Methods

How AI, data and BI systems are designed

Publications on how AI, data and BI systems are designed and taken to production, with evidence from projects and dated sources. Each one states a problem and sets out the method: the phases, the decisions, the controls and the evidence. The evidence comes from this site; the state of the field, from dated sources; and the judgment is labeled as the author's position.

9 publications

Topic T1How AI should be brought into an organizationBringing AI into an organization that is just starting so that what gets tested reaches operation, people adopt it and its value can be shown beyond the pilot.Read the publication
Topic T2How an agent system should be structuredStructuring a multi-agent system that works in production, and being able to show, before and after putting it into production, that it works.Read the publication
Topic T3How an LLM should be made to answer with the company’s data, without making things upMaking the model answer with the company’s data and not hallucinate, and measuring that it does.Read the publication
Topic T4How an AI prototype should be taken to production and kept runningTaking an AI prototype to production and keeping it healthy, cheap and safe once it is published.Read the publication
Topic T5How AI and data should be governedGoverning AI on top of data governance that works: every piece of data with a meaning, an owner and a purpose, every AI system inventoried, assessed and monitored, and an organization able to prove it without claiming more than it has.Read the publication
Topic T6How the life cycle of a data project at scale should be structuredDesigning a data platform and the life cycle of what runs on it so that it grows through integrations without multiplying copies, without mixing clients and without taking untested changes to production.Read the publication
Topic T7How a BI solution people actually use should be designedDesigning a semantic model and the dashboards built on it so that whoever decides actually uses them: comes back to them in every decision cycle, trusts every figure and never has to ask how it was calculated.Read the publication
Topic T8How decisions should be made with modelsMaking decisions with models without fooling oneself: every model has to beat a baseline, be validated the way operations will face it, reach the decision maker with its limits and be monitored after delivery, and optimization has to be formulated so that its result can be verified.Read the publication
Topic T9How a process should be turned into an intelligent operationTurning a process into an intelligent operation without automating a bad process: modeling it before touching the technology, deciding what to automate and with which tool, and giving each decision the human control its risk calls for.Read the publication