ConSol

Pflegeausbildungsfonds Bayern GmbH is the designated authority appointed by the Free State of Bavaria to finance nursing education in Bavaria. In simple terms, we determine the funding requirements, collect the corresponding levy amounts, and manage these funds in the clearing fund. We then make the предусмотренные equalization grants to the training facilities and nursing schools.

I’m personally the Deputy Managing Director, and in addition to my overarching responsibilities I also oversee strategic and organizational projects—especially those focused on further developing and digitizing our processes.

The project with ConSol essentially consisted of two building blocks.

The first was an AI-based chatbot for our website. It was designed to draw on the content from our knowledge portal and support users with their questions. A particular challenge for us is that funding for nursing training is a very specific—and in some cases complex—subject area with many technical and legal nuances. That’s why it was so important that the AI understands these relationships and provides as precise answers as possible.

The second building block was an internal AI solution in a secure, data-protection-compliant environment. On the one hand, it was meant to serve our employees as an intelligent reference tool, and on the other, to support concrete internal work processes.

One example is maintaining our knowledge portal. Together with ConSol, we developed an AI-powered wiki editor that can compress, summarize, and prepare extensive content for use in our portal. In addition, we use the solution, for example, to create process descriptions or to help draft suggested replies for emails.

For the project, we ran a selection process with multiple providers. ConSol stood out for us very early on. The proposal was prepared quickly, and the content was communicated clearly and specifically.

From the start, we had the strong impression that our requirements were truly understood—and that ConSol recognized what mattered most for our particular use case.

After that, we held an requirements workshop together. Building on that, ConSol developed a well-structured and easy-to-follow presentation of the proposal. Ultimately, it was the complete package—understanding of our requirements, subject-matter expertise, and a clear approach—that convinced us.

After the project was commissioned, we moved into the concrete implementation very quickly. We started with a kickoff session where we aligned on the timeline, the individual work packages, and the next steps. After that, we held fixed project meetings every two weeks.

On our side, our IT Manager and I were mainly involved. Depending on the topic, we also brought in colleagues from the relevant business areas.

At ConSol, we had clear points of contact, which made communication very straightforward. We were regularly updated on the current development status, and we could see early progress almost immediately. After about four weeks, we already had a first test version available.

During the testing phase, for example, we identified that we needed to adjust certain content and structures in our knowledge portal so the AI could work with it even better. That was an important learning process, and ConSol supported us closely throughout.

The internal AI solution also went through a test phase with our employees. Our colleagues could try the application in their day-to-day work and provide feedback right away. We then reviewed this feedback together and incorporated it into the next stages of development.

Another strong positive was budget adherence. We completed the entire project within the planned budget. In fact, there was even some budget left over, which we used right away for initial discussions about potential follow-up projects.

On ConSol’s side, we essentially had two dedicated points of contact.

One developer supported the technical implementation continuously. In addition, there was a fixed point of contact who handled overall coordination and the communication between us and the development team.

Especially at the start of the project, another developer was brought in from time to time so that certain topics could be handled in parallel and implemented faster. Once those tasks were completed, the team was scaled back accordingly.

Throughout the entire project, communication with our contacts was excellent. We always knew who to reach out to with questions—whether they were technical or subject-matter related.

We’ve had very positive experiences both with internal use and with the external chatbot.

Internally, we especially get positive feedback from colleagues who work with the AI regularly—for example, using it to prepare content for our knowledge portal. In particular, when it comes to structuring, summarizing, and organizing large amounts of information, the solution provides a real advantage.

In addition, throughout the project we learned a lot about how to use AI in a meaningful way. ConSol didn’t just deliver a technical solution—they repeatedly explained what’s happening behind the scenes, what options are available, and what we need to pay attention to when using it.

The external chatbot has also met our expectations. For us, it was definitely an eye-opener to see how well the solution already handles our very specific, sometimes complex, subject-matter content and questions. From our perspective, the original project goal was fully achieved.

The collaboration was very reliable in this regard. Questions and feedback were typically addressed the same day.

The project also went very well on the timeline. The agreed milestones were met, and the overall project was even completed earlier than originally planned.

What was especially helpful was that feedback wasn’t just collected—it was quickly turned into concrete adjustments. That way, we always had the feeling that the project was continuously evolving.

