How can data help resolve conflicts in systems design?

Data is essential in resolving conflicts in systems design, particularly with Generative AI. A key challenge is determining the initial extent of model training to support a business case. Early-stage design can benefit from strategies like implementing a proof of concept with a generic model API and a small dataset. This approach allows for validating the use case and assessing potential user benefits without significant investment. Based on these initial insights, more comprehensive training and investment can be directed towards larger datasets later. This method ensures decisions are data-driven, balancing model training needs against practical constraints, thereby aiding in navigating the complexities of systems design with precision.

Data is an intrinsic component of systems design. At the end the system design is a response to combined requirements of process, people and data pertaining to particular business objective/challenge to be addressed by the new system. As such data provides a fundamental cornerstone for the design - set of principles and anchors that describe elements of the business process, its inputs and outputs. It helps determine common, unambiguous set of definitions to be used during system design and consequent future use. Clarity of those definitions, principles and sets of initial values are fundamental in reducing potential issues during testing, deployment and run operations.

The cells of the human body have amazing intelligence and this intelligence causes the absorption of frequencies that will stimulate the cells. When external messages cause conflict and disruption, cellular functions will be disturbed and various diseases will arise. This rule also applies to data. Note that the fields must be defined correctly and fit the needs. The data must then pass through the right filters that detect conflicts with the existing data and be added to the database in order to be used correctly in decision making.

Los datos es uno de los requerimientos MAS importantes que buscan tener los usuarios "clientes" para la oportuna toma de desiciones, contar con las características de los mismos como: formato, veracidad, continuidad, seguridad, etc. etc. es información relevante para el dise?o de los sistemas, por lo que no es adecuado llevar a acabo el dise?o de un sistema sin cerciorarse que es lo que están buscando los usuarios o clientes, y más aún, a nosotros de TI requerimos de los mismos para la alimentación de otras posibles aplicaciones que tengan que ser retro alimentadas de el sistema que se dise?e.

By leveraging quantitative and qualitative data from user interactions, performance metrics and stakeholder feedback, designers can pinpoint specific areas of a system that may need improvement or reevaluation. This approach allows for evidence-based decision-making, reducing subjective disputes and aligning stakeholders' perceptions with actual user experiences and system performance.

System design does require continuous feedback loop to make it fit for purpose. By using agile approach in system design, every step of the process can be stress tested using selected data to obtain insights on its performance and evaluate its quality. You can also use pre-existing data sets to feed them into new system design and observe the outputs you get to compare and contrast with existing system performance. Designing comprehensive testing scripts and routines is as important as building system itself - using data to scope and size it, allows for more informed decisions to be taken during entire process.

Como lo mencioné en el comentario anterior, los datos de una forma u otra siempre nos sirven como retri alimentación ya sea directamente a nosotros, a los usuarios, clientes, o bien para retro alimentar a otros sistemas dependientes entre si mismos, por lo que es de suma relevancia verificar la oportunidad de creación de los mismos, su veracidad, formato y entrega garantizada a cada ente que los requiera.

Decision making is based on set of facts or opinions. To be able to make fact based decision one has to collect enough data and converge it into meaningful information necessary for discussion on the given subject. This way emotions and opinions are being, hopefully, eliminated from the decision equation and final outcomes is represented by the most likely result that can be achieved with given data sets. Moreover, data used in one decision process can be reused in the next one if content collected justifies its adequacy. Data before opinions and emotions shall be considered when evaluating system design and its fitness for purpose.

Hasta hace algunos a?os, aunque ya existían los MRP′s y ERP′s, realmente los fabricantes de estas herramientas olvidaban la generación de información, o bien que fuese fácil obtener los datos necesarios, hoy día no es sencillo, pero hay mas herramientas que nos permiten la extracción de la información necesaria, el punto es garantizar que sea provista en el momento en que se requieren, en lugar, el formato, medio para entregarse, periodicidad, etc. etc. es aquí donde TI debe hacer lo necesario garantizar este echo, y no solo eso, sino de igual forma asegurar que la información no se pierda ante cualquier desastre.

In use cases such as GenAI model learning is critical for trust and probability approximation of the resulting outputs generated. The more relevant data is being fed into the model the better (more natural) results are being generated by the system. This rule is fundamental for ethical reasons as well, as GenAI is not in a position to validate own outputs to satisfy truth test. Therefore to achieve improvement in system design one has to collect data along the way and catalogue it for reference and comparison purposes to see how resulting outputs are changing based on inputs provided. This is typical alpha-beta-prelaunch-launch sequence status of system design where different components are designed, tested and integrated together.