeq because we are catching production issues in a faster manner than we had before Seeq.

My advice to others looking into using Seeq is to pursue it. There are more opportunities with your own equipment than you know without examining them through the lens of Seeq. I would rate this product a 9 out of 10.

My main use case for Seeq is process optimization, any coding tool for Python, and any work that is related to time series data. We upload the data into Seeq and do the machine learning, visualization, and many other analytical tasks.

I have more to add about my use case for Seeq because it is consolidated to one place. It provides visualization, machine learning tools, any places where we can call the API to pull the data, and a proper way that we can integrate everything in just one place and use it easily.

When I mention the drag and drop and right-click features for visualizing and analyzing large sets of time series data, I see that it has changed my workflow significantly. Before the age of AI and LLM models, this was the very first tool that introduced optimization work to non-IT engineers who had not been able to code it, bringing them into a more optimization environment. Currently, with the age of generative AI and LLM world, we can accomplish similar tasks with other tools as well, but Seeq remains the core foundation for us. I would say it serves as a complement to other tools. We just need to supplement other data into Seeq and we can leverage it anyway.

In my workflow, Seeq integrates with other tools or systems as we try to incorporate more of Seeq into our daily operation, meaning the operational decision-making.

When we try to optimize or make any adjustment to see the equipment performance, we use Seeq extensively, and then reliability comes into play. So, we need to maintain the reliability of Seeq server much more because we make it up to the same level as our control panel and control system. I would say Seeq is coming to integrate much more and has more influence on our operational decision.

Seeq helps support collaboration among my team or across departments; I appreciate their sharing function and their journal, which allows us to record every comment, every action, and every analytics step that we did. When we share it with our partners and colleagues, they can view those journals and understand how we arrived at those formulas, how we built up the analytic dashboard, and how we came up with operational conclusions. I would say Seeq does a great job on this.

Seeq helps my team make faster and better decisions; for example, in an exchanger optimization problem, when we need to determine the time to clean the exchanger or shut down the plant for cleaning, Seeq is able to build the visualization dashboard to monitor that very quickly. I would say it takes less than an hour to do that. It is a tool for manipulating, analyzing time series data, and visualizing it up to an endpoint, which is the optimization problem to make a business decision. For anything about time series, I think of Seeq as the first choice.

In my opinion, the best feature Seeq offers is the ability to visualize time series data in an efficient way. If we code it by ourselves, it will take a very long time or require a lot of server capacity, and you cannot simply plot a time series of, say, 100,000 records on your own simply because it takes a lot of effort to do that. Analyzing it when we apply any formula or algorithm takes more time to finalize everything. Seeq does this through drag and drop and point and click actions. So, it is much easier to do it by using this tool.

Seeq has positively impacted my organization because I see many people using it, compared to the past five years when we had only PI Vision for visualizing time series data. Manipulating time series data was such a critical task that not many people were familiar with and were afraid to do. This task remains a core critical function for everyone to do it efficiently in order to complete anything. Process optimization and reliability analysis would address all failure modes.

I appreciate the question on how Seeq can be improved. The first thing is their graphical user interface. They invest so much in their backbone, but without a good dashboard or visualization tool, it feels insufficient. They focus too much on technical aspects. Many managers and people just want a simple dashboard that can show everyone in the control room or whatever places that when we finalize the tool, these are the final product. Seeq did not do it as beautifully compared to other applications.

I really want Seeq to improve their graphical user interface to be more user-friendly.

I have been using Seeq for about five years.

In my experience, Seeq is stable on production servers. Recently, we tested a new function on a test server that is still in the development phase, and while I can understand that, I expect a bit more. Testing should be reasonable to production as much as possible, but it is treated differently and lags a lot.

Seeq's scalability is really great. Any similar assets or problems can be copied and pasted from one business unit to another, applying them to another business unit with the same function, just changing the tags.

The customer support from Seeq is very great. They have a ticketing system and on-site support engineers located almost everywhere around the world. I am located in Thailand, and they have a Malaysian support engineer dedicated to my project, who gave me his WhatsApp number so I can reach him whenever I need.

I previously used PI Vision, Excel, or PI add-on tools with Excel. I used Python and R before I met Seeq. Seeq handles all of that in one place, providing every tool that I need. That is the main reason we use Seeq, and I love it a lot.