What stood out to me most, though, was the human side of the collaboration. From the very beginning, we got along really well with ConSol’s team, and we had an open, straightforward exchange.

Just as important for me was the way communication was handled. I’m not an IT specialist myself. Even so, ConSol managed to explain technical relationships and the progress of the project in such a way that they were easy to follow—even without a deep technical background.

Especially in an AI project, where some questions are quite technical, this skill is, in my view, absolutely crucial.

Honestly, nothing substantial comes to mind here. We were very happy with both the technical and subject-matter implementation, as well as the communication and our collaboration.

ConSol was also highly flexible on organizational and contractual matters and met us halfway. That allowed us to make the best possible use of the existing budget and even tackle a few additional topics—and potential follow-up projects—together right away.

From our perspective, there’s therefore no specific point we’d highlight as something that needs improvement based on this project.

Bauer Sanierungen is a hands-on trades business with more than 30 employees, specializing in the restoration of fire, water, and mold damage. In addition, we also handle services in plumbing, heating, technical drying, and painting. At Bauer Sanierungen, I’m responsible for the back office as well as coordination, and in the AI project I served as the central point of contact and decision-maker on our side.

Our core goal was a secure, structured, and consistent way for all employees to access generative AI—rather than the previous ad hoc, inconsistent, and potentially risky use (shadow AI). We also wanted to significantly speed up the very time-consuming creation of statements and reports. Another absolute requirement was a privacy-compliant setup with local data storage and European AI hosting options.

What mattered most was that ConSol took our requirements from the very beginning: an on-premise AI platform on our own servers, based on the open-source solution Open WebUI, with a GDPR-compliant connection to the European AI provider Mistral. Beyond that, we were convinced by the licensing model without user-based fees, which brings cost transparency, as well as the flexible architecture that allows a later switch or the parallel use of additional AI providers.

The approach was clearly structured: In the first phase, we worked with you to prioritize specific use cases—leak detection reports, statements, knowledge bots for property management and devices, as well as the digital accessibility of training materials. In the second step, the platform was set up on our own Ubuntu server, Mistral was connected as the AI provider, and the first assistants were developed. In parallel, we built a granular role- and permissions model for a secure, access-controlled deployment for all employees. The project was rounded out with a testing phase, training, and a gradual rollout to the team.

On our side, only a few people were involved: me, as the coordinating point of contact, along with subject-matter experts from the team who defined the use cases and reviewed the results, as well as our IT team for server operations and access. On the vendor side, one person worked on the project (project management, development, and integration).

Yes. Today, our employees use a central, secure access point to AI-assisted work, which has significantly reduced turnaround times for recurring tasks like statements. The privacy-compliant setup on our own servers with European providers gives us legal certainty, and avoiding user-based licenses improves cost transparency. With its open architecture, we also keep the option to switch AI providers or run multiple in parallel. Additional enhancements are already planned as well—such as integrating our internal planning tool and our ERP systems to embed the AI platform even more deeply into our workflows.

Yes, the agreed milestones were met. Our feedback was taken on board promptly and in a clear, understandable way: adjustments were implemented right away, and communication about the project status was clear and transparent at all times.

What impressed us most was how straightforwardly the provider implemented our demanding privacy requirements—own servers, open-source foundations, and European providers—without compromising any functionality. They also took a very practical approach: as a skilled trades business, we weren’t burdened with technology for its own sake; instead, the AI tools were tailored to our real working processes.

Not much to complain about—our collaboration was consistently professional. If I could suggest one improvement, it would be a more thorough onboarding, along with a central collection of templates and common prompts for the key use cases in mid-sized businesses. That would have really helped us.

I’m the Managing Director of Bauer Sanierungen. We repair fire, water, and mold damage, and we also handle services related to drying, plumbing, heating, as well as painting and coating work.

As managing director, I make sure our job sites run smoothly, our teams are deployed effectively, and we work efficiently and economically.

We ran into a real problem with scheduling. Our drying and trades teams were planned mostly manually, sometimes across split tables. It was hard to keep track of and kept leading to unnecessary trips and extra coordination effort.

