文章
文章
文章

Talking with Robots About Architecture

分享此文章

This article dives into the future of building: automation, communication, and whether robots will change everything. It offers an informed, useful, and realistic overview of how architects, engineers, and builders use automation today, and how they may use it tomorrow. Learn better ways to think about automation, how to apply it within your work, how you can use it, and how you will be likely disrupted by it in the near future.

Talking with Robots about Architecture See many real-world examples of automation in use today — both on the design side via Building Information Modeling (BIM) and on the physical side via robotics. Gain a better understanding of how to empower yourself through automation, and how to avoid wasting lots of time on dead ends — or worse, getting replaced altogether.

It is both for those doing the hands-on implementation of automation within their companies and for those who are managing those people and/or businesses. It starts with a general overview of the pluses and minuses of automation in general, then gets into ways of thinking about automation and tips to follow when automating things, and finishes with a big-picture take on how to make yourself more ‘automation proof.’

Automation Is Both Awesome and Terrible

Automation is both an awesome, empowering force, as well as a disruptive, wasteful one sometimes. It can be a profitable technology to leverage both personally in your day-to-day work and within the larger context of your company. But automation, like any other technology, when misapplied and/or misunderstood, can turn from an empowering force that will help your efforts to a huge waste of time and resources.

Automation is no more a magic bullet than any other technology. Robots only do exactly what you tell them to do, mostly. While it may seem that robots are taking everything over, until ‘strong AI’ comes along (which I personally doubt will ever happen) there is still much they are, and probably always will be, terrible at as well.

Automation Is Awesome

It’s Empowering

By automating difficult and/or repetitive tasks, we can empower ourselves and co-workers to be able to do bigger and better things. We can free ourselves from grunt work to focus on the more fun and interesting tasks.

It’s Interesting

Automating things is an awesome test of problem-solving skills, systems thinking, and detailed work. It’s a really fun thing for any designer, engineer, or builder to get into.

It’s Exciting

While all-software automation is somewhat exciting, physical robots are way cool. And it’s way more fun to automate some boring task than to sit in drudgery brute-forcing things.

It’s Profitable

You can create a whole lot of value via automation, which can then lead to more money. And if you can automate things, you can typically demand better positions and jobs. And if you own your own business, it can make you a lot of money.

Automation Is Terrible

It’s Expensive

Automating things takes a lot of time, money, and upkeep, so it better be worth it. And it’s hard to know when it’s worth it.

It’s Difficult

Programming and robots are already hard, figuring out the right things to automate is even harder sometimes.

It’s Fragile

Terminator this ain’t. Most automation only works if everything is exactly the way it’s supposed to be. A single thing out of place and it breaks. Or worse, when the client changes something, and breaks your whole automated workflow.

It’s Disruptive

Sometimes fully leveraging automation requires you to change the whole way you do business, which can be really challenging.

It’s All about Created Value

Because automation is typically expensive (in time, or equipment, or both!), and the most expensive thing we can do is make something that no one winds up really using, we need to be careful about how we automate things, and what we automate. We need to make certain we’re creating value by automating something, and that we’re in turn capturing that value. By focusing on, and properly measuring, the value we create via automation we can make certain it’s worth it, and that we’re really making things better.

Four Types of Automation

We’ve found that automation of tasks tends to fall into one or more of these four categories, each with its own special challenges and issues. By understanding which type of automation you’re dealing with, you can better tackle it, and make certain you’re actually creating value.

Type One: Same Job, Only Cheaper

This is where we’re replacing a task or even a whole job done by a person with automation, purely because it will be cheaper. The automated task won’t be done that much better, or something really complex won’t be made easy, but instead the automated task is just going to happen faster or take less people to do than the manual way of doing it. It’s pure economics, so with this sort of automation, you really have to be focused on making certain it really is cheaper in the long run.

With this type of automation, the mistake you can make is simply spending more automating the task than it would have cost to just pay someone to do it in the end. Setup and ongoing maintenance costs will be very important to keep under control. It’s easy for them to spiral out of control. You might have the best intentions, but get seduced away by what might be ‘better’ but really is just a waste of time or money. Or what you might think is easy turns out to be way harder to do. Doing whatever is the easiest and simplest thing you can do it probably best. Higher-level languages then, like Dynamo or built-in scripting within your apps, might be your best bet even though they aren’t as ‘good’ as ‘real’ applications written in lower-level Visual Studio or ‘better’ cloud-based apps.

