Artificial Intelligence: Yes! But no Dependencies on Data Scientists!
You can reach a lot with Artificial Intelligence (AI). But which follow up costs will occur to adapt
your AI solution to changed processes?
One thing is clear: Data Scientists are rare and expensive. Therefore, our self-learning AI standard
software PREDICTIVE INTELLIGENCE is automating some time-consuming tasks which Data Scientists execute,
normally.
On the one hand, this reduces implementation costs for your AI solution and on the other hand, also your
operating costs. How does this work? Our self-learning algorithms, which have been awarded many times,
represent the core of our standard software PREDICTIVE INTELLIGENCE.
In this way, your AI solutions get scalable and can be used over years, without hidden follow up costs.
Unsuspected Possibilities with Self-Learning Artificial Intelligence (AI) and Cognitive Robotic Process
Automation (RPA)
On the one hand, nowadays, companies are exposed much more to cost pressure, quality requirements,
unpredictability and increasing process complexity.
On the other hand, companies have more and more data available form their processes, machines and
resource operations.
However, availability of (big) data does not yet realize efficiency increases, hoped for – only smart
data discover inefficiencies and disturbing factors in both, business and technical processes.
Predictive Analytics and Control: Scalable, Self-learning and Un-supervised
PREDICTIVE INTELLIGENCE is an unsupervised self-learning AI standard software. Even in complex processes
and dynamic data structures, you get foresighted recommendations for daily
operations. Machine to machine communication allows direct process and machinery control – with proven
efficiency increase. Dynamic simulation methods discover hidden optimization potentials. Disturbing
factors are discovered early. In this way, you avoid inefficiencies before they occur!
In contrast to other AI approaches (i.e. Deep Learning), PREDICTIVE INTELLIGENCE gives transparency
which factors influence your processes in a positive or negative way. Thus, you can sustainably improve
your processes.
Our solutions, which have been granted with various innovation awards, increase
efficiency of humans, machineries, materials as well as energy.
Predictive Quality (Production) Assess and predict quality of production steps. Discover interwoven disturbing factors and
hidden machine anomalies. Control machinery in a predictive way (machine-to-machine control or as
recommendation for machine operators).
Predictive Maintenance Understand initial signs of future machine failures. Discover reasons for wear and tear.
Recommend action for machine maintenance and machine control.
Predictive Energy (Industry) Predict energy. Discover disturbing factors for energy efficiency. Automate energy trading in a
predictive way.
Predictive Energy Grid Realizing full potential of renewable energy usage. More precise energy purchase and sale, also for renewable energy. Predictively automated energy trading.
Excellent production quality is the figurehead of each production company. Innovative companies which
are ready for the future avoid minor quality and discover disturbing factors on quality or on machinery,
early in advance.
Our self-learning Artificial Intelligence Software Solution PREDICTIVE INTELLIGENCE supports you on all levels of
quality optimization:
Assess and predict quality of production steps
You can skip time consuming quality checks (i.e. ultrasonic). Instead of checking only a fraction of
critical production steps, 100% off all relevant production steps is checked. Manual checks, which
might be required by ISO standards, are targeted to borderline processing.
Result:
Significant cost reduction of quality tests and, at the same time, 100% test coverage.
Discover interwoven disturbing factors and hidden machine anomalies
Our self-learning algorithms discover complex disturbing factors, for example in complex variant
diversity which normally hardly give you a chance to find out reasons for minor quality. Or in
processes, which are exactly the same (same machineries, same calibrations, same supplier material,
…), but still deliver – sometimes – bad quality.
Result:
Sustainable adjustment of production process to avoid low quality.
Predictive control of machinery
Machinery should always run in such a way that it produces high quality. However, batch changes or
high variant diversity often make it difficult to always produce high quality. Our self-learning
algorithms discover complex and interwoven correlations.
Result:
Machine operator is given recommendations for optimal machinery set up. Or it is possible to
automatically control machines.
PREDICTIVE INTELLIGENCE automatically adjusts algorithms to understand changes in machine behavior (i.e.
changed production or machinery utilization). Customized algorithms are automatically rolled out to all
machinery instances.
Below mentioned reference projects prove which value add our self-learning analytics software solution
PREDICTIVE INTELLIGENCE realized:
Bosch
Predictive Quality Analytics in Production
T-Systems
Predictive Quality Control
ZF
Self-learning AI Improves Production Quality in Complex Variant Processing
World-class products require world-class production line management. Increase machinery availability and
optimize technical service management by predictive and demand-oriented maintenance.
Maintaining machinery on the basis of pre-defined cycles and condition monitoring belongs to the past.
Our self-learning Artificial Intelligence Software Solution PREDICTIVE INTELLIGENCE supports you on all levels of
predictive maintenance:
Discover anomalies in machine behavior.
Get transparency which factors influence machine inefficiencies. In this way, you can optimize machine operation
and sensor usage.
Predict accurate machine failure and inefficiencies.
PREDICTIVE INTELLIGENCE automatically adjusts algorithms to understand changes in machine behavior (i.e.
changed production or machinery utilization). Customized algorithms are automatically rolled out to all
machinery instances.
