We at DHTMLX are excited to continue helping early-stage companies not only with significant discounts within our startup program but also by covering their success stories. Today, we are happy to welcome Jan Koch, the co-founder of Kobra Dataworks, the startup platform that uses DHTMLX Gantt for logistics scheduling. As the company continues to grow, they are actively integrating AI technologies into its solution, with plans to eventually enhance the Gantt chart with smart, AI-driven features. In this conversation, Jan shares how our JavaScript Gantt component fits into their daily workflows and what’s next on their innovation roadmap.
Hello Jan, and thank you for accepting our invitation for the present interview. Could you tell us a bit about your startup, your current goals, and the role of DHTMLX Gantt in your project?
Our startup is called Kobra Dataworks and we are specialized in the transport and logistics industry in Germany. Essentially, we try to use AI to simplify processes for logistics companies. We used DHTMLX Gantt to build the platform, where dispatchers of our customers can monitor trucks and their driving tours along the timeline using the Gantt view. And we’ve been very happy with the Gantt library so far.
What is the role of AI in your application?
We use AI for multiple purposes. Our users receive different types of transport orders as email attachments in PDF or Excel files. We use AI to extract the data on truck tours and put it into a structured format, so we can use it in our dashboard. We use AI for communication in the support chat embedded into the platform, where AI-driven notifications are sent (e.g. whether a driver has sent a proof of delivery document). AI also helps us to understand what the day of a dispatcher looks like i.e. the recurring steps that they have to take every single day and we automate those activities. Our phone AI can give calls to truck drivers or recipients of loads, and organize things on behalf of the dispatcher. The only thing that we are not touching yet is planning out driving tours with AI, because currently it is a very tricky problem to solve.
We also consider possible options for integrating AI into our components, including Gantt, that can be really helpful to users rather than an eye candy.
Actually, right now we are working on one such feature. We think about using AI to adjust the start and end dates of truck travel paths displayed in the Gantt chart based on GPS data. For instance, if a truck has to reach a specific location at 6 PM, and using GPS data of this truck we can update the Gantt chart in real time, providing dispatchers with the relevant information on tour progress. There can be parameters attached to the Gantt taskbar based on certain data that would update the properties of a tour (such as the task completion status). Then, AI would automatically redraw the bar of the chart, making life easier for us.
The other idea that comes to my mind is related to the timescale. Currently, we have the timescale set to various breaking points like daily, 12h, 6h, 1h. The other day, one of our customers asked about the possibility to dynamically resize the timescale. Your library has the features to make this function even more dynamic but we haven’t implemented it yet.
How did you choose DHTMLX Gantt among other competitors and solutions available on the market?
The review process of my team led us to you. We did a brainstorming phase at the end of last year, where we compared a couple of libraries. Then, the dev team started building prototypes using these options. DHTMLX Gantt performed best in this process.
DHTMLX Gantt in Kobra Dataworks platform
Did you face any challenges on the way to implementing DHTMLX Gantt in your project?
Currently, we are tinkering with the alignment thing. It is not urgent, but it is the challenge that we are facing right now. In a traditional Gantt chart, it is totally fine to have overlapping tasks, but in our scenario, it is almost physically impossible to have things overlap. In our case, tasks overlapping mean that a truck reaches the destination point of one tour and the point of departure for another tour at the same time. Currently, we are fine-tuning our approach to redrawing the start and end points of the taskbars, so they do not overlap and go next to each other. This problem may not even be related to the library, but rather refers to the implementation issue on our end.
One more thing is asynchronous loading. I’m pretty sure that our current solution does not make the best use of the library. But not too many customers complained about that, so we didn’t touch that again.
You’ve mentioned that your solution is aimed at German customers. Do you have plans to expand into other countries or markets?
Yes, right now we are focusing on Germany because it is a large market with 12,000 companies involved in the industry. Currently, we are in talks with German billion-dollar companies interested in our services. It is not about this platform as of now, but we are getting our foot in the door with these big companies, and it is plenty for us.
In theory, we could offer this platform globally, but there are several reasons we haven’t done it yet. First of all, we don’t need to, because we are growing in Germany. Secondly, all the data points in our platform come from third-party applications, so we would have to integrate the whole set of new APIs from all the truck vendors, including GPS and driver data services. This step will add a lot of complexity, and we want to make sure that we are ready for it. We need to be 50-60% through the German market before entering other markets.
Is your application a cloud-based solution?
Yes, it is. But we have some customers who work in certain industries with high data privacy standards. For such customers, we offer the on-premise solution.
Are there any features or improvements you’d like to see in Gantt in the future?
I’m not involved in the technical details of the Gantt implementation in our platform, so I can just speak from a high-level business perspective. One of the things that can help make our lives easier is a faster way to add data filters to the Gantt chart. For instance, right now we have a task to add a new date filter to simplify interaction with the timeline.
Which frontend and backend technologies are used in your project together with DHTMLX Gantt?
On the front-end, our application is based on Laravel and Livewire frameworks written in PHP. We also use Python elements for AI components. The back-end leverages MySQL database and Redis cache.
Recently, we’ve been integrating a lot with the AI agent platform called Aide. Also, we use Bunny CDN for storage.
Have you already started earning money from your startup project?
With this platform, we are close to hitting the break-even point. We do have some other services, where we do AI-based process automation that should serve as an entry door to selling this platform (the sales cycle lasts from 6 to 12 months). The other services render revenue to keep the company afloat.
Jan, thank you very much for such a comprehensive interview. We wish you further growth, prosperity, and more innovative ideas for your startup platform.
Visit our customer spotlight gallery to read more great stories about the use of DHTMLX components in real projects.