Beware of ‘sunk costs’ too, where a big investment today into automating something turns into tomorrow’s albatross. Later I’ll talk more about the concept of ‘technical debt’ but for this sort of automation, you really want to calculate your opportunity costs and try to keep getting locked into expensive proprietary solutions that are hard to migrate away from in the future.

Basic Dynamo scripting to renumber rooms or to renumber doors to match their rooms in Revit is a good example here. Or simple scripting in Fusion or automating part modeling with parametric iFeatures in Inventor. It’s not that hard of a task, but it’s not something that’s going to be done that much better than a person doing it. It’s just simply faster.

A good physical example is an item like CNC-automated saw stops and fences for saws. Tigerstop is one of several companies that make one kind of these ‘automatic fences.’ You walk up to a chop saw or band saw equipped with one, punch a length into a connected tablet, and a stop for the saw automatically moves to that length. It’s nothing a skilled fabricator couldn’t do on their own, it’s just faster. And you can still easily cut the material wrong. And while you may think automation is always more error-free, typically that’s not as much of a sure thing. There are still plenty of things that can go wrong, from operator error, to GIGO, to misplaced items or data. So while automation may cut down on some of the errors, don’t count on it eliminating them.

For Type One automation, cheaper really has to be cheaper.

Type Two: Automation Does It Better

For certain sorts of tasks, automation simply can do a better job than a person sometimes, sometimes being the key word here. While a CNC mill can certainly outperform a human machinist many times over, it only can do so on certain tasks. And only when they are making the right things, in the right way. With this type of automation, the mistake you can make is automating the wrong thing, thinking you’re making things better when you’re actually not.

Automation can certainly be an incredibly powerful technology, empowering you to do something way better than ever before. Here it’s more than just that the automated task is faster, or cheaper done that way, it’s genuinely better in some way. Higher quality, significantly fewer errors, or more precise, it’s where you’re doing something that you could do, but that the automated solution is just simply constantly better at it.

However, it still holds true that automation is expensive. For this type of automation, getting it right is more about fully and truly understanding the nature of the task and workflow than the technology itself. Automation can just as easily do a whole lot of the wrong things just as quickly as the right things. It’s also easy to fall down rabbit holes or waste time focusing on the wrong things, spending money to automate something that didn’t really need it.

When you’re thinking of automating a task that falls into this type, there are three things you want to focus on. First is to use some techniques from the lean manufacturing and agile software worlds, in that you want to really study the task to make certain you understand it. Process mapping, asking five whys, and actually timing the current task and its upflow/downflow dependent task times can really help to solve if you’re automating the right task. Because it’s possible that making this specific task you’re thinking of happen faster and better might just create bottlenecks, backlogs, or piles of waste waiting for someone else to have to get to. This specific task might be done better, but it’s not making things really better.

Secondly, you should think if you really need for this particular task to be that much better. Sometimes better isn’t really needed, and perfection is the enemy of good enough. Just because you could automate something to be better doesn’t mean that it’s a good idea. Again, like with type one automation, you should do some cost benefit analysis before starting. But you should also think about how the task fits into the larger picture, and if there isn’t something else you might be able to tackle that could be more meaningful.

Lastly, you can’t automate away stupid. As humans will always be in the loop, don’t just assume that automation will solve issues that actually have to do with training, personnel, company culture, or unrelated industry problems. Like any technology, it’s not a magic bullet, and you’ll want to carefully consider if you’re actually automating the right things.

Model-based, BIM-to-fabrication workflows are a good example of this type of automation. Going from design, to detailed model, to actual fabrication can make things faster, more accurate, and with better quality and tolerances. Structural steel, HVAC ductwork, millwork and cabinets, curtain walls, precast concrete, and more are being designed with BIM tools, and then semi-automatically fabricated with automated tools. It’s a very powerful way to work.

For Type Two automation, better has to actually be better.

Type Three: Automation Makes It Easy

For some kinds of tasks, it’s more about how automation can make what used to be extremely difficult easy. It’s not just faster to do it this way, or results in a better end product, it’s that it also makes it just much easier to produce overall. Generative design is a great example of this type of automation. While it may have been possible with traditional CAD apps to design something impossibly complex, it would have also taken a frightening amount of time. By automating parts of the design itself, it can empower you to be able to make buildings far more complex than before, but in less time. Or have an app automatically easily generate thousands or millions of variations on a design to study it.

