The informed organisation

..or how super-human intelligence is being achieved today

Originally Published on July 6, 2018 on LinkedIn

Today, we live in a digitalized world. Digital transformation has reached into every aspect of our lives, definitely into every aspect of our businesses.This means that every activity we conduct in business is aided by computer-systems, and the computer-systems are becoming ever more capable and to some extent "intelligent".

How fantastic would it be if computers could take over all the errands we have to run every day and leave us to only pursue what attracts our interest? I leave it to others to answer if this is a utopian or a dystopian vision. In the meantime however there remains an ever-growing amount of human-intelligence work to be done.

Computers these days are exceptionally good at very distinct, precisely specified tasks, surpassing human capabilities by far in their (narrow) areas of expertise. While computers keep constantly improving their capabilities to solve gradually wider-scoped problems, for the next couple of decades at least it remains human duty to perform the

General-intelligence tasks:

  1. categorize any upcoming task or problem into the right problem domain so to pass it into the respective business-subsystem. Is it an exception/bug/error/problem or is it a "normal" instance of a (potentially desirable) event that was planned for? Does it belong into e.g. supply chain, hr, production, service, sales, r&d, or "strategy" (e.g. everything else)?
  2. translate information into the form that the respective subsystem requires to process it correctly
  3. gather and aggregate the systems output into a form that is then typically delivered to another human as a unit of value

The overall system performance is determined by a couple of factors

  1. How good is the quality of the first decision into what category a "thing" belongs? E.g. how many instances get classified correctly / incorrectly as a percentage? How well does the modularization of the business-systems reflect the stream of incoming events? How well are the decision-makers equipped (in terms of training/education combined with information availability and quality) to make quick and correct decisions?
  2. How much effort is required to key in the information? And again how high is the error rate?
  3. How much additional effort is required after system-processing to deliver the value unit? And how good is the quality / how high is the delivered value?

A good enterprise information management

..should start with developing the above metrics and subsequentially delivering them in an actionable form.

  1. The metrics need to be as simple as possible and easy to understand.
  2. It needs to be clearly visible from which raw sensor-data they are derived and how
  3. Otherwise human beings will not trust them and will not base their decisions on them, rendering them useless.

Subsequently the system should provide context and trend-data on the metrics support leaders in deciding where to allocate effort and investments

  1. Do the people who make the first contact have the appropriate training / skillset and tools to classify correctly? Has the organisation an efficient routing-process? How often are things re-routed before they are being processed the first time? Does the system give them sufficient information to make an informed decision or do they need to do guesswork / trial and error until something "sticks"?
  2. How much labour is required to complete a certain amount of task? Is the rate increasing or decreasing? Is the work being processed in a timely fashion or are long queues the norm? Is the workforce underutilized? How much does it cost in time and money to improve the computer-system's performance vs. how much would it cost to hire additional labour? Is the scalability sufficient to beat the market development? Is the error-rate increasing or decreasing and what are the main root causes (80/20 rule)?
  3. How good is the quality of the computer-system's output and what would it require to improve the delivered value?

Human- and computer-intelligence can form a powerful collective "super-intelligence" today by means of a good communication infrastructure. Individual human intelligence still outperforms machine intelligence when it comes to general reasoning and problem solving. However we all know that 10 managers, deciding together are not 10 times more intelligent than one manager deciding alone. Personal interests, principal/agent problems, differing world views, norms and values, communication-errors and a whole host of other problems limit drastically our capability to co-operate. However - in our digitalized world, AI or computer systems in general have matured. They can act as very effective information brokers, that solve or mitigate most of those problems, thereby giving an order of magnitude boost to the overall organizations' intelligence. And this is not science fiction, but the reality of how super-human intelligence is being achieved today.