My Reflections on Experience Design in the Industrialization of AI
-
At the beginning of 2018, I joined the AI department of my company, responsible for experience design in multiple AI implementation areas such as AI retail, AI office solutions, and facial recognition.
AI projects differ significantly from internet-based ToB products in terms of industry chains, users and scenarios, product forms (primarily AI ToB integrated software-hardware products), and experience carriers.
Despite my extensive experience in design, my background in mobile, PC, and web design left me somewhat at a loss when tackling these projects. The design team went through a prolonged exploratory phase before gradually developing a mature methodology, continuously expanding the boundaries of experience design.
Compared to internet-based ToB products, the industry chain is longer (Figure 1), involving more R&D roles, necessitating the establishment of efficient collaboration processes.
The inclusion of external hardware partners requires precise communication of experience standards and strict quality control in deliverables.
The diversity of B2B merchants and the richness of C2B application scenarios demand higher capabilities from designers to identify core issues across complex scenarios and user groups.
Figure 1: Comparison of Industry Chains Between AI ToB Integrated Software-Hardware Products and Internet-Based ToB Products
Product forms include software products, integrated software-hardware products, solutions, and data services, often with high business thresholds. Designers may lack familiarity with users and scenarios, exceeding their existing cognitive abilities (e.g., developer habits, potential hardware integration issues).
From the lifecycle diagram of AI ToB integrated software-hardware products (Figure 2), it’s evident that business objectives vary at each stage, as do design goals.
Figure 2: Changes in Business and Design Objectives Across the Lifecycle of AI ToB Integrated Software-Hardware Products
Obtaining user characteristics and scenario information is challenging, and accessing ToB competitors is difficult, adding complexity to design work.
End-to-end design proves more effective in addressing products that span online and offline, involve multiple roles, and feature numerous touchpoints.
As mentioned, AI implementation involves multiple business nodes, spans online and offline, and includes diverse roles, each presenting unique challenges and design goals. This expands the boundaries of experience design for AI ToB integrated products, extending beyond core software and hardware design to encompass full business chain design.
For example, one product involves seven key stages from R&D to deployment (Figure 3), each with distinct user roles and objectives, requiring targeted design efforts.
Figure 3: Analysis of Key Focus Areas in the Business Chain of AI ToB Integrated Software-Hardware Products
Many stages in AI implementation remain uncharted territory for design, lacking experience-driven guidance. This offers designers greater opportunities to identify issues, uncover opportunities, and contribute to business goals through design.
B2B product design thus becomes highly diverse, covering software experience design, hardware industrial design, software-hardware interaction design, and even service processes, sales materials, and post-sales workflows.
Take the commercial promotion/negotiation/pre-sales phase as an example. The goal of experience designers is to support sales efforts, with design scope including but not limited to demo presentations, branding, sales materials, and exhibition design. To ensure optimal AI performance, designers may also provide experience standards for software/hardware deployment to clients with limited development capabilities.
Full business chain design means designers assist in achieving commercial objectives through end-to-end design, shifting from 'user-centered' to 'audience-centered' design.
Note: 'Audience' refers to different stakeholders at various stages, such as corporate decision-makers or trade show attendees.
In AI ToB integrated projects, most requirements are built from scratch, making user research critical for market feasibility. The inclusion of human factors, hardware, and structural design expands roles beyond traditional interaction/visual design, redefining responsibilities (Figure 4).
Figure 4: Evolution of Design Team Responsibilities in AI ToB Integrated vs. Internet-Based ToB Product Development
Unlike internet designers focused on interaction and interface design, experience designers in integrated projects address broader scopes, anchored in two core principles: 'goal-oriented' and 'audience-centered.'
Experience designers must flexibly apply 'user-centered' methods and principles to solve problems in AI industrialization, adapting to 'audience-centered' and 'goal-oriented' approaches.
These are my humble reflections as an experience designer in the AI field. The AI exploration process is full of possibilities, demanding flexible use of existing knowledge—painful yet fascinating. It has revealed a broader horizon for experience design, and I look forward to sharing more insights in the future.
Note: Some details in the diagrams are emphasized, and the processes are still under validation. Feedback and discussions are welcome.