3D printing, while overhyped, is another good example of this type of automation. While it was possible to sculpt, mill, or cast materials into complex forms, the automated process of 3D printing makes producing things like this easy. With this type of automation, the mistakes you can make are more in the design of the automation itself; if it’s too hard to use, too confusing, too complex, then it’s not really easier. It’s more a problem of the user interface and user experience being good than it is about the underlying technology being as good as it can be.

With this type of automation the focus should be on empowerment. You want to keep people on what they are good at, and the automation on what it’s good at. There’s a newish term for this sort of thing, called a Centaur. It’s automation and a human, working together, each doing what it’s best at. It’s much more collaborative than the other types of automation.

For Type Three automation, easy has to be easy.

Type Four: Automation Makes It Even Possible

Finally, for some sorts of things, automation makes it even possible in the first place. Without it, you simply couldn’t work that way at all, or make the things you can by using it. This tends to be leveraging a newly available technology, or new combinations of technologies, and because of that, it tends to also require large business-wide shifts to make best use of it. It’s not just faster, or better, or easier, it’s a whole new way of doing something. And it can be awesome, or disruptive. Or both!

A great example of this type of technology is when CAD, and then BIM, first became commonly available to our industry. It automated a great deal of our drafting production and gave everyone a whole new way to work, but required new ways of working, new roles within your company, and new costs and opportunities as well. The mistake you can make with this type of automation is that it tends to require a systematic, holistic approach, that might require great changes to your overall business or industry to really make use of it. Just because it’s now possible, it might not be really useful, or not useful for your specific situation. Sometimes whole businesses have to be built off this sort of automation for it to really work and create enough value to make it worth it.

For Type Four automation, just because it’s now possible doesn’t mean it’s worth it.

Tips When Automating Things

No matter what type of automation we’re looking at, and whether you are personally working on automating a task or managing someone who is, there are some general things you’ll want to keep in mind.

Handle — Body — Tip

Automation is just a tool. And whenever you are designing a tool, it’s good to break it down into three parts, and focus on how to make each part the best it’s able to be. That is because these three parts have different goals, yet have to work together, to make for a good tool. I think the first thing to consider is the ‘Handle’ of the tool, or rather the part that faces the worker. What’s the User Interface and User Experience supposed to be like? What goals do we have for it, and how to we measure if we’re meeting those goals? Are we looking to make something very custom for a small group of experts, or something anyone can pick up and use?

Then the ‘Body’ of the tool, or rather the part that’s the actual code or physical robotics. What systems are you going to standardize on? What trade-offs and compromises are you making by doing that? How are we going to generate ‘unit tests’ so we can test the important parts of our solution to be certain it’s meeting our goals for production? Is there another platform or system that we should test as well?

Then finally, the tip, or rather the part that actually touches the work. For code, this might be the specific API or file formats you’ll be working with, for physical robotics you might be looking at the tooling, cutters, or actuators; either way it’s the bit that touches what you’re working on. How does it interface with the work? Does it introduce errors? And how do we measure that? Is there a better solution available off-the-shelf, or do you really need to custom design something? By defining these things before you start you can save a lot of time effort, and in the end make much better tools.

Measuring and Metrics

So again, if automation is expensive and complex, we need to be focused on the actual value created by it. One good way to do that is by smartly measuring things along the way. One way to do this is by focusing on just a few important metrics to measure before, during, and after something has been automated. And it’s best to couple these together into what are sometimes called cohort metrics where you measure two or more different things, and compare them versus each other over time, instead of just looking at one or two measurements in isolation.

For a Type One automation, you might focus on task time, or the overall time it took to complete the task prior to automation and then after automation. However, you’d also want to keep track of machine downtime, total units produced per year, and maintenance costs over the same timeframe, and compare all three numbers side-by-side, so you get a real picture of what’s going on. Just because the time to complete that task might have gone way down, if it’s also actually producing the same or less than before because the machine is down all the time then it’s vital to understand for really measuring the value produced. A really great book that digs into this concept well is High Output Management by Andrew Grove, which I highly recommend everyone read. Several times!

Technical Debt

This is a wonderful term from the software industry. It reflects the implied cost of additional rework caused by choosing an easy solution now instead of using a better approach that would take longer or cost more. It’s a great way to put an actual dollar amount, even if it’s fuzzy, on the compromises you may be making as you design your automated system.

For example, if I’m looking at buying two different CNC routers to automate production within a woodshop I could buy a cheap slower entry-level one or a much more expensive fast fancy one. If I go with the cheap one, with the assumption I might have to upgrade later when things get busier, I’ve incurred a technical debt equal to the difference in price between those two tools (and the cost of their setup). A debt I’ll have to ‘pay off’ in order for my automation plan to move forward once production hits a certain demand.