Both, machinery operators and machinery producers benefit from next generation maintenance.
New business models are created to offer value add services for predictive maintenance analytics, i.e.
via cloud.
Below mentioned reference projects prove which value add our self-learning analytics software solution
PREDICTIVE INTELLIGENCE realized:
German Rail
Avoiding total damage thanks to predictive maintenance
T-Systems
Next Generation Maintenance
Siemens
Predictive Analytics Partner for Predictive Services
NTT Facilities
Failure Prediction for Critical Air-conditioning System
Without energy, production stops. Electricity, heat, compressed air, … are significant for production
processes.
For energy intensive industries like steel or aluminum production, energy costs are ca. 25% of entire
production costs. Thus, it is obvious that saving energy costs is mandatory. However, how can it be done
without having negative effects on your production processes and product quality?
Our self-learning Artificial Intelligence Software Solution PREDICTIVE INTELLIGENCE supports you on all levels of
energy optimization:
Predicts highly dynamic energy demands and energy generations.
Analyzes machinery behavior and gives transparency which influences have negative effects on energy
demand.
Gives recommendations for machinery operator how to produce the same product volume and quality with
less energy. Or automatic control is done directly, machine-to-machine.
Gives recommendations for plant operator.
Gives recommendations for energy trading, which products should be sold in which volume and price at
energy trade.
PREDICTIVE INTELLIGENCE automatically adjusts algorithms to understand changes in machine behavior (i.e.
changed production or machinery utilization).
Below mentioned reference projects prove which value add our self-learning analytics software solution
PREDICTIVE INTELLIGENCE realized:
Opel
Self-Learning Predictions for optimizing Power Plant Operation and Energy Trading
Different energy generators and consumers, which are combined to a multi-utility control center, should
simply run in an optimal way. However, what is “optimal”? There are conflicting targets, like increasing
usage of own power supply, reducing energy costs overall, cutting peaks, using machinery more
efficiently, ...
If renewable energy is also generated, complexity is increased. Energy trading can be a lucrative
building block, too, if trade fair risks are reduced.
Our self-learning Artificial Intelligence Software Solution PREDICTIVE INTELLIGENCE supports you on all
levels of predictive energy grid:
Predict energy demand and generation.
Calculate machinery efficiency, and thus, also costs.
Simulate best possible multi-utility control center and control machinery accordingly, either by
recommendations to operator or by automated machinery control.
Realizing full potential of renewable energy usage.
More precise energy purchase and sale, also for renewable energy.
Predictively automated energy trading.
PREDICTIVE INTELLIGENCE automatically adjusts algorithms to understand changes in machine / human behavior.
Customized algorithms are automatically rolled out.
Below mentioned reference projects prove which value add our self-learning analytics software solution
PREDICTIVE INTELLIGENCE realized:
EnBW: Full Potential For Renewable Energy Usage
Innovative analytics approach enables future-oriented solutions for renewable energy
Opel
Self-Learning Predictions for optimizing Power Plant Operation and Energy Trading
German Rail
Energy Peak Cutting
Stadtwerke Saarlouis: Renewable Energy Prediction
Flexible and Precise Predictions despite High Volatility
Digitalization does not stop in front of buildings. Our self-learning Artificial Intelligence Software
Solution PREDICTIVE INTELLIGENCE discovers needs of building users and equipments, like future
maintenance demand for air
conditioning systems. Self-learning algorithms optimize maintenance, energy efficiency, comfort,
security, building usage and quality of living.
Discover behavior patterns
Building operators and users have individual requirements. PREDICTIVE INTELLIGENCE Solution for
Smart Building discovers these behavior patterns, even in highly dynamic environments.
Simulate optimum
All possibilities of available building control are simulated in order to reach optimization goals.
For conflicting gloas, dynamic conflict resolution strategies are applied.
Predictive building control
Best strategies for reaching defined targets are realized. Depending on the individual building,
this can be automated, for example for energy-efficient control of heating/cooling or for optimal
comfort. However, also recommendations are given to building users, such as for predictive
maintenance of machinery or how to use office resources most efficiently.
PREDICTIVE INTELLIGENCE automatically adjusts algorithms to understand changes in building and machine
behavior. Customized algorithms are automatically rolled out to all building and machinery instances.
Below mentioned reference projects prove which value add our self-learning analytics software solution
PREDICTIVE INTELLIGENCE realized:
NTT Facilities
Failure Prediction For Critical Air-conditioning System
Even the best products or services are successful only if they meet your clients' demand. Therefore, it is essential
to understand your clients.
Here, too, self-learning PREDICTIVE INTELLIGENCE can help because more and more sales-related data is available.
Self-learning algorithms learn your clients' behavior and predict with high accuracy, for example, how your client
will behave in the future.
This enables you to adjust to individual client wishes, although your clients have not mentioned them. In addition,
it is possible, to partly automate customer sales care, for example in placing client-specific advertisements in
online shops.
Our self-learning Artificial Intelligence Software Solution PREDICTIVE INTELLIGENCE supports you on all levels of
Predictive Sales:
Predict product sales.
Discover influences on selling.
