With respect to the ongoing trend of AI, robotics, deep learning, machine learning and a number of data analysis kinds. The key enabler of this industry is data, information. An era of technologies to be part of the future, already a multi billion industry. Referenced to this, one often hear “information is the new fuel”. So, therefore let us stop there a second.
What is fuel to you?
Think of it as a methapor.
Then think of what fuel is, what function it have, in a generic perspective.
What can the cost be not to challenge the current association to fuel? If the relation and understanding of information and data is not very clear, how can one ensure sustainable development of the new technologies? I wan’t to challenge this in the higher view, and explain why I want you to care about it.
For instance: A composition of atoms that you put into combustion. After combustion we have a new composition of atoms, something else. A catalythic process. Or so. Irreversibly blended with other similar kind of atoms. Unable to trace back to it’s source or original structure. The cost of fuel is producing, not by using it. Once fuel is there. it’s ready to serve. Does this fit into the description of data to IT? If yes, might you violate architectural foundational principles by neglect traceability?
For completeness, I choose to introduce with IoT. The reason is that IoT matter as an newcomer as enormous producer of data. But do IoT produce data? Well, yeah! To a cost? Maybe not obviously in first though, but it’s quite easy to associate a cost immediately to IoT. Hardware to produce is for me, cost. The engineering of IoT hardware is in infancy. Also to mention the feasibility of the implementation phase. Of course I include the security aspect in that comment. IoT is meaningless and belong to kids corner until it ensure security and information receipt acknowledgement. Where is the business value here?
Machine Learning and Deep Learning is definitively about data. But does it produce new data through cost, such as IoT? Not directly, one can say. This may be seen as the instrument used to transform fuel to power. As a technology. Still, building and maintain it is connected directly to cost. The algorithms and analysis of information or data to produce new data by creation of scenario figures. Figures that come with human effort, training and evaluation. And on top of that, one need data quality recognition and classification. Every time you end up in a change of an algorithm, the old data might be useless and need to be recreated. Should I care to mention cost of power during computing? Where is the business value here?
Taking persistence into account. DW and Big Data plus dozen of “localized” technologies for the structured data, in addition to the physical storage to persist data. As it is for IoT, all might agree that this is engineering. Obviously not about fuel! But you might be tempted to associate information to the fuel persisted on it, as gasoline in a gas tank. But wait. This storage mechanism is there to keep the power until it’s released. Such as a battery or transmission/gearbox. To the cost of this objective; think of the mandatory methods and styles to agree on transportation, data format, availability and quality selection. This breeaathing red colored dollars sign. Costs just to produce availability/usefullnes of data for decision making. Where is the business value here?
To this picture, add information. So, now most of you might disagree or say that of course there is business value behind all this. I bet you are right. Information is common in all areas, but is it fuel in any of them? or are they just supporting or consuming functions where information is a part?
Is automation or AI about replace humans with machines or services? Automate services without direct human actions. Invent services in areas where human can’t or won’t as of today. Oh yeah, here we will find increase of income or increase of defense. Or increase of service level. Reduced cost. Increased ROI. Happy business user and the CFO that love to invest and see technology help to create value. The key is automation. More about that soon. Now time for a mandatory parable.
For more than hundred years ago, automation could be to put cogwheel between a winch and a arm to rotate. This to make spans of iron ore transport from a hill to ground. A person need to rotate this. One may think that it was a horse or two, but a horse must be managed by men. And it take a one to one relationship between a human work hour and efficiency. One day, one attached a rotating axis from a completely different invention domain, steam engine, to make the rope winch rotate. Suddenly iron ore could be transported 24h a day. Never get tired, always transport, regardless of time to deliver. It now took one person 5min one time an hour to put coal into steam engine owen. As of a sudden, little later, a good income could be doubled by install another line of iron ore transport, with same purpose. Eventually little faster, more reliable and need less coal to run. Invention, optimization, cost reduce.
Is this an obvious business case? Similarities to what we do with IT? Yeaah, little. And this is finally the point for rest of the article. Robotic, Machine Learning, deep learning is all about IT. IT is mechanic, strategy, methods to produce, use and re-use the information or data. Purpose? convert to power, together with fuel. The fuel itself is not in any of the earlier metaphors. In the parable above, the very transport (movement = service) of iron ore could be seen as the product of all mechanical composition. Compare this to IT deliver a web-based renting service. Fuel would continue be the business decisions and ideas, that drive the innovation and invention of technology and usage of information, converted/combusted into power that push energy into the transmission and gearbox.
The questions I will let you take with you is (from IT perspective); Are we those who should define and invent AI or automation? To which cost or increase revenue do we replace or incorporate automation that make sense? That is a business question. IT now sit on extremely powerful platform of technology. Maybe you can’t have control over the effects, when apply it in a greyzone where business decisions is not present. Think intentionally, IT is here to deliver effecient technology strategy to business decisions. While business is here to provide adequate services to customers, whatever their usage is. So what I can see, Master of Business Administration is there for business development and interface to the customers. Right? It’s may be tempting that IT might take share of the market to increase the “IT drive the business” perspective. The deeper technical knowledge of the tools can make IT advocate for business, what capabilities they should design services for.
It didn’t work well at dotcom-bubble at about year 2000. I assume that the finance and IT around the world is more mature and can’t repeat the dotcom-bubble like it was, but i’m certain that we can (by greed or by accident) repeat troublesome and time consuming decision patterns and styles where IT is used to meet business. Eventually because we are new to read the volatile consumer market as of today. Remember the intention of the technology strategy. I trust business to make decisions and requirements automation and AI, but IT to provide the technology strategy. With this said, this is an interpretation of automation. One can say it’s wrong, one can agree, one can choose both or none of it. Reflect, analyse and comment! Thanks.
Bottom note: If you like more peoples to read this, just press a like or a comment. For LinkedIn it’s enough for spreading the word, in contrast to many other networks that require sharing.
View Jonas Nordin’s professional profile on LinkedIn. LinkedIn is the world’s largest business network, helping professionals like Jonas Nordin discover inside connections to recommended job candidates, industry experts, and business partners.
PS: Interesting analysis of emerging technologies, with respect to this article. The scope in the link would be far outside the year 2017, but indeed this is the year to start watch out for it. Also as you understand, I still see AI as the services to business and end users, not a technology by itself.. =) http://cognitiveworld.com/article/emerging-technologies-watch-2017 DS.