I have seen a return on investment as I saved a lot of millions of dollars for my business unit since we started using Seeq. Many business decisions, operational troubleshooting, and optimization have been accomplished with Seeq. My organization is small, and while we have improved our efficiency, we have rarely had to lay off employees because of this improvement. I would say there are some recognized results about time saving in our team, but they are not quantifiable numbers.

Before choosing Seeq, I evaluated other options like Power BI. When discussing dashboards, people think about BI solutions like Tableau or Power BI or software like RStudio, which are usually used for data analytics work. However, comparing them with time series data, nothing beats Seeq. Those software handle relational data relatively easily, but the volume is small compared to Seeq, which specializes in time series.

Seeq handles large volumes of data or high data velocity very well. I used to use Python before I started visualizing my time series data, and even with just not many variables, about 10 to 12, my computer worked very slowly compared to Seeq. I think they have a special algorithm to process and select the appropriate data points to visualize and process high volumes. It requires a lot of servers and special tools; it is not just a simple Excel file typically used in our learning. Seeq handles a large volume, and time series data is not simple data; it contains every detail and requires specialized software to manage that.

My advice for others looking into using Seeq is that if you are looking for a finished product for time series, Seeq is a good one that you should try. It is easy to build, and if your team members or engineers are not familiar with coding, Seeq is a good tool to start. You can do machine learning and advanced analysis that you cannot perform simply because you do not know how to code. You can just try, click, drag and drop and see the results, and you will be obsessed with it once you see the results. I would give this product a rating of eight out of ten.

My main use case for Seeq is mostly advanced analytics including machine learning and data science to solve industrial problems such as golden batch analysis, soft sensor modeling, hybrid modeling, and condition-based monitoring.

A quick specific example of how I have used Seeq for golden batch analysis is that Seeq has some of the most handy and fantastic features for analytics. They call it Capsules, which is used to create a condition in a matter of a few clicks. You can create conditions as complex as multiple nested loops. The way we use these conditions is to identify the best historical batches, good batches meaning, out of perhaps two to three years of data, and we identify which were the best batches within certain specifications. We average out those batches, create a standard deviation, and we use a reference profile from Seeq. Seeq really comes with out-of-the-box features and functionalities which can be used readily without even knowing about the mathematics behind it, without even doing the coding. You can quickly build your entire workflow for the analytics in a matter of a few hours.

Using Seeq for workflows like identifying the best batches and creating reference profiles really enables decision-making. Now I know for my current batch what the efficiency is in comparison to the best golden batches from history. If there are any deviations in real-time, Seeq throws out an alert or notification which can be utilized for me to know that there has been a point where I need to decide what needs to be done, if I need to make modifications to any of my process parameters or something of that sort.

I think there are plenty of other use cases such as heat exchangers, compressors, and reactor modeling. Using Seeq for a heat exchanger is the most prominent one across chemicals, oil and gas, and pharmaceutical industries. Using physics-based calculations, such as thermodynamics, to estimate the fouling factor. Once you estimate the fouling factor based on the data, you forecast it to the future and identify at which point in time it is going to go below the minimum design value or minimum acceptable value. If you take the delta difference from the future point where it intersects the minimum value and now, that is what the remaining useful life is. It has been very handy when you are dealing with a fleet of heat exchangers. You now know which heat exchanger needs to be attended to at what point in time because you can prioritize the exchanger performance based on the fouling factor looking into the future.

The best features Seeq offers include a module called Workbench, and Workbench comes with abundant tools, pre-configured tools that you as a process engineer—I mean, you do not need to be a data scientist—can readily use for your analytical workflows. You may have some thought process or some modeling approach in your mind which gets started from cleansing of the data to modeling the data all the way to conditions, machine learning modeling, and dashboards. The way Workbench is structured is a very concise way of putting the tools together. As an engineer, I just need to go from top to bottom, and it helps me to build the model. In my daily workflow, the specific tools or steps in Workbench I find most valuable are Conditions and Capsules.

Seeq is doing what they can and they should. I do not have anything more to add about the needed improvements.

I have been using Seeq for seven years.