Another example is that you’ve decided to stick with your largely AutoCAD-based automated scripting for shop drawing production. In order to upgrade in the future, you’d not only have to buy Revit or Inventor or whatever, you’d also have to hire someone to rewrite everything for the whole new platform. So that decision is ‘borrowing from the future’ and creating possible large expenses in the near-term that could sink you when a competitor shows up that doesn’t have that debt.

So carefully consider your choices, and the future costs they may have. Technical debt, like any other debt, isn’t inherently bad. Debt can be very helpful for a business to grow faster or jump on other opportunities. But it has to be managed properly, otherwise it can quickly destroy any gains you might have produced with your automated solution.

Larger Issues around Automation

There are some fundamental issues around automation, and some that are specific to AECO, that are important to keep in mind.

Automation Is Fragile

Terminator this ain’t. Trading redundancy for efficiency is a great deal until a single small part stops working, and you have no other way of producing the same work until it’s fixed. Software and hardware break all the time, either due to internal or external forces. So thinking that automated solutions are robust and stable is foolish, for even fully approved software updates can completely sink your systems, and just simply time is always working against you as companies stop supporting those systems or even just as important parts wear out.

Automation Is Stupid

Robots only do exactly what you tell them to. Your automated solution can’t also invent a better way of doing something later on, but a person totally could. While automation can add a lot of value once, people can add value constantly.

Automation Is Boring

People honestly only care about the things other people do. A great example here is that Mozart, the famous composer, made a game where you roll dice, follow some rules, and the game writes a waltz a quartet could play. Even a beginning programmer can take this, write it as a program, and produce more waltzes than you can listen to in a lifetime. And no one cares about any of those songs. Because we as people just care about what other people have done.

So even if you could fully automate something as complex as producing a building, those buildings would be totally boring and no one would really care about them very much. And even if you could fully automate the production of something like apartment complexes, it would be the people living in them (and thus customizing them!) that would actually give them any real value. Or make anyone care very much about them at all.

When we started our business we made a big deal about how we used automation to produce all the creative elements our company makes. And not one of our clients cared at all about that fact. It’s mildly interesting for a moment, but they honestly care that we’re great to work with, are going to deliver as promised, and that they like what we do.

Buildings Aren’t Cars

Why can’t we build buildings the way that Tesla makes cars? Well, cars are mass-produced items made of mostly totally custom parts, while buildings are totally custom items made of mostly mass-produced parts. Automation for AECO is going to look very different than it does for other industries. Also buildings are a result of four sometimes-conflicting, hyper-local forces:

Politics - Building - Performance

Culture is the design itself and how it relates to its location, users, context, aesthetics, etc. There are some things here that can be generally automated, such as form generation or generative design, but it tends to be very project-specific, and what is useful now might be ‘dated’ in the near future.

Performance is how well the building ‘works’; it’s energy consumption, it’s efficiency of materials, how it performs during events like fires or earthquakes. Lots can be automated here, both in the design and operation of the building.

Economics is not only how much did it cost to build, it’s also its operating costs, rents it’s able to collect, cost to alter in the future, etc. While some automation might help reduce the construction or operating costs, there are large factors here (such as rents you can charge) that just aren’t problems you can address via automation.

Politics is all the regulatory issues that always surround buildings. While some permit drawing production can be automated, things can vary so much from city-to-city that such automation can quickly become so complex as to be not worth it.

Strategies for Surviving Automation

Want to not be replaced by rbots? Here are some general ideas and tips from someone who’s spent the last year working alongside robots in AECO.

Empowerment vs. Disempowerment

Do you tell the robots what to do, or do they tell you what to do? And, more importantly, are those robots working for you, or are you working for them? For example, as a small business owner, if I automate some task, even if that automation replaces something I currently do, even if it’s now telling me what to do, I’m still being empowered by that automation. For the extra value it creates I’ll capture in some way.

Whereas if I’m working for a large company, and whatever I’m doing, or whatever my boss is doing, can easily be replaced by automation I’m much more likely to be disempowered by that. While Uber and Lyft drivers are ‘empowered’ to work their own hours (which is way cool) they also have little control over Uber or Lyft themselves, which could easily decide they don’t need that particular driver anymore for reasons that drive has no control over. So putting yourself into positions where more automation just empowers you more is where you want to be. How can we do that?

Wicked Problems

One way to do that is by focusing on wicked problems. These are problems that are complex, dynamic, with lots of interrelated parts, near-infinite possible solutions, and/or fuzzy victory conditions. Turns out computers are just bad at solving these sorts of problems usually, while humans are awesome at them.

