--- title: Driving01 tags: Consultoria slideOptions: theme: white transition: 'fade' --- # Driving01: AI Maturity Model | AI Readiness Model Persona de contacte: Ricard Vela <ricard.vela@driving01.com> --- ## Models de maduresa Un model de maduresa (MM) es defineix com un **artefacte | eina de disseny**. En el seu origen (2009) estava dirigit a la millora de processos de software i posteriorment es va adaptar als sistemes d'informació, transformació digital, industria 4.0, etc. Per construir un MM es segueixen diverses fases: + Definició del problema. P.e. "*AI as an economic success factor*". Es defineix: + **On** es vol arribar i perquè. P.e. "adaptació a l'entorn", "optimització", "escalabilitat", etc. + **Com** s'hi pot arribar: estratègies **defensives** (*data as a single source of truth*: data quality + data governance + AI) i estratègies **ofensives** (*data-driven company*, canvi cultural, noves competències, orientació al canvi). Exemple: full de ruta de Gartner. + Estudi de les alternatives. En el nostre cas hi ha *maturity models, readiness models* i *capability models*. + A nivell corporatiu (grans empreses), els més importants són [Gartner](https://www.gartner.com/en/documents/3982174) (2020), [Intel](https://www.intel.com/content/dam/www/public/us/en/documents/white-papers/ai-readiness-model-whitepaper.pdf) (2018) i [Ovum](https://www.amdocs.com/sites/default/files/filefield_paths/ai-maturity-model-whitepaper.pdf) (2018). + A nivell acadèmic: [Yams 2020](https://timreview.ca/article/1399), [Alsheibani 2019](https://aisel.aisnet.org/pacis2019/46/), [Holmstrom 2021](https://www-sciencedirect-com.sire.ub.edu/science/article/pii/S0007681321000744). Les paraules clau que usen són (activities, analytics, data, decisions, ecosystem, knowledge, mindset/culture, organization, people, risk, strategy, technology) i el seu enfocament es poden resumir en aquesta taula: ![](https://i.imgur.com/DkYByZH.png) + Definir una estratègia de desenvolupament: poden crear-ne un de nou o fer evoluacionar un d'existent. + Desenvolupament iteratiu. + Identificar les dimensions sobre les que es treballarà de forma seqüencial però iterativa. Les dimensions més usades són: Culture/Mindset, data, ethics, organization, privacy, strategy, technology. + Determinar i definir els estadis de maduresa per cada una de les dimensions. El més normal són 5 nivells: novice, explorer, user, translator, pioneer. + Implementar. + Evaluar. ### Exemple AI Readiness Model: | | Level I | Level II | Level III | Level IV | Level V | | ---- | ---- | ---- | ---- | ---- | ---- | | Mindset | No AI friendly culture | Workforce discovers the benefits of AI | Evidence of an AI-friendly culture | Culture enables AI innovations | Employees boost AI innovation | | Data | No criteria for collection and structuring of data | Criteria for data infrastructure defined | Prototypical implementation of the data requirements | Data is largely collected and structured | Data collection and structuring optimised | | Ethics | No awareness of AI ethics | AI Ethics policies are evolving | AI ethics guidelines applied in single cases | AI ethics rules are widely established | AI ethics principles are holistically applied | | Organization | Structure and resources not aligned with AI | Creation of initial structures and resources for AI projects | Piloted structures and resources enable AI projects | Established structures and resources support AI projects | Structure and resources are optimised for AI projects | | Privacy | No awareness of data protection | Data protection is partially considered in AI applications | Privacy is taken into account by the AI teams | Data protection is internalised and widely applied | Data protection is fully integrated and considered | | Strategy | No AI vision and strategy available | Vision and Strategy are pushed internally | Vision is established and actions are defined | Strategy is clearly defined | Strategy is perceived as leading in the industry | | Technology | No application of AI Tools | Awareness of AI technology | AI technology is partially used | AI applications are adopted | Use of AI technology is standardised | ## Exemple Data Science Readiness model **Analytic Opportunities**: Considers new and existing use cases to apply data science to improve organization mission and operations. + Who in your organization is thinking about how data science can help it operate better, smarter, faster, or more efficiently? + How do employees notify leadership of ideas for operational improvements? + How are proposals for new analytic solutions reviewed for investment? + Is someone researching new capabilities that can drive better decision-making? + Who evaluates the effectiveness of ongoing analytics projects? + How do you capture lessons-learned from your ongoing analyses? **Data**: Considers opportunities to use new and existing data sets and better manage and govern data in support of data science projects. + Where do you begin when thinking about your organization’s data? + Does your organization have a vision for how to use its