Enable client-specific advertising.
Reduce return shipment.
Recommend optimal prices.
PREDICTIVE INTELLIGENCE algorithms understand behavior changes and algorithms adapt automatically.
Below mentioned reference projects prove which value add our self-learning analytics software solution
PREDICTIVE INTELLIGENCE realized:
Big Data turns into Smart Data if you analyze large data streams for highly complex data patterns in
real time. In this way, you can avoid - future – inefficiencies.
PREDICITVE INTELLIGENCE analyzes extreme data masses which are
delivered, for example, by machinery sensors. You get results at once, without time delay, even for
complex
analytics. Our solutions are specialized for highly-complex and fast calculation methods.
5,000 billion gigabyte of data was produced in 2003, worldwide. Only 8 years later, the same data volume
was
created in just 2 days. Today, it takes around 7 minutes to gather the same amount of data. Data – on
its
own – does not deliver value add. Intelligent analytics optimizes processes and cuts off inefficiencies
before they occur.
However, it is important not to lose time with data management and analyzing. In many processes, time is
the
critical success factor, i.e. in machinery operation, in service level monitoring or tracking of
logistics
processes. Big data streaming is analyzed according to required rules and anomalies or (future)
inefficiencies are discovered, at hoc. You are informed in real time, or your operative control is
improved, automatically and at once.
You can define your own rules in a flexible way, or you can use self-learning algorithms to discover
hidden
rules and generate them automatically.
Below mentioned reference projects prove which value add our self-learning analytics software solution
PREDICTIVE INTELLIGENCE realized:
Our standard software PREDICTIVE INTELLIGENCE with its self-learning Artificial Intelligence algorithms
delivers highly accurate predictions even in complex processes and discovers hidden anomalies in machine
and human behavior.
In addition, recommendations are given to avoid future inefficiencies. These recommendations are given
to humans, i.e. machine operators in production, or control is done without human interaction, thanks to
machine-to-machine communication.
Depending on your analytical task, relevant modules are made available.
USP: When your processes change, our software understands future consequences of those changes,
automatically. Therefore, no Data Scientist has to modify your Artificial Intelligence solution,
anymore.
Analytics Discovery
Initial discovery of complex patterns
Continuous learning of changes
Automatic adjustment for instances
Analysis
Discovery of complex root causes
Transparency how to improve processes sustainably
Prediction
Highly accurate prediction of complex KPIs
Reliable planning of even strongly interwoven processes
Avoiding inefficiencies before they happen
Anomaly Detection
Discover anomalies in processes or machine behavior
Clearly assess process anomalies and those anomalies which lead to bad performance, failure,
...
Simulation
Assess changed processes before implementing organizational or technical changes
Execute What If Scenarios and find best solution for future process execution
Predictive Control
Integrating optimization into operative control
Automatic and continuous optimization of operative processing
Self-Learning Predictions for optimizing Power Plant Operation and Energy Trading
ZF
Self-learning AI Improves Production Quality in Complex Variant Processing
Police
Discovering communication patterns in organized crime
EVO: Machinery Control / Simulation
Heat Recovery saves 25 % of Energy
NTT Facilities
Failure Prediction for Critical Air-conditioning System
LG: Predictive Heating/Cooling
Adaptive Hotel Solution to Minimize Energy Costs
RES-COM
Research project Industry 4.0
Dillinger Hütte
Energy Demand Analysis
Koehler Paper Group
Predictive Quality Analytics in Production
Stadtwerke Saarlouis: Renewable Energy Prediction
Flexible and Precise Predictions despite High Volatility
Bartz Werke
Reduction of Energy Resources in Casting
Bundesministerium für Bildung und Forschung
Research project Industry 4.0
Saarstahl
Logistics Optimization
Netzwerke Saarlouis: Gas Prediction
Twice as Precise than State of the Art Solution
Wittenstein
Predictive Maintenance
DFKI
Research project Industry 4.0
Hager Group
Predictive Building Control
Mobil saar
Research project Mobility 4.0
CyProS
Research project Industry 4.0
bpE
Process Analytics
Guided AB
Predictive and self-learning house automation
Bundesministerium für Verkehr, Bau und Stadtentwicklung
Research project Industry 4.0
Vicar
Customer Analytics and Recommendations
Thanks to self-learning Artificial Intelligence, our standard software has proven itself in different
industrial branches. On the one hand, it improves our clients operative results. On the other hand,
costs over years are reduced significantly, because necessary modifications – caused by changed
processes – are understood automatically by Self Learning Algorithms. Thus, there is no need for a Data
Scientist to adapt the solution again and again.
Although IS Predict is still a young company, its software solutions have been awarded with several
innovation
awards. This acknowledges our solutions' innovative character.
PREDICTIVE INTELLIGENCE is an unsupervised self-learning AI standard software. Even in complex processes
and dynamic data structures, you get foresighted recommendations for daily
operations. Machine to machine communication allows direct process and machinery control – with proven
efficiency increase. Dynamic simulation methods discover hidden optimization potentials. Disturbing
factors are discovered early. In this way, you avoid inefficiencies before they occur!