A great example here I think is the game of Go. It’s a complex, interrelated game of near-infinite solutions. But it’s got a well-defined victory state in that there are rules for the game that you can’t change. So thus how ‘good’ a game was is something that can be measured. And thus automated. While Go remained a hard problem for AI researchers, a clever machine learning team just made a Go robot that is now at grandmaster status, and it even plays in a way no human ever would. They did this by having it play countless millions of games against a copy of itself until it learned to play well. It wasn’t really a wicked problem, just a very hard one to solve until recently. And even with it ‘solved’ it’s not like anyone is going to stop playing Go, even professionally, as Go itself is interesting enough of a game for people to feel it has value. The super-smart Go robots will likely just get their own league and everyone will keep playing Go.

But if Go was Calvinball, and the rules could change as the game was played, this approach to automating it simply wouldn’t work, for there wouldn’t be a way to measure who ‘won’ as easily. Or if you made a Go board that was twice as big, exponentially expanding the number of possible moves beyond what can be easily computed currently. Or if you added some more social rules to Go, like bidding in poker, where you have to tell if someone is lying or not. Or if Go was more complex in how the pieces interrelated, in that some pieces change in relation to how other pieces are played.

So focusing on wicked problems is one way to stay relevant no matter what happens with automation. Design, engineering, and construction are full of such problems. Become an expert in one, such as local building codes (and the politics within), or software optimization, or cost-efficient structural design for a particular complex building type, or in existing facade renovation for mid- or high-rise buildings (which are all going to need it in the next 20 to 50 years) and even if a robot comes along, it won’t be able to keep up…

Soft Skills

The other way to ‘future proof’ yourself is to learn some soft skills. People like to hire their friends. Even when all else isn’t equal. So gaining some skills in sales or other customer-facing area is a sure way to stay relevant. People like to work with their friends. So learning some great management skills to get the most out of the people you work with, and who work for you, is another great way to stay relevant.

And robots really don’t buy anything. So understanding how to get people’s attention, communicate a story, and some marketing basics is another great soft skill to have. Robots are terrible business people. Learning how to be a business-focused, value-creating, creative problem solver is another way to stay relevant.

Finally, robots are bad at coming up with things the world’s never seen before. A/B testing would have never designed the first iPhone, for example. Gaining some design skills, learning what people really would want, and how to deliver on that will make you irreplaceable.

Jeffrey McGrew is co-founder of Because We Can, an award-winning design-build architecture studio in Oakland, California. As co-founder and lead architect, his responsibilities for design direction and digital fabrication strategies are balanced with business demands of sales, marketing, and project management.

______
icon-svg-close-thick

Cookie 首选项

您的隐私对我们非常重要,为您提供出色的体验是我们的责任。为了帮助自定义信息和构建应用程序,我们会收集有关您如何使用此站点的数据。

我们是否可以收集并使用您的数据?

详细了解我们使用的第三方服务以及我们的隐私声明

绝对必要 – 我们的网站正常运行并为您提供服务所必需的

通过这些 Cookie,我们可以记录您的偏好或登录信息,响应您的请求或完成购物车中物品或服务的订购。

改善您的体验 – 使我们能够为您展示与您相关的内容

通过这些 Cookie,我们可以提供增强的功能和个性化服务。可能由我们或第三方提供商进行设置,我们会利用其服务为您提供定制的信息和体验。如果您不允许使用这些 Cookie,可能会无法使用某些或全部服务。

定制您的广告 – 允许我们为您提供针对性的广告

这些 Cookie 会根据您的活动和兴趣收集有关您的数据,以便向您显示相关广告并跟踪其效果。通过收集这些数据,我们可以更有针对性地向您显示与您的兴趣相关的广告。如果您不允许使用这些 Cookie,您看到的广告将缺乏针对性。

