AI in Design: Q&A with our founder and CTO
Alyson Hallander / April 1, 2019
PlanIT Impact is the first tech firm in the AEC industry to leverage AI to enable easy, fast, collaborative, and smart design of more efficient buildings.
Our energy module, driven by our Discovery Assistant, employs a Machine Learning (ML) model, which is a type of AI, to suggest optimized design strategies that balance energy use and return on investment (ROI).
As new AI fueled technologies like PlanIT Impact hit the AEC market, many question how AI works, what it means for the future of building design, and how it will impact our daily professions. By the end of the technological revolution in the not so distant future, our lives will be, assumingly, radically changed for the better, but how?
To get answers, I interviewed our founder, Dominique Davison, and one of the AI masterminds behind PlanIT Impact, David Scheer, to dig into how they view the future of AI in the AEC industry and to shed light on how PlanIT Impact uses AI. Read on!
What role do you see AI having in the way buildings are designed?
Software enabled with Machine Learning, a subset of AI, is exciting because it profoundly speeds up computation and analytics. It allows us as a profession to accelerate solving the very complex problems of optimizing building efficiency, something that is more critical than ever.
However, just like any piece of software or technology, if you don’t bring intelligence to the tool, it can lead to the old “garbage in, garbage out” idiom. With every major shift in technology, there was a fear of it reducing the agency of the architect or engineer. The introduction of Computer Aided Design back in 1961 and easier 3D modeling with SketchUp decades later, among others, had a similar response. Even if any intrepid individual that wants to delve into these tools might be able to design something more holistically and comprehensively, if you don’t bring a strong, basic understanding of good design and building performance principles, the result won’t be great.
The role of AI in engineering is more easily understood than its role in design. With that said, AI’s role in design is twofold:
- It helps automate tedious tasks by predicting and fulfilling typical and repetitive needs of the design (i.e. cutting a hole in a BIM model wall for a window, laying out an automated grid for lighting fixtures, and assisting with space planning according to predicate rules). This helps the designer focus on creativity without being distracted by technical limitations.
- AI can do so much more than guide the designer’s workflow; it also creates a richer design environment in which to work. PlanIT Impact uses AI to enhance rather than replace the work of the designer – AI provides a layer of performance information at the speed of the designer’s normal workflow, as well as a multitude of alternative suggestions.
My architecture design professor told us to extend the lines we draw beyond the corners of the walls – suggesting other options and influences of that wall line…leading to the next design revision. With PlanIT Impact’s AI, additional performance possibilities are presented in the same way.
In PlanIT Impact, what does the AI do? How does it work?
PlanIT Impact uses AI in two ways:
- It is the predictive ML model for project performance, allowing us to provide performance results at the speed of the designer’s workflow versus breaking up the design flow to set up a simulation and wait for it to run.
Although complex, building energy performance has a predictable connection to the parameters of a building design. With enough training data, a ML model can generate performance answers as quickly as the designer can make the changes – placing a new window in a wall provides not only the visual feedback of the BIM model, but also the energy performance feedback of the design change.
- The ability to provide a sandbox space according to the outcomes defined by the designer. In other words, simply calculating and showing the performance effect of a design move after it’s made allows AI to generate a range of ideas for the next design move. It provides information to the designer to help him/her make the next design decision faster and with greater success.
AI increases the speed of a new model, and all possible solutions that meet either stated or assumed performance goals can be generated before additional design decisions are made.
How does PlanIT Impact’s engine continue to “learn” to refine its methodology and results?
A machine learning model is “trained” using loads of data that represent possible design options. This allows solutions not already calculated to be predicted based off patterns learned from the training data. At any time, the training data can be enhanced and refined with even more data that will improve the accuracy of the predictive model.
In an existing building, for example, where training data may make assumptions about occupancy schedules, build quality, weather data, and more, PlanIT Impact can continuously retrain the model with actual metered data to help improve future predictions.
We can also use AI to improve the definition of the performance goals used to drive suggested solutions. PlanIT Impact is working on developing this feature even further.
What is your mission for PlanIT Impact in helping architects leverage AI for the better?
PlanIT Impact’s mission is to make a significant contribution to the industry – reducing carbon impact and water use by removing barriers of efficiency and sustainability to design.
Because PlanIT Impact is the fast, easy, collaborative, SMART, and most powerful solution to building performance assessment, we are able to remove some of the most common stumbling blocks (i.e. specialized expertise, knowing the costs, too much time, and complexity), which prevent clients from proper modeling and achieving higher levels of performance. With PlanIT Impact, building professionals are empowered with the answers they need to help building owners and communities make smarter, more sound decisions toward a carbon neutral future.
Dominique Davison, AIA LEED BD+C, is the founder and CEO of PlanIT Impact, and founding principal of DRAW Architecture + Urban Design, an award-winning practice based in Kansas City, MO. Winner of the AIAKC Architect of the Year, she is an author and frequent national speaker in the sustainable and smart city space.
Originally from the San Francisco area, David Scheer brings an array of experience to PlanIT Impact as a .NET developer, energy modeling expert and architect. Scheer has been a consultant in architectural design, and energy and automation, and was a senior product manager at Autodesk.