As a partner, I cannot share the information about whether I have seen a return on investment.

I am satisfied with Seeq. I have looked into many other platforms, and Seeq has been one of the most useful and very easy to adopt and consume by the users, by plant people. Being a data scientist, I can love products which could be more complex, but at the shop floor, the process engineers need something that is useful for them. Even before that, they need something that they can use without solving complex mathematical problems.

My main use case for Seeq involves monitoring data trends and studying historian data to conduct modeling.

A specific example of how I use Seeq for monitoring and modeling is through data monitoring, which is quite important because Seeq has many functions to filter out data that is not needed, such as plant stoppage or abnormal conditions. It performs excellently in terms of filtering. Based on the formula, I can also create many predictive models to determine how the predictive line looks when certain parameters are increased. Additionally, Seeq can generate very good reports based on trends with explanations. These three aspects are what I focus on primarily with data monitoring.

Regarding modeling, because of the filtering features, it helps me quickly develop a historian model, producing a good trend line for modeling. For example, as a process control engineer, whenever I implement automated systems, I have to monitor the trend to see whether the scripting I put in aligns with the actual scenario. Seeq can present several trends on a single page for quick comparisons, and the timeline is quickly accessible, allowing me to see values such as the average, minimum, and maximum efficiently.

In my opinion, the best feature Seeq offers is the filtering. Previously, when using traditional tools such as Excel, extracting data caused the system to lag, and filtering required setting specific conditions, often more than one.

Seeq stands out for me in filtering compared to traditional tools such as Excel because it is easier and more powerful.

Seeq has positively impacted my organization by making work more effective. It used to take months to collect data, model, and filter, but now Seeq saves time, allowing us to complete tasks in a week or less. Additionally, we used to utilize many tools to conduct monitoring and alert generating, but now it is just a single tool that everyone can access quickly and easily.

In terms of needed improvements, for machine monitoring, we used to create our own software to conduct monitoring. Previously, this involved modeling in Excel, generating models, and using that software to detect and report. Machine alerts have sometimes happened for a month, which can be critical. With Seeq, we can generate weekly reports for our customers or stakeholders to highlight important notes and necessary actions, especially in terms of maintenance. One quick example would be that we have saved a significant amount of money in machine downtime due to these alerts, and we also extended maintenance periods by monitoring machine conditions and determining safety margins to extend maintenance during critical supply periods effectively.

As a beginner with Seeq, I think one improvement would be to have more simplified formulas. While there are already AI agents to provide guidance, the formulas can still be a bit confusing. Additionally, it would be beneficial to have more training sessions for the company—not just formal certificates, but ongoing certification at different levels to enhance user quality and skills. This is similar to how Aspen software values certification, which can enhance acceptance in the job market.

I have been using Seeq for about six years.

Seeq does experience occasional breaks in data extractions, but I find the service provided is quick enough to cover any downtime. Since we are using an on-premises cloud, the services offered are more direct, which is beneficial. However, improvements in downtime could still be made, although I consider the service sufficient.

Regarding Seeq's scalability, when users and data grow, the stability is not as great as it was initially, which we monitor. Nevertheless, the communication between customers and Seeq is good as they strive to improve stability. I would say it is not as stable as it began, but they are working on enhancements.

The customer support from Seeq is responsive and helpful when I have issues or questions. They provide hotlines for users to contact monthly, as well as urgent lines for quick technical support. The addition of the AI feature for reaching support conveniently is also a great recent enhancement.

Before using Seeq, we previously utilized solutions such as Google Sheets and Excel for data extraction, filtering, and basic modeling associated with control systems. We switched to Seeq because it offered quicker data filtering and generating alerts that added significant value for our team, making it easier to respond to machinery issues. Initially, we were testing the waters with Seeq, but we found it very effective for reducing workload and improving efficiency, which is why we made the switch.

My advice to others looking into using Seeq is that you need at least a basic understanding of how to handle data before diving in. While they provide basic training, I find that the two to three-day training is quite intensive, and without prior data knowledge, it can be challenging. Additionally, understanding the background cloud system, service provider, and contract details is crucial for building the needed infrastructure to connect with Seeq. I would rate my overall experience with Seeq a seven out of ten.