AI Service
AI service to accelerate workflows. As a technical AI service provider, we connect and enrich your systems and employees with relevant, new, explainable knowledge. Suitable for agile organizations in all areas.
Your company's knowledge accessible to all employees in plain language
What if all your employees could understand what's in those tricky contracts?
Or what features are in software developers' code?
What if your complaint management system could automatically check incoming complaints for completeness? Many companies lose a lot of time handling complaints and have to communicate with customers several times before all the information is complete.
Have you already enabled your customers to provide AI-based access to your API? The code they need to implement on their site is already generated and customised.
What if your customers could ask questions about your services or products on your website?
Save on support costs by providing your customers with technical documentation and white papers in simple, clear language?
HighPots offers you the ability to make all information and data from all your systems accessible via AI. We also show you which information you should not make available to any AI system. Or only to certain user groups.
Retrieval Augmented Generation (RAG)
At HighPots, we implement the use cases described above using a method called RAG. Retrieval Augmented Generation (RAG) is an advanced technique in the field of artificial intelligence that enhances the usefulness of Large Language Models (LLMs).
RAG combines the text generation capabilities of LLMs with the ability to retrieve specific information from external data sources and incorporate it into the response generation process. This allows LLMs to draw not only on their trained knowledge, but also on current and specific data.
We represent your organisation’s knowledge in vector spaces/vector databases. You decide where these vector databases are stored. The majority of HighPots customers store these vector databases in private clouds or on their own servers. This is often valuable knowledge (intellectual property) that gives the company a competitive advantage.
HighPots offers other security-relevant AI hybrid solutions. For example, running high-performance LLM AI models on your own servers. Only these have access to the security-relevant data in the vector databases. We connect less critical systems to ChatGPT via API. If the technology companies change their terms and conditions and decide to use your company’s data, your intellectual property is safe.
More than 400,000 internal wiki pages and more than 10,000 whitepaper PDF pages as a database. With a factual accuracy of 83% and a high level of coherence with few logical breaks in longer texts, HighPots’ GenAI hybrid model won us over. The company’s in-house knowledge of vector databases combined with a local Llama model and the API to the ChatGPT cloud. Privacy and intellectual property protection combined with a low error rate were our key requirements.
Mateo Garcia
Team leader Data Science
Astara Madrid
AI for Process Optimization:
Analytical Artificial Intelligence beyond ChatGPT & Co.
With the development of Analytical Artificial Intelligence, we at HighPots conscientiously integrate large amounts of data. Analytical Artificial Intelligence shows exactly what is in your data and provides well-founded predictions and recommendations for action. Although Analytical AI is also based on an artificial neural network, it is possible to understand and explain exactly how the logical conclusions are reached. As a data science service provider, HighPots develops analytical intelligence to identify correlations and optimize processes. As an AI service provider, we often combine the development of analytical AI systems with generative AI, so that HighPots customers can query the knowledge of analytical AI in natural language.
AI Expert Systems
HighPots develops expert systems for diagnosis, decision making and knowledge representation. Thanks to Bayesian statistics, even when little data is available. Especially in the fields of medicine, engineering and finance. AI expert systems developed by HighPots are also used for market analysis. Companies that want to investigate an entry into a market where data is scarce use HighPots AI expert systems.
Discover our Technical AI Services
Our artificial intelligence services include machine learning combined with generative AI services. On your server systems as on-premises AI, in our AI open source cloud, in your cloud instance with third-party providers such as Microsoft Azure, Google Cloud or Amazon AWS Cloud.
KI B2B - Fast Connectivity & AI Training
HighPots offers fast connections to pre-built generative AI models. By “AI-ready” we mean pre-trained and fine-tuned AI systems such as Open AI ChatGPT, Google Gemini or Google Cloud Services Platform (CSP) for on-premises artificial intelligence on your own servers. These systems only need to be trained with your company’s knowledge. HighPots helps you with AI implementation and AI training. This type of AI is often used for chat bots or internal complaint management.
AI Services from HighPots - AI Application Example
Expert system from HighPots for medical imaging processes. Conversion of 2-dimensional images into 3D images.
2D Scan
Digital Twin
On behalf of an international medical equipment manufacturer, we developed an expert system for MRI, CT and X-ray machines. The images from these medical imaging devices are converted from 2D to 3D.
Technical Process
1
Choosing the deep learning architecture: Use of a CNN (Convolutional Neural Network)
2
Selection of the CNN: Based on the parameters network width, network depth and network resolution, we use the EfficientNet CNN.
3
Pre-training: Dataset architecture based on ImageNetvia, creation of a dataset based on images validated by radiologists
4
Fine-tuning with TensorFlow: Testing the dataset with the ImageNet Large Scale Visual Recognition Challenge, then fine-tuning the pre-trained EfficientNet with TensorFlow (first defining the layers that should remain frozen, then defining and creating the new layers for your specific task).
5
Reward modeling and reinforcement learning: creating agents for image classification
6
Result evaluation after the first iteration of reward modeling and reinforcement learning: Results of the agent evaluated by radiologists (positive/negative)
7
Model adaptation and partial retraining
8
Agent adaptation during reward modeling and reinforcement learning: Storage of a more complex scoring scheme (point system)
9
AI model refinement - reinforcement learning: renewed confrontation of results with radiologists
10
Repetition of points 5 to 9: Increase the probability that the results are true; define true (correct) when SixSigma error rate is achieved
11
Development of an interface to the diagnostic software of the device manufacturer
12
Connecting the AI expert system with generative AI; natural language queries and questioning of results by doctors and patients.
Generative AI was unable to achieve our 4-sigma error rate in early AI diagnostics. HighPots was the only service provider with predictive and classification models to achieve an error rate of less than 6.2 errors per million diagnostic cases in our imaging procedures.
Sophie de Jong
Product Development
Philips Healthcare Eindhoven
Learn continuously from data, use results in real-time - machine learning with HighPots
As an artificial intelligence service provider, we expand the knowledge of B2B AI systems in real time. With our machine learning methods, we constantly improve the quality and probability of occurrence of AI systems. We use stochastic methods and sophisticated algorithms for supervised learning.
Transformer models without natural language processing
As an artificial intelligence service provider, we expand the knowledge of B2B AI systems in real time. With our machine learning methods, we constantly improve the quality and probability of occurrence of AI systems. We use stochastic methods and sophisticated algorithms for supervised learning.
Build, train and tune AI
Start generating real, demonstrable business transformations using AI technologies with our AI experts.
Optimize your data for AI
Create a strategy with HighPot's AI consultants to build your optimal data set. Across the entire data science and data analytics lifecycle. Scale your AI workload with us.
Reduce AI risks and act in compliance with the EU GDPR and EU AI Act.
Together with our AI governance specialists, you can control, monitor and maintain AI systems in compliance with the law. Integrate AI responsibly into your overall business operations and combine generative AI, analytical AI and machine learning. HighPots can help you with a wide range of AI use cases.
