No current, only approach.
KnowFab GmbH i.G. · Services
Industrial AI for engineering, quality, and production
We deliver industrial AI as productized and project-based services for engineering, quality, and production. Available as SaaS, in your cloud, or on premises.
SaaS · Cloud · On premises - explainable, integrable, scalable.
Knowledge graph + neural network
Explainable results
Our difference
Artificial intelligence that understands engineering.
We combine process data, model logic, and domain expertise so that a robust technical assessment emerges from only a few signals. This interplay is our know-how and the difference compared with generic AI approaches.
Interactive visualization
From joining process to a safeguarded quality statement
Using resistance spot welding as an example, we show how a few representative process variables are condensed first, checked against domain knowledge, and then turned into a technical statement.
01 Input
Representative joining process
The visualization shows how a real joining process produces robust input data for the assessment.
01Positioning
Parts and electrodes align.
02Force + current
Electrodes contact, current starts.
03Hold + cool
The nugget stabilizes.
04Open
The joint remains.
Process windowAnimated joining process
F
F
Molten lens
Weld nugget
Upper cap
Lower cap
01A Signals
A few process variables as input
The example highlights current, voltage, and resistance.
Voltage follows the contact.
Resistance follows the contact.
02-04 Flow
Hybrid of model and domain knowledge
The model derives internal signals from a few process variables. These signals are then checked against domain knowledge about material, process, and typical defects.
Activationinternal model signal
Knowledge graphconnected domain knowledge
Plausibility checktechnical consistency check
05 Output
Technical not-ok finding
This example shows a not-ok case with an undersized weld nugget diameter. The output combines quality statement, defect pattern, cause, and corrective action.
not ok
Weld nugget remains below target size
The finding is not ok because the nugget diameter stays too small. Remaining wall thickness and electrode indentation remain within the acceptable range and support the diagnosis of worn electrode caps.
Target at least 4.6 mm
above limit of 1.5 mm
within the window of 0.10 to 0.20 mm
Cross-sectionRepresentative not-ok geometry
The finding is not ok because the nugget remains too small. Remaining wall thickness and indentation depth stay within the acceptable range.
The weld nugget does not reach the required load-bearing size.
The flattened cap contour increases the contact area. As a result, current density drops at the joint and the nugget remains below target size.
The caps should be replaced so that the effective contact area returns to its target range and the required nugget diameter can be achieved again.
Project business
Project business for open industrial processes
Products
Products at a glance
Two standardized product lines for recurring needs in engineering, quality, and production, complemented by custom projects for open use cases.
KnowFab Design
KnowFab Design
Supports engineering and planning tasks with a robust, data-driven basis for decisions.
Fewer manual coordination loops during the development phase.
- Digital support for planning and evaluating production and joining processes
- Uses existing design and process data
- Reduces manual analysis and coordination effort
KnowFab JoinTech
KnowFab JoinTech
Analyzes and monitors live production processes to stabilize quality and throughput.
Detect scrap before it happens.
- Evaluation of process and quality data from manufacturing
- Detection of deviations and anomalies
- Support for root-cause analysis and process stabilization
Project business
Custom Projects
Complements standardized products with robust implementation for open use cases in engineering, quality, and production.
Faster start with direct application relevance.
- Analysis of specific customer requirements and technical constraints
- Implementation of tailored AI solutions with integration focus
- Step-by-step path from pilot and reference building to scalable product logic
Deployment
Deployment models
Criterion
Complete SaaS solutionRecommended start
Deployment in your own cloud
On-premises operation
Operating approach
Full operation through the KnowFab infrastructure and the KnowFab account.
Provisioning in your cloud infrastructure with connection to existing data and security standards.
Provisioning in your own infrastructure for environments with strict requirements around data sovereignty and network separation.
Technical advantage
Fast go-live without providing your own platform
Integration into existing IAM, network, and governance concepts
Full control over data, systems, and access paths
Integration advantage
Operations, updates, and monitoring from one source
Direct connection to existing data pipelines and systems
Suitable for regulated or isolated production areas
Scale / usage
Scalable across sites, lines, and use cases
Flexible for customer-specific integration requirements
Adoption without fundamentally changing existing operating procedures
Process
How we work
From the first alignment meeting to delivery, we work with clear technical criteria. Early in the project, we document transparently what can be implemented robustly and what cannot.
01
Scheduling & target clarification
Joint clarification of the target picture, use case, boundary conditions, and relevant stakeholders.
02
Data review & system picture
Review of data sources, data quality, interfaces, and technical integration feasibility.
Gate 1Data go / no-go
03
Potential analysis & feasibility
Assessment of value potential, technical feasibility, risks, and expected result quality.
Gate 2Business case go / no-go
04
Implementation & integration
Implementation of the solution, system integration, and iterative validation in the target process.
05
Acceptance & delivery
Technical acceptance against defined criteria, handover to operations, and a documented delivery state.
Early transparency in the project
- Clear distinction between feasible and non-feasible requirements
- Early identification of technical risks and data gaps
- Binding decision points before the next expansion stage
Contact & next step
In the initial conversation, we clarify the use case, data situation, and target picture. After that, you receive a clear assessment of feasibility, effort, and the most sensible next steps.
- Technical assessment instead of non-committal general statements
- Early clarity on limits, risks, and prerequisites
- Defined starting point through to robust delivery