The automation of any processes can be done in two ways

All single steps can be automated in only a structural way.

The whole process can be automated by the use of AI.

For example: To reduce the workload in the recruiting process, a company installs an online system for applications. The system itself has an automated routine that separates the application by quality using information provided by the candidates in the application form.

If the automation was done without using an AI, it is fast and easy to integrate and quite cheap most of the time. On the other hand, if there is just a small change in the process or a candidate uses unusual data, the automation doesn’t work anymore and it needs time and invest to fix the automation. This is necessary with each change or problem in the process.

By using an AI, the starting process needs preparation and planning. However, as soon as the AI is integrated and ready to work, the processes work automatically and even changes or problems are not an issue. There no longer is a distraction for your HR experts, thus they can concentrate on tasks machines cannot fulfill. Ideally, learned processes or parts of it can be reused in other processes by the AI.



Machine Reasoning is a principle used by machines for problem solving through drawing conclusions from previously learned knowledge applied to new environments with logical techniques like deduction and induction. The system has to be taught (you will learn more about that in the part on Knowledge Items) and the taught knowledge has to be stored. Therefore, a semantic map is created that a process engine relies on for understanding the world through semantic reasoning.

Machine reasoning has the ability to dynamically react to change and by doing this, reusing existing knowledge for new and unknown problems. With machine reasoning, problems are solved in ambiguous and changing environments. The AI dynamically reacts to the ever-changing context, selecting the best course of action. Thus, machine reasoning is the basis for a general artificial intelligence (General AI).


Knowledge is taught in atomic pieces of information that represent individual steps of a process in so-called Knowledge Items (KIs) including what, when, where and why of each step.

They are reusable and flexible, so instead of writing vast amounts of scripts or runbooks, you create and use a comparably small amount of KIs from which the AI can deduct a huge numbers of recombined solutions for occurring problems. That means that the AI reaches a high automation rate with little mentoring by humans.

Exponential Teaching

Exponential Teaching describes the human work with HIRO™, teaching KIs to HIRO™, enabling the engine to solve problems autonomously. The KI pool grows approximately linear while the recombined solutions grow exponentially.

Follow the Rules of Exponential Teaching and you can make the best from HIRO™ and its taught knowledge:

  • Write KIs incrementally, atomic and reusable
  • Use Task to Knowledge to organize your knowledge
  • Separate decisions from actions.
  • Use as many existing KIs and variables as possible.
  • Avoid chaining of KIs and use minimal binding and trigger conditions.
  • Store output properly and ensure good error handling.

A real life example

How does Machine Reasoning change the way the IT department works?

Back in the days, long before HIRO™ was introduced in Tom’s company, his working routine was basically a real routine consisting in completing repetitive tasks. When a ticket came in regarding a Webserver not responding, Tom had to single-handedly solve the problem:

... request restart of service ...

... request configure network ...

... configuration on linux ...

... request restart of service ...

... ticket update: Webserver responds ...

Now, things have changed: It’s Toms first day with his new AI-colleague HIRO™ and, again, the Webserver is not responding. The ticket comes in and Tom solves it. Business as usual, but since he has a new colleague he wants to share his knowledge with him. Tom transfers his knowledge into KIs HIRO™ can use. Teaching HIRO™ enables the engine to solve tasks on its own.

The Webserver is down again! So far, that was bad luck for Tom. But now that he has trained HIRO™, things are different: The ticket comes in and HIRO™ can solve the problem on its own.

One year later: Tom has taught HIRO™ 156 KIs in 90 teaching sessions. HIRO™ has solved 21.234 tickets with 1.791 recombined solutions. And Tom has time to think about the use of HIRO™ in other parts of his company. (The figures are from an actual case.)