Knowledge driven chatbots

About

Overview Chatbot Author is based on our original thinking of transcending rule-based knowledge from the intangible characteristics of content into tangible, working and measurable knowledge assets. This involves a circulatory process consisting of create, share, measure and evolve. Create Our aim is to enable others to simplify and streamline rule-based knowledge complemented by an audit trail for transparency, traceability and new forms of measurements. The means for achieving this transition is a shift from the monologue of content to the dialogue of rule-based knowledge, which aligned with published futurist thinking . Our long-held belief, is that the target state for transformation requires software to be developed in the form of chatbots. Rule-based knowledge, as the targeted market, is so large that there was serious doubt that the 23m software developers worldwide have sufficient spare capacity to take on such a quest, especially as there are so many other competing priorities. Our novel alternative was to use analysts / consultants, now known as Citizen Developers that have the skills to deconstruct and reconstruct the knowledge. The barrier to this transition involved the traditional design and processing of data driven solutions, which always has been a problematic challenge for end user software development. To overcome this challenge, we decided to look at the rule-based knowledge differently. In essence, the documents containing rule-based knowledge is in content form and does not have any dependency upon a database for the narrative to be read and understood. The document simply contains knowledge. A good case in point, is that the regulations have no dependency upon a database. This led to an important guiding principle, which was that the solution required to be knowledge-driven and not data-driven. Being knowledge-driven fundamentally challenges the convention of software development, which has evolved over the past 60 years. Consequently, we developed the notion of Knowledge Maps. These are quite distinctive as they are not driven by data, thus removing the complexities of building apps. The nearest correlation to our thinking we could find was by Stephen Wolfram’s New Kind of Science whereby it was discovered nature uses simple programs to handle complexity. By implication nature does not use databases it uses simple programs. Wolfram’s insights concluded complicated programs cannot cope with the complexity of constant change, whereas simple programs can. This underlying revelation is counter intuitive to the traditional thinking of technologists. Our Knowledge Map provides, in essence, the means for simple programs to be generated. Naturally, a Knowledge Map is underpinned and bound by the algorithmic nature of choices, pathways and outcomes. It should be made clear that simple programs should not be confused with the knowledge being simplistic. Naturally, this led to the issue that not all knowledge could be captured in one map. This was overcome by providing the means to dynamically link Knowledge Maps. The dynamic linking is based on each Knowledge Map being represented as a self-contained, stateless software object . The ability to dynamically link stateless software objects meant we could cope with any permutation of rule-based knowledge complexity, whilst masking the complexity through path-driven dialogue. Using Knowledge Maps provided the basis for automating the building of the software as a stateless object. It is stateless as there is no dependency upon the need to process data. The dynamic linking of the Knowledge Maps enables a loosely coupled ecosystem of any size to be developed and evolved. This approach reflects more accurately knowledge, which is in a constant state of change. Share The generated software representing a Knowledge Map is designed for dialogue interaction using scripted dialogue. This means a person is guided through the knowledge fabric via choices, which drives the selected pathway to reach the best fit outcome. Measure Every interaction is captured as a dialogue-step, which provides the basis for new forms of measurements and analytics. Evolve Using the emergent evidence generated from the captured dialogue-data provides the real-time feedback loops for enriching and extending the Knowledge Maps.