data? + Who is thinking about how to maximize the potential of your data? + Is your data regularly used to drive both strategic and operational decisions? + Do your employees all treat data the same way? + Does your organization have data definitions and categories for all collected data? **Analytic Techniques**: Considers the analytic tradecraft and techniques to be applied to generate insights from data. + Are you able to quickly and easily access the specific pieces of information your organization needs at any given moment? + Do your analyses typically include historical data and predictions for the future? + Does your organization know which analytic techniques to use to generate actionable insights from your data? + Are data results presented in a way that makes the conclusions obvious, or do they require explanation? + How confident are you in the statistical accuracy of the analyses on which your organization is making business decisions? **People**: Considers the set of human capital programs required to develop a talented and capable team of data science practitioners. + Do you know who the data scientists are in your organization? + How does your organization find, recruit, on-board, develop, and retain an analytics workforce? + Do you know where to place data scientists so that they are best positioned to help the organization answer its most critical questions? + Is your organization continually reviewing its analytics goals to understand its talent needs? **Technology**: Considers the optimal ways to use existing and new technologies including applications, data platforms, and infrastructure to perform data science projects. + Does your organization ingest and store data in a way that enables it to be easily managed and analyzed? + Does your organization have tools that allow users of varying skill levels to interact with data? + Do you put as much emphasis on the front-end user interface and visualizations as the back-end technology? + How does your organization think about, plan for, and integrate new and emerging technologies that may make certain analyses more efficient? **Culture**: Considers the set of mechanisms that communicate, share, and reinforce the value of data science across an organization to change the behavior of the staff + Has your organization’s leadership made it clear how data science plays a role in furthering your organization’s mission and strategy? + Do your organization’s policies and procedures promote employees’ active engagement with data science? + Is it an organizational norm for all employees to use evidence-based decision-making? + Does everyone across your organization, including non-analytics staff, agree upon and understand the value of data science? --- ## Data Factories vs Data Warehouses The data warehouse is a broken metaphor in the modern data stack. We aren’t loading indistinguishable pallets of data into virtual warehouses, where we stack them in neat rows and columns and then forklift them out onto delivery trucks. Instead, we feed raw data into factories filled with complex assembly lines connected by conveyor belts. Our factories manufacture customized and evolving data products for various internal and external customers. ![](https://i.imgur.com/Eoa2m39.png) As a business operating a data factory, our primary concerns should be: + Is the factory producing high-quality data products? + How much does it cost to run our factory? + How quickly can we adapt our factory to changing customer needs? To establish data **quality control** in our metaphorical factory, we could test at four points: + The raw materials that arrive in our factory. + The machine performance at each step in the line. + The work-in-progress material that lands between transformation steps. + The final products we ship to internal or external customers. --- ## Decision intelligence A practical discipline used to improve decision making by explicitly understanding and engineering how decisions are made, outcomes evaluated, managed and improved by feedback. (Gartner) A designed decision — for example, one presenting a series of action to outcome pathways — presents an opportunity to clearly identify where and how data or AI can support downstream pre-decision processes (e.g. modeling), or how to determine whether an action will have a desired outcome. * https://towardsdatascience.com/a-framework-for-embedding-decision-intelligence-into-your-organization-f104947651ae * O’Reilly Media, “AI Orchestration Enables Decision Intelligence,” Medium, 19 January 2021. [Online]. Available: https://medium.com/oreillymedia/ai-orchestration-enables-decision-intelligence-2a88d8306ac9. [Accessed 26th September 2021]. * Gartner, “Top Trends in Data and Analytics, 2022", Rita Sallam, 11 March 2022. GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved. --- https://martinfowler.com/articles/data-mesh-principles.html Data Warehouse ![](https://i.imgur.com/KeY0LtM.png) Data Lake ![](https://i.imgur.com/IWubzwT.png) Data Mesh ![](https://i.imgur.com/2ZVquKF.png) Data Products ![](https://i.imgur.com/lwcL685.png)