|
Intelligence ProgrammingWhat's the next big thing in IT after SOA, Virtualization and Web2.0 ? Intellligence Programming, maybe ? Imagine now everybody has adopted SOA, Web2.0 and Virtualization of everything so computing is green, scalable, agile. Using composite applications, mobile technology and Web2.0 all end users are smoothly integrated in automated end-to-end business processes. Companies have managed to outsource the right business taks to their partners and customers thanks to rich Internet Clients. The IT Center is optimized and is automatically monitored. It can detect problems and repair them automatically : you just need the perform the switch of the redundant broken part of your data center. In other words, imagine companies have achieved the supreme level of maturity : the full agility ! What we than could see is the further consolidation of IT datacenters into MEGA-data-centers wich can run the IT of 100 companies all together. These will monsters, 100 times more powerfull and bigger than Google. Those data-centers will have two different parts. The majority of the datacenter will be ocupied by the mainstream virtual servers : the normal "banking services", "internet blogs"', "interner shops", CRM applications and ERPs. An other part of the data-center will be the sandbox of innovation. A bunch Virtual servers will be left to the good will of some marketeers who will create there the ?YouTube or the SecondLife of the future which will aim at attracting the customers to the good old servers. These servers will be not anymore be SOA like : They will have to adapt to change in real-time. The main characteristic of those servers will be their contained intelligence. Marketing Intelligence Agents will, with the help of the seach and index services, closely watch the competiton. Whe the competitor's information will change, a competing strategy would be elaborated in real-time in order to diminish the first mover's advantage of the other party. This will only be possible if semantic servers are able to make sense of what has changed at the other side. Let's give an example : let's imagine that a site selling cars lauches a new promotion. A competitor's vitual agent is watching this site.Someone designed the virtual agent to systematically make counter offers - it's a volume busines and you want to kille the competition. How does this offer come to the end-user ? The end-user doesn't always watch both sites but has probably subscribed to a car news rss feed somewhere. As soon as the initial announcement is made, the second manufacturer needs to make his counter-offer on that particular feed. Of course if you do that, your business must be aligned. Therefore a new selling process will be instanciated on the fly for the customers who will refer that that counter-offer. This exceptional process must make use of the traditonal services and organization in place, but as the tradional IT is SOA and Agile and if all services contract service are respected, the only specific step to allow this would be to add some go/no-go decision points : Go/?NoGo with the counter-offer publication and new sales process roll-out. As a consequence, on top of those innovative business processes layers, will come an Artificial Intelligence layer. The explosion of competition in a globalized world will make it impossible to allow a human to pilot all decisions on line. Determing the company's strategy and ethics will be the fisrt step of a new kind of programming : decision programming or intelligence programming. This programming will have several aspects. First one needs to model the information needed to take a decision and make sure that those information can be gathered at the right moment. Secondly, one needs to make sense of these data and model decision trees whiche determine in which options are possible for the real-time adaptation of business processes. Thridly, heuristics need to be developped which translate the company's strategy and values. Furthermore, these rules will need to be prioritized such that the system knows which options can be taken in extreme situations and which can NEVER be taken, whatever the situation. Finally, in extreme cases, it will not be able for the machine to determine how the business process should be adapted. In that case, the possible options will be reported to a human and a meeting will be needed to select the right one. This meeting will of course be virtual. The next big thing after SOA is so the Intelligence Programming (Marketing Intelligence, Operational Intelligence, Competitive Intelligence, Strategy and Values Programming): programming the business processes to adapt themselves automatically - for instance by allowing them to increase capacity and remove bottlenecks if the ROI is >1. After SOA the focus will shift from on "How should my business run" to "How should my business adapt to change".
This sounds to me like a Darwinan Evolution of Software and IT... |
