The majority of professional design software today allows designers to automate specific actions, which are manually scripted to generate a specific and predictable output. This type of automation optimizes the speed of performing routine tasks, but does little else. The future of automation for the professional designer will instead open up completely new possibilities.

This new kind of automation will augment the designer's toolkit by improving the work, surfacing hidden connections, and inspiring in novel ways. Automated design tools will not just save time; they will amplify the designer's capabilities. They will help analyze a problem and explore possible solutions at a speed, volume, and level of sophistication outpacing anything that would be possible for a human designer. Future designers will do better work as automation begins to add real value to the design process.

This is an optimistic take on automation, presented at a time when others in the professional community are expressing concern about applying artificial intelligence and machine learning to creativity and design. Advancing the role of automation in design does not mean creating robot designers able to fully reproduce and extend the work of human designers. Google's self-driving cars can parse a lot of inputs, and Deep Mind's video game playing agents can learn through trial and error, but they do not exhibit creativity or an understanding of the human context. Design is a very different type of challenge for artificial intelligence because it is initially concerned with identifying meaningful problems to solve, and only then invests itself in producing valid answers. Formalizing that process, in order to reproduce it through automation, is a lofty goal.

There are, however, promising developments that indicate automation in design can create value. For example, TheGrid—a startup in San Francisco—is already building technology that will allow designers to introduce algorithms in the design toolkit. The company is currently beta-testing what they call the "first artificial intelligence platform that creates websites," which means that their product will build a unique, self-optimizing website around the content and constraints fed to it. That type of technology has the potential to put an end to low-value activities currently performed by designers (e.g. personalization of templates, data entry, etc.), empower non-designers to achieve quality results, and give professional designers the possibility to scale design approaches through the use of algorithms. 

[Editors Note: Leigh Taylor, Chief Creative Officer of TheGrid, explains how scalable design will change the way we build websites here.]

This type of pioneering work will bring a radical paradigm shift in design: the designer provides instructions to automate design decisions. Automation no longer only serves up repeatable and predicable solutions, but becomes a tool that expands design possibilities. As this type of transformative automation continues to advance, here a few perspectives on how it is going to improve the designer's toolkit.

1. Design Research: Richer and More Insightful

In the research phase designers build a deep understanding of the target audience. During this phase designers capture a vast array of information in written, audio, and visual form. The material is analyzed and synthesized in combination with other data (e.g. market analysis and quantitative research), with the goal of uncovering patterns and making connections that generate insights for the creation of relevant design concepts.

Designers would benefit from tools that automate the collection and organization of data. At frog we envisioned a new kind of research toolkit based on distributed sensors, which enables designers to not only amplify the quality of information but also extend the research into new areas (e.g. accurate measurements of air quality, noise levels, and the strength of data coverage). This initiative addresses the challenge of pre-processing information captured during research, in order to make it available while the team is still in the field. Automating the preliminary organization and display of information improves both the quality and the efficiency of the research.

2. Design Inspiration: Broader and Faster

Design research and other discovery activities lead to a set of data that must be reviewed for patterns. In the same way algorithms currently support music discovery (e.g. Rdio, Pandora, Spotify), the next generation of design software will automatically curate knowledge and inspiration relevant to the design objective.

Designers will define the scope of automated searches using a spectrum that is very specific to their design goals on one end, and very broad on the other end. The software—or inspiration engine—will organize the retrieved assets based on the criteria defined by the designer, as well as categories and frameworks suggested by the tool itself. Designers will work together with the software to make connections between different types of content and highlight both analogies and patterns.

The search will continue in a more discrete range once designers identify meaningful connections by combining and prioritizing sets of inspirations. Custom scripts, modules, plug-ins, and additional content libraries will increase the software's relevance to specific fields of design. Integration with computing platforms will maximize accuracy, collaboration, and processing speed.

3. Design Exploration: More Concepts, Less Labor

The exploration of design concepts is a process that combines several types of knowledge: business context, design intent, user insights, relevant design inspirations and best practices.

In the near future, advanced algorithms will take the designer's germinal concepts—perhaps about a digital service, or a physical product or space—and explore possible variations and combinations at an extraordinary volume and speed. In the time that it currently takes a designer to envision one concept, an automated exploratory design tool will generate thousands of design options based on the same concept description, and even rank them according to several success criteria such as extensibility and performance. The designer will then select the most promising design alternatives for further development and validation.

Designers will have more time to come up with concepts because automated tools will take on the labor-intensive task of exploring all of the valid options according to given constraints and variables. Autodesk is already a pioneer in this space with Project Dreamcatcher, and it follows that others should pay close attention to the possibilities in this area as new tools are developed.

4. Design Productivity: Personal and Flexible

Once the most promising concept or combination of concepts is finally defined, the designer needs to translate the concept into a coherent solution built on an extensible design language system. It is a complex and labor-intensive process that requires several design iterations. Currently that process is shaped in large part by a series of rigid design tools available as different applications, which are loaded with features rarely needed at the same time. Design tools should instead adapt themselves to the designer's approach.

In the same way that modern web-based applications are developed with open architectures that adapts to user preferences (thanks to a modular structure), design software will be reduced to its constituent programs and delivered as interoperable services. A new kind of automated design platform—or orchestrator—will provide the designer with the needed tools at the right time in the creative process. Functions that are today available only in one form, and exclusively as part of software packages, will be served instead as individual and interconnected services. The predictive logic and machine learning capabilities of the "orchestrator" will surface, recommend, and combine design tools as needed.

Designers will assemble their own toolkit, which will act like an assistant constantly learning how to be more efficient at meeting the designer's needs.

5. Design Development: Autodevelopment

The design of digital products and services is one with software development. Coding is required to turn concepts into interactive simulations and prototypes, improve the design, and eventually make it real. Development frameworks and automated tools such as Macaw and Webflow aim to reduce the labor associated with common activities and are becoming increasingly sophisticated. Design-to-code algorithms will eventually support the conversion of a user interface design pattern or template into usable code across different frameworks. In the future automated tools will change the way products are built by allowing developers to write code, rules, and constraints at a high level of abstraction.

By exploring the areas of opportunity for automation, we will move toward a new era where designers will have unprecedented tools that are personal, smart, and able to assist with design work in novel ways. These new tools will not, however, change the fundamental nature of the designer's work; determining the right problem to solve, and identifying the design solution that best advances the human experience, requires distinctive human qualities.