The enthusiasm for artificial intelligence (AI) as a source of solutions to problems is not new. In law, from the early 1980s until at least the early 2000s, considerable work was done on developing ‘legal expert systems.’ As the DataLex project, we participated in those developments, through research and publications, commercial and non-commercial systems, and teaching students application development. This paper commences with a brief account of that work to situate our perspective. The main aim of this paper is an assessment of what might be of value from the experience of the DataLex project to contemporary use of ‘AI and law’ by free legal advice services, who must necessarily work within funding and other constraints in developing and sustaining such systems. We draw fifteen conclusions from this experience which we consider are relevant to development of systems for free legal advice services. The desired result, we argue, is the development of integrated legal decision-support systems, not ‘expert systems’ or ‘robot lawyers’. We compare our insights with the approach of the leading recent text in the field, and with a critical review of the field over twenty-five years. We conclude that the approach taken by the DataLex project, and now applied to free legal advice services, remains consistent with leading work in field of AI and law. The paper concludes with brief suggestions of what are the most desirable improvements to tools and platforms to enable development of free legal advice systems. The objectives of free access to legal information services have much in common with those of free legal advice services. The information resources that free access to law providers (including LIIs) can provide will often be those that free legal advice services will need to use to develop and sustain free legal advisory systems. There is therefore strong potential for valuable collaborations between these two types of services providers.