Artefact AI & Analytics Squad: 10-weeks Proof of Concept

Artefact is a global data-driven services company specializing in consulting for data transformation and data & digital marketing. We help companies transform data into business impact by delivering tangible results across the entire value chain.

Our LATAM technical team has developed deep expertise in Azure since it's the cloud provider of our main clients and we have developed multiple projects in all data areas and industries. We are experts in the development of end-to-end Data Products using Azure, from the ingestion of data in Azure Data Factory to the data visualization in Microsoft Power BI, going through the data manipulation, data modelization and product industrialization in Azure Databricks. Most of our Data Engineers are equipped with the DP-203 Data Engineering on Microsoft Azure Certification. Also, Artefact helps our customers in getting started with Azure and we help them in configuring the clusters, the environment and managing the costs.

Artefact and Microsoft bring an end-to-end digital solution and our AI Data Factory methodology can deliver fully customizable Data products in 8~18 weeks divided into 4 main steps. However, the main objective of this Proof of Value is to deliver a Minimum Viable Product integrated into a Power BI visualization, in other words, a prototype of a product with a proven AI value by our Data Scientists and Data Engineer.

All the duration below is estimated and it could be longer or shortened. STEP 4 IS NOT INCLUDED IN THIS POC OFFER

1) Business Discovery & Framing - (2 weeks):

#Objective:

• Prioritize, understand and define the roadmap for one Business Use Case

#Activities

• Identify use cases with a proven potential to create value and a real chance of success from a technical point of view
• Confirm the business potential of the use case by quantifying its added value and validate its technical feasibility through preliminary analysis
• Identify data owners and receive access to the necessary data in the Azure Data Lake Storage, Azure DevOps and Azure Databricks;
• Define the governance and management strategy for use cases
• Define a high-level plan for the highest valuable Use Case

2) Most Valuable Model (MVM) design - (5 weeks):

#Objective

• Prepare, build, and test AI models to answer to respond the business question

#Activities:

• This is where we will make all the data ingestion and cleaning, data pipelines, and model selection using Azure Databricks and Azure DevOps as versioning control;
• Deliver incremental value to the user every week, from the beginning of the project, always interacting with business client teams;
• Ingest, prepare, and harmonize all the necessary data defined in the previous phase;
• Design and structure the technical environment for the use case;
• Build the models and algorithms that best respond to business decisions (MVM - Most Valuable Model). We use the native MLflow experiments on Databricks to track the Machine Learning models and increase their performance as the project goes on.

3) Minimum Viable Product (MVP) design - (3 weeks):

#Objective:

• Prove the business value of the model through the analysis of the models

#Activities:

• Define the key performance identifiers to evaluate the model;
• Run tests in the “real world” to assess the effectiveness of the model;
• Track tests KPI graphics using Databricks Dashboards.

4) Product Industrialization: (NOT INCLUDED IN THIS POC OFFER)

#Objectives:

• Scale the scope of impact for each MVM.
• Scale the Digital Products to activate and make the complex algorithms transparent for the business to accelerate day-to-day decisions.

#Activities:

• Structure and automate the data process (ETL) to capture, iterate and accelerate the execution of models. We automate the MVM using Databricks Workflows. It provides us with a bird's-eye view of each run, failures and Git integration;
• Escalate the impact scope of the created product;
• Design and create digital products for large-scale activation;
• Scale the product to business areas to accelerate day-to-day decisions.

Artefact Analytics Squad References:

• Global Beer Company: strong partnership counting with more than 15 Data Specialist to deliver +10 end-to-end data projects at scale
• Global CPG Company: the creation of an AI Factory Data Hub in Brazil to deploy multiple AI projects at scale
• Global Retail Company: the creation of an AI Factory Data Hub in collaboration with Google to deploy multiple AI projects at scale