icon-svg-close-thick

第三方服务

详细了解每个类别中我们所用的第三方服务,以及我们如何使用所收集的与您的网络活动相关的数据。

icon-svg-hide-thick

icon-svg-show-thick

绝对必要 – 我们的网站正常运行并为您提供服务所必需的

Qualtrics
我们通过 Qualtrics 借助调查或联机表单获得您的反馈。您可能会被随机选定参与某项调查,或者您可以主动向我们提供反馈。填写调查之前,我们将收集数据以更好地了解您所执行的操作。这有助于我们解决您可能遇到的问题。. Qualtrics 隐私政策
Akamai mPulse
我们通过 Akamai mPulse 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Akamai mPulse 隐私政策
Digital River
我们通过 Digital River 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Digital River 隐私政策
Dynatrace
我们通过 Dynatrace 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Dynatrace 隐私政策
Khoros
我们通过 Khoros 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Khoros 隐私政策
Launch Darkly
我们通过 Launch Darkly 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Launch Darkly 隐私政策
New Relic
我们通过 New Relic 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. New Relic 隐私政策
Salesforce Live Agent
我们通过 Salesforce Live Agent 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Salesforce Live Agent 隐私政策
Wistia
我们通过 Wistia 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Wistia 隐私政策
Tealium
我们通过 Tealium 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Tealium 隐私政策
Upsellit
我们通过 Upsellit 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Upsellit 隐私政策
CJ Affiliates
我们通过 CJ Affiliates 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. CJ Affiliates 隐私政策
Commission Factory
我们通过 Commission Factory 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Commission Factory 隐私政策
Google Analytics (Strictly Necessary)
我们通过 Google Analytics (Strictly Necessary) 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Google Analytics (Strictly Necessary) 隐私政策
Typepad Stats
我们通过 Typepad Stats 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Typepad Stats 隐私政策
Geo Targetly
我们使用 Geo Targetly 将网站访问者引导至最合适的网页并/或根据他们的位置提供量身定制的内容。 Geo Targetly 使用网站访问者的 IP 地址确定访问者设备的大致位置。 这有助于确保访问者以其(最有可能的)本地语言浏览内容。Geo Targetly 隐私政策
SpeedCurve
我们使用 SpeedCurve 来监控和衡量您的网站体验的性能,具体因素为网页加载时间以及后续元素(如图像、脚本和文本)的响应能力。SpeedCurve 隐私政策
Qualified
Qualified is the Autodesk Live Chat agent platform. This platform provides services to allow our customers to communicate in real-time with Autodesk support. We may collect unique ID for specific browser sessions during a chat. Qualified Privacy Policy

icon-svg-hide-thick

icon-svg-show-thick

改善您的体验 – 使我们能够为您展示与您相关的内容

Google Optimize
我们通过 Google Optimize 测试站点上的新功能并自定义您对这些功能的体验。为此,我们将收集与您在站点中的活动相关的数据。此数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID 等。根据功能测试,您可能会体验不同版本的站点;或者,根据访问者属性,您可能会查看个性化内容。. Google Optimize 隐私政策
ClickTale
我们通过 ClickTale 更好地了解您可能会在站点的哪些方面遇到困难。我们通过会话记录来帮助了解您与站点的交互方式,包括页面上的各种元素。将隐藏可能会识别个人身份的信息,而不会收集此信息。. ClickTale 隐私政策
OneSignal
我们通过 OneSignal 在 OneSignal 提供支持的站点上投放数字广告。根据 OneSignal 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 OneSignal 收集的与您相关的数据相整合。我们利用发送给 OneSignal 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. OneSignal 隐私政策
Optimizely
我们通过 Optimizely 测试站点上的新功能并自定义您对这些功能的体验。为此,我们将收集与您在站点中的活动相关的数据。此数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID 等。根据功能测试,您可能会体验不同版本的站点;或者,根据访问者属性,您可能会查看个性化内容。. Optimizely 隐私政策
Amplitude
我们通过 Amplitude 测试站点上的新功能并自定义您对这些功能的体验。为此,我们将收集与您在站点中的活动相关的数据。此数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID 等。根据功能测试,您可能会体验不同版本的站点;或者,根据访问者属性,您可能会查看个性化内容。. Amplitude 隐私政策
Snowplow
我们通过 Snowplow 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Snowplow 隐私政策
UserVoice
我们通过 UserVoice 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. UserVoice 隐私政策
Clearbit
Clearbit 允许实时数据扩充,为客户提供个性化且相关的体验。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。Clearbit 隐私政策
YouTube
YouTube 是一个视频共享平台,允许用户在我们的网站上查看和共享嵌入视频。YouTube 提供关于视频性能的观看指标。 YouTube 隐私政策