We also looked at standard solutions on the market. The issue was that none of them really understood—or captured—how our day-to-day operations work. On paper, some things looked like a match, but in practice key workflows, dependencies, and special requirements were missing.

So it was clear that we’d need our own solution. In the past, this kind of custom development would likely have been too expensive and too time-consuming for us. ConSol showed us that with AI support, you can develop tailored software much faster and at significantly lower cost today. That’s what made a custom solution realistic for us in the first place.

Our goal was a simple, well-fitting tool for our dispatch team. It should help us stay on top of things, plan the teams more effectively, and avoid empty drives.

ConSol quickly understood what we’re about. We didn’t need some big, off-the-shelf software package—we needed a solution that fits our day-to-day reality.

What mattered most to us was that ConSol took our actual workflows seriously. This wasn’t about forcing our operations into some ready-made system. It was about building a tool that supports our planning in the way our real business requires.

We were especially impressed by how much faster and more cost-effective it was to develop the software with the help of AI than before. For us as a trades business, that was a key point: we didn’t want years of an “IT construction site” or a massive budget—we wanted a practical solution that pays off economically. ConSol made exactly that possible.

First, we reviewed together how our planning was going so far and where the problems were. It wasn’t just about creating polished process diagrams—it was about real day-to-day work: which teams are available, which assignments depend on each other, where empty runs happen, and where we’re losing time in the coordination.

After that, we developed a planning and dispatching tool step by step, tailored to our workflows. ConSol used modern AI tools in the software development process, which helped us get visible results much faster than we would have expected from traditional software development.

We were also able to give feedback whenever something still didn’t quite fit. Those points were picked up right away and improved.

On our site, the management team and the responsible people from planning were involved. I personally was mainly there to make sure the solution makes sense for our operation.

A technical contact and developer from ConSol were also involved.

We have a much clearer picture of the planning today. It’s easier to coordinate the routes, and we were able to cut down on empty trips.

That saves time, reduces stress, and lowers costs. What mattered most to us was that the tool fits our way of working and isn’t just an off-the-shelf solution. With standard software, that wasn’t possible before.

Another important point is that we’ve seen that custom software doesn’t have to be automatically prohibitively expensive or painfully time-consuming anymore.

The collaboration was straightforward. Our feedback was taken seriously and incorporated quickly.

What stood out most was that our real day-to-day business requirements were at the center of it. That’s exactly where standard solutions had failed before.

But what surprised me most was how quickly, with AI support, individual software can be created today. In the past, for a custom solution I would’ve immediately thought of high costs, long timelines, and a lot of coordination. This project showed that things can be different now—if the service provider brings the right tools and the necessary know-how.

I can’t think of much more. Maybe we could define upfront a bit more precisely what success we want to measure later—for example, fewer empty trips or less time spent on planning.

Overall, we’re very happy with the result and the collaboration.

Consult for Rare supports companies in the pharma and healthcare space with strategic questions related to rare diseases. As Managing Director and consultant, I’m responsible for client projects, business development, and the day-to-day management of the company.

I wanted to use modern AI tools in a targeted way for my day-to-day consulting work. My focus was on researching and organizing complex information more efficiently, creating presentations and concepts, and streamlining administrative tasks.

What made the difference was the hands-on focus and the ability to translate the potential of AI into concrete use cases in my day-to-day work. From the start, the program centered on the real challenges I face in consulting and management—not generic training content.

At the beginning, my existing workflows were analyzed and suitable areas for AI use were identified. After that, specific use cases were developed and implemented together in multiple coaching and sparring sessions. Over the following period, the AI use was continuously expanded and adapted to new requirements. The collaboration is ongoing to this day.

The collaboration took place directly between me, as the managing director, and the responsible AI consultant.

KI is now a permanent part of my daily work. In particular, when researching, structuring complex content, creating presentations, and handling administrative tasks, I can clearly save time. At the same time, the quality and consistency of many of my work outputs have improved.

The collaboration was consistently reliable and flexible. Feedback was taken on right away and integrated into further guidance.

What stood out as especially positive was the ability to translate technological possibilities into concrete, immediately usable ways of working. Instead of theoretical discussions, practical solutions were always at the center.

The collaboration was extremely successful. As the AI landscape continues to develop, new opportunities are constantly emerging—opportunities we can evaluate and explore together.