Our Impact
Customized AI model based on chemical structure data
Partial automation of chemical structure elucidation
Simplified analytical processes with increased throughput
Mission
NMR spectroscopy (nuclear magnetic resonance spectroscopy) is a method used for analyzing molecular structures. A molecule is placed into a strong magnetic field, causing the atomic nuclei of certain elements (such as hydrogen or carbon) to respond to this magnetic field.
These responses provide insights into the chemical environment of the atoms, from which the structure of the molecule can be derived.
Our client posed the question of whether an AI model could predict the outcomes of NMR spectroscopy for certain chemical structures. This would allow structure elucidation by comparing predictions with actual measurements.
Solution
In close collaboration with the domain experts at BASF, axio concept’s data science specialists developed an AI model in just four months, capable of generating the desired predictions. The PoC created during the project will continue to be refined through our ongoing partnership.
Technologies:
-
Microsoft Azure
-
Microsoft Azure Machine Learning
-
Keras
-
Python
-
rdkit
Highlights
-
Training of a high-performance AI model to analyze molecular structures in .mol files
-
Elastic and scalable cloud-based training using Azure Machine Learning
-
Empirical comparison of multiple neural network architectures to achieve optimal results
Testimonial
The axio team impressed us with their quick comprehension and deep expertise in data science and machine learning.
Within a very short period, they grasped complex chemical contexts and developed a powerful AI model for automated structure elucidation. Their professional and focused collaboration was a key factor for the success of our project.
Dr. Tassilo Grün
Lab Team Leader NMR-Spectroscopy, BASF SE
Facing a Similar Challenge in Your Lab?
The NMR prediction model for BASF is just one example – AI can do far more in analytical chemistry:
-
Automate Spectra AnalysisNMR, IR, GC-MS, Raman – AI accelerates structure elucidation and increases throughput.
-
Generate QC Reports AutomaticallyAI creates time-consuming reports from raw data – including plausibility checks and LIMS integration.
-
Search LIMS & Archives IntelligentlySimilar analyses, historical references, method suggestions – in seconds instead of hours.
Know What's Possible in 2-3 Weeks
Our AI assessment for analytical labs: clarity on your opportunities before you invest.
2-3 Weeks
Compact assessment period
From €3,900
Credited toward implementation
2-3 Hours
Your internal time commitment
What You'll Have After 3 Weeks
-
Spectra analysis: A clear answer on whether and how your NMR, IR, GC-MS, and Raman workflows can be automated
-
Additional AI opportunities: e.g. automated report generation, intelligent archive search, LIMS integration
-
Business case: ROI calculation and effort estimate for each identified use case
-
Decision-ready report: Prioritized recommendations with a concrete implementation roadmap
The Process – Lean and Efficient
1
Week 1
Kick-Off & Data Review
NDA in place, access to your systems, initial data analysis
2
Week 2
AI Analysis & Interviews
Algorithm testing, stakeholder input, process mapping
3
Week 2-3
Report & Decision
Structured findings, clear recommendations, ROI projections
Let’s Discuss Your Opportunities
No obligation—just an honest, realistic view of your organization, your data, and your goals.