icon-svg-hide-thick

icon-svg-show-thick

定制您的广告 – 允许我们为您提供针对性的广告

Adobe Analytics
我们通过 Adobe Analytics 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Adobe Analytics 隐私政策
Google Analytics (Web Analytics)
我们通过 Google Analytics (Web Analytics) 收集与您在我们站点中的活动相关的数据。这可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。我们使用此数据来衡量我们站点的性能并评估联机体验的难易程度,以便我们改进相关功能。此外,我们还将使用高级分析方法来优化电子邮件体验、客户支持体验和销售体验。. Google Analytics (Web Analytics) 隐私政策
AdWords
我们通过 AdWords 在 AdWords 提供支持的站点上投放数字广告。根据 AdWords 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 AdWords 收集的与您相关的数据相整合。我们利用发送给 AdWords 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. AdWords 隐私政策
Marketo
我们通过 Marketo 更及时地向您发送相关电子邮件内容。为此,我们收集与以下各项相关的数据:您的网络活动,您对我们所发送电子邮件的响应。收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、电子邮件打开率、单击的链接等。我们可能会将此数据与从其他信息源收集的数据相整合,以根据高级分析处理方法向您提供改进的销售体验或客户服务体验以及更相关的内容。. Marketo 隐私政策
Doubleclick
我们通过 Doubleclick 在 Doubleclick 提供支持的站点上投放数字广告。根据 Doubleclick 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Doubleclick 收集的与您相关的数据相整合。我们利用发送给 Doubleclick 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Doubleclick 隐私政策
HubSpot
我们通过 HubSpot 更及时地向您发送相关电子邮件内容。为此,我们收集与以下各项相关的数据:您的网络活动,您对我们所发送电子邮件的响应。收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、电子邮件打开率、单击的链接等。. HubSpot 隐私政策
Twitter
我们通过 Twitter 在 Twitter 提供支持的站点上投放数字广告。根据 Twitter 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Twitter 收集的与您相关的数据相整合。我们利用发送给 Twitter 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Twitter 隐私政策
Facebook
我们通过 Facebook 在 Facebook 提供支持的站点上投放数字广告。根据 Facebook 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Facebook 收集的与您相关的数据相整合。我们利用发送给 Facebook 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Facebook 隐私政策
LinkedIn
我们通过 LinkedIn 在 LinkedIn 提供支持的站点上投放数字广告。根据 LinkedIn 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 LinkedIn 收集的与您相关的数据相整合。我们利用发送给 LinkedIn 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. LinkedIn 隐私政策
Yahoo! Japan
我们通过 Yahoo! Japan 在 Yahoo! Japan 提供支持的站点上投放数字广告。根据 Yahoo! Japan 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Yahoo! Japan 收集的与您相关的数据相整合。我们利用发送给 Yahoo! Japan 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Yahoo! Japan 隐私政策
Naver
我们通过 Naver 在 Naver 提供支持的站点上投放数字广告。根据 Naver 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Naver 收集的与您相关的数据相整合。我们利用发送给 Naver 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Naver 隐私政策
Quantcast
我们通过 Quantcast 在 Quantcast 提供支持的站点上投放数字广告。根据 Quantcast 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Quantcast 收集的与您相关的数据相整合。我们利用发送给 Quantcast 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Quantcast 隐私政策
Call Tracking
我们通过 Call Tracking 为推广活动提供专属的电话号码。从而,使您可以更快地联系我们的支持人员并帮助我们更精确地评估我们的表现。我们可能会通过提供的电话号码收集与您在站点中的活动相关的数据。. Call Tracking 隐私政策
Wunderkind
我们通过 Wunderkind 在 Wunderkind 提供支持的站点上投放数字广告。根据 Wunderkind 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Wunderkind 收集的与您相关的数据相整合。我们利用发送给 Wunderkind 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Wunderkind 隐私政策
ADC Media
我们通过 ADC Media 在 ADC Media 提供支持的站点上投放数字广告。根据 ADC Media 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 ADC Media 收集的与您相关的数据相整合。我们利用发送给 ADC Media 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. ADC Media 隐私政策
AgrantSEM
我们通过 AgrantSEM 在 AgrantSEM 提供支持的站点上投放数字广告。根据 AgrantSEM 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 AgrantSEM 收集的与您相关的数据相整合。我们利用发送给 AgrantSEM 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. AgrantSEM 隐私政策
Bidtellect
我们通过 Bidtellect 在 Bidtellect 提供支持的站点上投放数字广告。根据 Bidtellect 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Bidtellect 收集的与您相关的数据相整合。我们利用发送给 Bidtellect 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Bidtellect 隐私政策
Bing
