50 semi-structured interviews conducted with legal services professionals in England over the period 2019-20. Interviewees were drawn from law firms, corporate legal departments, and alternative legal services providers. The interviews explored the opportunities for, and constraints experienced on, the adoption of artificial intelligence in interviewees' legal services work and organizations.Our motivation is to help chart a path toward comparative advantage for the UK's legal services sector by unlocking the potential for AI in legal services. Our objectives are to make progress in understanding, and relaxing where possible, three types of constraint on the implementation of AI: First, the need for complementary investments; second, the need for legitmacy in the application of AI to law; and third, the limits of technological capability. More specifically, our objectives are as follows: 1. Complementary investments in human capital: To understand the training and skills needs for lawyers and computer programmers working with AI applications in the legal sector, and to design, jointly with private sector partners, education and training packages that respond to these needs for delivery by both universities and private-sector firms, details of which will be made available for adoption by other organisations across the sector. 2. Complementary firm and professional structures: To understand how firm-level governance choices and the structure of legal professional knowledge affect the implementation of AI technology in the legal sector in a way that contributes not only to academic scholarship but also provides best practice guidance for firms and professional bodies. 3. Complementary policy choices: To understand differences in labour mobility, skills and technology transfer in high-value services sectors between the UK and principal competitors, and derive policy implications for the UK. 4. Legitimacy: To map how the various possibilities for legal challenge of automation in dispute resolution relate to types of dispute resolution technology and types of claim; to use this to identify opportunities for maximum benefit from use of automation in dispute resolution. 5. Technology: To push outward the frontier of what is techologically possible in the application of AI to legal reasoning through use of ontologies, knowledge bases, and semantic systems, so as to increase the range of possible applications.

Interviews conducted primarily face-to-face with a smaller number remotely. Interviewees were sent a list of topics for discussion in advance of the interview. Interviews focused on these topics but were only semi-structured so as to permit discussion of other topics raised by the subjects.