The job description for an AI service designer isn’t yet well defined. In brief, an AI Service Designer is somebody who uses the tools of Service Design to design services powered by artificial intelligence.
To perform this role effectively, they need a solid understanding of Service Design process and an ability to facilitate this for a variety of stakeholders. They also need to understand how to apply Artificial Intelligence technology effectively. For example, they can help ensure AI has sufficient high-quality training data to operate on, and that the AI’s output is used appropriately, including handling different types of failure.
To help us understand the role of an AI Service Designer, let’s look at how the role is positioned in relation to other related roles that they typically collaborate with. In the table below, I illustrate where I take a lead, versus areas where I play a supporting role.
| Phase | Steps led by an AI Service Designer | Collaboration with other roles |
| Discover | I identify and map the interactions between key stakeholders, inside and outside of the business, e.g. customer, staff, suppliers. I will identify what data is being collected, any limitations on how it can be used, its completeness and selection bias. (For AI projects, this is crucial) I interview and conduct activities with the stakeholders to reveal their needs. I build an understanding of how their needs might be in tension, e.g. a marketing professional’s desire to capture, model and personalise every customer interaction vs. a consumer’s desire for privacy. | Stakeholders – It’s essential to engage key stakeholders from an early stage, because nobody wants to feel their job is being automated away, or reduced to labelling data so that AI can perform the “skilled” work. Also, stakeholders often have deep industry knowledge that they are happy to share. In return, I offer them curiosity and a ‘fresh pair of eyes’, which can help them to acknowledge that existing processes can be improved. |
| Define | Supported by AI tools, I synthesise and communicate a clear overview of the stakeholder needs. I also broadly outline how current or emerging AI technology might be able to meet these needs, using the data we have access to. I also mention any major risks that should be considered. This helps to encourage leadership to not prematurely discard opportunities that they assume would be infeasible, but are actually made possible by AI. | Leaders – It’s up to the leaders of the business to decide which opportunities to pursue, based on their deep understanding of their business, its goals and its place in the marketplace. My role is to ensure that their decision is well-informed, both by a clear understanding of the stakeholder needs and of AI’s technical capabilities and limitations, both now and in the future. |
| Ideate | I facilitate brainstorming and other idea generation activities with a diverse group of professionals, to elicit numerous ideas of potential solutions. I also contribute my own solution ideas. I provide inspiration by sharing examples of how AI is solving similar challenges in other contexts, to encourage participants to consider these options even if they’re not yet familiar with the technology. | Diverse stakeholders and experts in the problem and solutions, including technologies |
| Select | I facilitate a feedback and selection process, and give my advice on the feasibility of AI technologies involved, including highlighting risk factors, ethical and practical considerations. | Diverse stakeholders and experts to give opinions Leaders and team members to make a decision on which solution(s) to implement and how. |
| Prototype or develop | I support the team in implementing a solution using AI technology, drawing attention to how an AI project needs to be approached differently. I help them focus development on the most important stakeholder needs, while managing the probabilistic and constantly-evolving nature of AI technologies. | UX designers – I help them adapt to designing for AI, which inherently involves significant variation and unpredictability. This requires significant changes to their normal process and artefacts. AI engineers and software developers – I ensure they have a clear understanding of the stakeholder challenges we’re addressing, and have considered common failure cases of AI systems. Data scientists – I ensure that they have a clear understanding of how the data was collected, any risks of bias or other data quality concerns. Also I check that they understand how the data can and will be used. QA Testers – I help them adapt their test methodology, which can be particularly challenging when using AI, and often requires using AI tools to test the AI solution. Team leaders – I work with them to establish new human processes that take advantage of the new AI capabilities and also manage its limitations. Users and trainers – I help users to understand the basics of how AI works and how to use it effectively. This helps them to feel empowered by the technology instead of threatened by it, and to trust it appropriately. |
| Test and monitor | I implement a process for initial and ongoing evaluation of the AI system. As AI systems are constantly learning, performing a round of user tests may be insightful, but is not sufficient. Processes must be implemented for continual monitoring of the AI system, and for dealing with failure cases. | Team leaders, particularly of customer and user support teams – I help them establish processes for supporting the AI system Leaders – I establish processes (such as metrics and dashboards) to give leaders an accurate overview of how the AI system is performing, to inform their decision making. |
If you’d like to discuss how an AI Service Designer can support your team, get in touch!