我们通过 Bing 在 Bing 提供支持的站点上投放数字广告。根据 Bing 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Bing 收集的与您相关的数据相整合。我们利用发送给 Bing 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Bing 隐私政策
G2Crowd
我们通过 G2Crowd 在 G2Crowd 提供支持的站点上投放数字广告。根据 G2Crowd 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 G2Crowd 收集的与您相关的数据相整合。我们利用发送给 G2Crowd 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. G2Crowd 隐私政策
NMPI Display
我们通过 NMPI Display 在 NMPI Display 提供支持的站点上投放数字广告。根据 NMPI Display 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 NMPI Display 收集的与您相关的数据相整合。我们利用发送给 NMPI Display 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. NMPI Display 隐私政策
VK
我们通过 VK 在 VK 提供支持的站点上投放数字广告。根据 VK 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 VK 收集的与您相关的数据相整合。我们利用发送给 VK 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. VK 隐私政策
Adobe Target
我们通过 Adobe Target 测试站点上的新功能并自定义您对这些功能的体验。为此,我们将收集与您在站点中的活动相关的数据。此数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID、您的 Autodesk ID 等。根据功能测试,您可能会体验不同版本的站点;或者,根据访问者属性,您可能会查看个性化内容。. Adobe Target 隐私政策
Google Analytics (Advertising)
我们通过 Google Analytics (Advertising) 在 Google Analytics (Advertising) 提供支持的站点上投放数字广告。根据 Google Analytics (Advertising) 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Google Analytics (Advertising) 收集的与您相关的数据相整合。我们利用发送给 Google Analytics (Advertising) 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Google Analytics (Advertising) 隐私政策
Trendkite
我们通过 Trendkite 在 Trendkite 提供支持的站点上投放数字广告。根据 Trendkite 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Trendkite 收集的与您相关的数据相整合。我们利用发送给 Trendkite 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Trendkite 隐私政策
Hotjar
我们通过 Hotjar 在 Hotjar 提供支持的站点上投放数字广告。根据 Hotjar 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Hotjar 收集的与您相关的数据相整合。我们利用发送给 Hotjar 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Hotjar 隐私政策
6 Sense
我们通过 6 Sense 在 6 Sense 提供支持的站点上投放数字广告。根据 6 Sense 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 6 Sense 收集的与您相关的数据相整合。我们利用发送给 6 Sense 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. 6 Sense 隐私政策
Terminus
我们通过 Terminus 在 Terminus 提供支持的站点上投放数字广告。根据 Terminus 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 Terminus 收集的与您相关的数据相整合。我们利用发送给 Terminus 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. Terminus 隐私政策
StackAdapt
我们通过 StackAdapt 在 StackAdapt 提供支持的站点上投放数字广告。根据 StackAdapt 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 StackAdapt 收集的与您相关的数据相整合。我们利用发送给 StackAdapt 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. StackAdapt 隐私政策
The Trade Desk
我们通过 The Trade Desk 在 The Trade Desk 提供支持的站点上投放数字广告。根据 The Trade Desk 数据以及我们收集的与您在站点中的活动相关的数据,有针对性地提供广告。我们收集的数据可能包含您访问的页面、您启动的试用版、您播放的视频、您购买的东西、您的 IP 地址或设备 ID。可能会将此信息与 The Trade Desk 收集的与您相关的数据相整合。我们利用发送给 The Trade Desk 的数据为您提供更具个性化的数字广告体验并向您展现相关性更强的广告。. The Trade Desk 隐私政策
RollWorks
We use RollWorks to deploy digital advertising on sites supported by RollWorks. Ads are based on both RollWorks data and behavioral data that we collect while you’re on our sites. The data we collect may include pages you’ve visited, trials you’ve initiated, videos you’ve played, purchases you’ve made, and your IP address or device ID. This information may be combined with data that RollWorks has collected from you. We use the data that we provide to RollWorks to better customize your digital advertising experience and present you with more relevant ads. RollWorks Privacy Policy

是否确定要简化联机体验?

我们希望您能够从我们这里获得良好体验。对于上一屏幕中的类别,如果选择“是”,我们将收集并使用您的数据以自定义您的体验并为您构建更好的应用程序。您可以访问我们的“隐私声明”,根据需要更改您的设置。

个性化您的体验,选择由您来做。

我们重视隐私权。我们收集的数据可以帮助我们了解您对我们产品的使用情况、您可能感兴趣的信息以及我们可以在哪些方面做出改善以使您与 Autodesk 的沟通更为顺畅。

我们是否可以收集并使用您的数据,从而为您打造个性化的体验?

通过管理您在此站点的隐私设置来了解个性化体验的好处,或访问我们的隐私声明详细了解您的可用选项。