 What I'm going to talk through today is around intelligent solutions with the focus on industry, around the vision versus current reality and really looking at how should we get from the current reality to that vision with a focus on high level decision making and an example I'll talk to is in the scheduling domain. So if we look at factories of the future there's a vision out there of totally automated factory robots doing everything even maintenance on each other and we are progressing towards that vision and factory data systems are getting better, factory automation is getting better but the current situation is we're in a transition phase, we're beginning to transform from human centered activities towards a more automated approach. But how long will this transition phase last? So if we look at how artificial intelligence is progressing, if we go back to 1980s deep blue created by IBM beat the world champion Gary Kasparov in chess. Now at this moment in time I certainly remember people thought artificial intelligence was going to take over the world in a very short time and that hasn't quite happened. If you look earlier this year AlphaGo created by Google DeepMind won a Go match against the world champion. Now Go is a game played with black and white stones, for those who aren't familiar with it and you're kind of trying to control a section of the board. Now it's a much more complicated game than chess, it requires judgment and strategy and there's a lot more potential moves than there is in chess and from an artificial intelligence point of view this wasn't expected just yet, this was probably expected a couple of years out so from an AI point of view this was very good progress. However let me put on another house and say if I was an industry leader, if I'm a factory manager I could look at this progress and say 1987 you won a game of chess, 30 years later you won a more complicated game, it doesn't really take a box for me. So when you look at artificial intelligence solutions today, would a manager of a complex factory give complete control to algorithms? I think we'd all be agreed that the answer to that is no and even people who don't agree on anything else would probably say the answer to that is no. So the question that is algorithms capable of running a complex factory like when do we think that might happen or more importantly how do we think that might happen? So AI at the moment is successful but it's in bounded situations and what I'm really talking about today is how we can begin to use AI for higher level decisions, decisions where you require judgment, you require strategy, allowing for the fact that AI solutions today do not understand a lot of context and I think what we need to recognize also is that it's not going to be one giant step and we're not going to wake up someday and we suddenly have an algorithm that's this good that can control all our factory so we need to look at what are the steps we could do today, how do we start this journey and remembering that the goal is not perfection, the goal is that we are better than we were yesterday. So one potential improvement and I think this was referred to this morning is to treat human operators and automated systems not autonomously but as team members in what's known as a joint cognitive system so what you're trying to do here is move from a situation where you have a human who's utilizing technology but not really collaborating with technology to a situation where you pretty much have technology having a conversation with the human and a really good example of that is how sat nav systems are and we probably don't even think of how sat nav systems work but if you're anything like me you probably use the sat nav to get here today and you took it out of your pockets, you put in where you're going, it gave you a number of options and there was a default option which you probably chose and just yeah I'm grand with that and you selected but you were in control of what you did and after that if there was a crash or anything like there was in the m4 this morning if anyone was using it it gave you an alternative it said maybe you should go this way instead so it started suggesting it's continuously working in the background and suggesting updates for you at any point again you're still in control you can choose a different path you could decide to go a different way you could make a mistake and go a different way but it just keeps telling you from where you are now here's the best thing you can do and the culture around that is really interesting and actually quite important and something we need to get into industry so the culture around a sat nav systems is very much about the human being in control if you remember when sat nav systems came in first there were stories in the newspaper how people drove off a cliff following a sat nav system and no one went well I'm never using sat nav systems again that clearly doesn't work the culture was well what an idiot who'd follow a sat nav system all the way off a cliff that's just crazy so that culture of saying well actually you use an artificial intelligence algorithm it suggests a solution to you but you actually do apply a bit of human cop on at the end of that to make sure you have the right decision so decision making is it perfect can we expect perfect of course not and this is something about expectations especially in industrial solutions there's a feeling sometimes that just because the solution is automated it is perfect which is not true so this is around decision support about augmenting the human not replacing the human about helping them make the right decision and saying that there's a few bots here