Andrew Anagnost, President and CEO, Autodesk talks about how cloud, machine learning, artificial intelligence and generative design are going to create a whole new world of opportunities in the next decade. Mayank Singh reports fromAutodesk University 2017, Las Vegas
The business model transition to software as a solution (SaaS) at Autodesk took place under its former President and CEO, Carl Bass. What kind of a business model transition would you like to lead at Autodesk?
Actually, I led the business model transition; it was mostly my work while working for Carl. He was actively involved with other kinds of transitions in terms of driving the company onto Cloud. The transitions that I am focused on right now is completing the subscription transition and by that I don’t mean completing it only for the company but also for the customer so that he understands the benefits of subscription. We expect to complete it over the next year and a half. The other focus is digitising Autodesk itself, as it will make us more responsive to customer needs. Right now it is too complex for customers to manage their relationship with Autodesk and it needs to be much more seamless. A customer needs to be able to do 80 per cent of the basic mechanics of managing their relationship with Autodesk digitally without having to talk to a human being.
Going forward, you will see us reimagining the downstream processes which follow the design processes like construction, manufacturing, video and game production. These are things that are becoming more intimately connected to the design process. There is a conversion of design and manufacturing going on and some of the automation and key technologies that we are using to create models based on the constraints of how they get made are changing.
Over my tenure of ten years, the world of design and make are going to be very different. We are really going to be in a push button world where people are designing and it can immediately be translated into instructions and capabilities for building something as the design would have already captured all the requirements for building something. It’s going to be a new world in ten years.
Emerging markets are being talked about as the next big market for software service. Is Autodesk creating specific programmes or working with governments in these markets to diffuse technology?
We are involved with emerging economies not only through our business, but also work with our partners to support developing countries. We are helping these countries with technological development and technical skill enhancement. Emerging markets are not ranked equally, in some of these countries economic opportunities have not been diffused completely but they are pretty mature markets. For instance, we look at Africa and think that it is going to be huge someday and we are trying to figure out what can be done to actualise that vision. One of the things that is true of Africa and a number of emerging markets is that they are skipping the wired infrastructure phase as it is unaffordable for them. As a result they are going to be at the cutting edge of the wireless business. Amongst emerging markets the place that is getting more attention is India, as it is economically stable, open to technology and external vendors coming in and collaborating and working on new ideas.
For example, India is already making policies to diffuse technology and we created a programme in which Autodesk gave out Fusion 360’s to one of ‘Made in India’ initiatives. In order for us to be in a position to do programmes like these there has to be a market need, a government or a local institution that is willing to work with us in a collaborative way and skill base to build upon. If anyone of these three are missing it is difficult.
You spoke about doing better and more for less at Autodesk University 2017. What kind of government policies do you require to support this stance, especially in smaller countries?
Policy makers can play a critical role. In countries like Singapore we have been talking about how to truly industralise construction to build more sustainable and efficient cities without a lot of waste. The construction process is highly fragmented with complex relationships between owners, architects, contractors and subcontractors. Currently, there are disincentives in the process that makes people who are involved to do the right things technologically. The contractual and legal environment is not conducive to accelerate industrialised construction, but governments can play a significant role in facilitating it.
They can start mandating Building Information Modelling (BIM) for projects. They can also define roles and approval process that is required to move through the various stages and get involved with the accountability models of owners, architects, contractors. Governments can start accelerating the right thing to happen, instead of having the wrong incentives in the process.
Most countries in the GCC have major infrastructure plans, but a number of these plans were made when oil was at $100 plus per barrel levels. Given the softening of oil prices since 2014, most of these countries are facing budgetary constraints. Are there ways in which these countries can complete planned projects at a reasonable cost?
Waste and inefficiency in infrastructure building drive up cost. In addition, there is corruption across the world in this industry. If we drive automation and dis-intermediate people who are siphoning off with the process, we can brings down costs. An increase in construction productivity will eliminate unnecessary players enabling us to build more. Material waste and process waste both need to be checked. If something can be built with three subcontractors instead of 20 there is going to be less graft, more throughput and less waste. It’s not here today, but it is coming. The industrialisation of construction is going to be huge for municipalities and countries that are trying to build things. It’s going to help countries in the GCC and elsewhere build the infrastructure that is needed. It’s not here now, but we are getting there.
Generative design started off as a design tool, but it’s getting wider in terms of its potential. What are the limits that you see as far as generative design is concerned and is the availability of more data going to aid such processes?
Let’s be clear about what we mean by generative design. Generative design is the ability for us to use computational power and machine learning algorithms to provide options for numerous types of design challenges. We are providing information to the designer that informs numerous types of choices. So there can be design based geometries which can instantly be machined or 3D printed. Generative design is going to go all over the place, as the algorithms get better it is going to help us make something faster compared to what would have been possible in a traditional way. Data is important but algorithms matter too and generative design capability is going to show up everywhere.
Can generative design and artificial intelligence be combined to improve each other’s efficiency?
Generative design is in many ways an artificial intelligence application. It is a machine learning application which is being combined with an automated manufacturing process. At Autodesk we are trying to prototype handoffs between generatively created geometry and automated manufacturing processes. For instance, a robot has been programmed with machine learning and artificial intelligence algorithms to recognise the next step without additional programming. We are starting with simple things and are working on those machine visualisation problems so that we can use them to automate hand-offs through the generative design process.
The entertainment industry has been working with traditional technology, which is time-consuming and costly. Can generative design disrupt the model and do you see this happening soon?
It is going to be fairly disruptive as we are going to help people generate characters from their requirements and then morph those characters when they are done. A lot of time is currently being wasted in entertainment in creating the initial character forms, while the purpose is to bring those characters to life. The idea is to expend more energy in creating life like interactions with these characters. Till now we have not seen a Generative design entertainment example, but we are working on generating a portfolio of characters based on a set of constraints that are going to be able to grab quickly.
In manufacturing, Autodesk has Inventor and Fusion 360. These design tools have had overlapping functionalities and did not work well together. Is there a rationale to continue with them in the current fashion?
For a while Fusion 360 just looked like a different version of Inventor, though it was a cloud based solution. Moving forward, Fusion 360 is going to concentrate on the downstream automation of the manufacturing process like design made convergence, automation of workflow enabling a manufacturable set of instructions or product. Inventor and Fusion 360 will also become more interoperable. Fusion will be able to consume data from Inventor and help customers manufacture what they are making in the cloud using cloud computing and cloud power capabilities. Fusion is going to be our automation solution for design make convergence. It will have all the design functionality and CAM capability that it has today, but it is going to do more design make automations, making a continuous work flow between inventor and Fusion 360
Is your product life-cycle management (PLM) strategy undergoing a change due to the new functionalities being offered by Cloud?
PLM is being completely reimagined in the cloud. There is no doubt that all the old PLM systems are going to be outdated in 10 years. One of the amazing things about Cloud is that many of the data problems that we have been dealing with can be solved robustly in it. This is also the reason why it can be frustrating for us not to deploy cloud solutions everywhere. You get one central data model, one is able to disaggregate that data model and serve data in many different ways. It is the ultimate solution for a lot of data flow problems that we have had in the past. You are going to see us invest a lot of our data flow resources in the cloud. We use public cloud and the more that infrastructure is available within various geographies, and it provides us with the security and reliability that we need the more you will see us deploying these technologies. Right now we are involved with Amazon Web Services (AWS) to roll out our technology. The rate and pace at which we get into various countries will depend on when the third parties that we work with like AWS and Microsoft Azure get into various countries.