How We Work
We focus on developing & deploying interpretable AI solutions that are transparent and trustworthy.
AI solutions builds on sound business understanding.
SMAI Focuses on Interpretable AI
We are an AI consultancy with a focus on tabular data and interpretable machine learning. In order to create value and insights with AI it is important to:
- Deploy AI solutions in systems and ensure models and data are continuously updated and available for use
- Succeed in integrating the solution within the organization
- Provide accurate and understandable end-user output
- We believe AI should be transparent in order to succeed
- The subject should have a right for an explanation that is trustworthy
- The solution should be robust against bias in the data
Interpretability is when users and businesss can understand the reasoning behind predictions and decisions made by the AI model.
SMAI Development
All solutions we develop are inherently interpretable such that a full explanation always is available which fulfills several purposes such as; building trust for the AI model, the right for an explanation for a subject, highlighting potential bias in the data, validation of the AI model and data and providing insights.
We structure your data and provide insights from historic and current data to predict the future through a fully interpretable AI approach.
Our development phase is defined by 3 key steps
Extract Data
Validate & Clean Data
AI Modeling
Data and business insights
Validated tabular data
Interpretable and trustworthy
SMAI Deployment & Implementation
We are experts in deploying AI solutions into systems and providing the end-user with insight and an understanding of the output.
Microsoft Azure Cloud
The interpretable AI model is a vital part of the implementation since the end-user needs to interpret the output and insights from the AI model. Without the interpretation in the implementation the value of the AI solution diminishes which is why business knowledge is the heart of the solution.
AI creates insights and value for solving business problems when it contains business knowledge, has high precision and interpretability. Implementation success relies on a solution that is trusted, understood and usable for the end-user running in a production environment where the technical deployment ensures models and data are continuously updated and available for use.
Our Deployment phase is defined by 2 key steps
Deployment
End-user Solution
Solution running in production -
continuously updated
Usable for the end-user
Our method
Iterative
Weekly meetings or check-ins
SMAI Consulting has a proven method and years of experience in working with and deploying AI solutions successfully in businesses.
We follow 6 iterative steps we believe are crucial in succeeding.
1.
Business Understanding
AI solutions builds on a sound business understanding and the process is iterative. Weekly meetings or check-ins.
2.
Idea Generation
Good hypothesis are crucial to locate the underlying causes. We translate your valuable business knowledge regarding the business problem and what might affect it into data and AI.
3.
Data Access, Transformation and Cleaning
We extract the relevant data, transform it and make it ready for an AI solution.
4.
Interpretable AI Model Development
We are experts in developing AI models that are interpretable and production-ready.
5.
Validate AI Model and Data
Knowing and trusting the solution will guide you in the right direction.
6.
Present Results
SMAI Consulting is strong in presenting and clearly communicate complex material to non-technical users and collaborating with stakeholders.
We work with Microsoft Azure Cloud
Create and scale
- Resources needed can be created in a few seconds
- Scale up when compute power is needed. Scale down when it is not
Infrastructure as Code (IaC)
- Pipelines from Azure DevOps orchestrates the resources in the Azure portal
- Utilizing infrastructure as code (IaC)
AI & developer independence
- Azure Machine Learning made for developing machine learning and running in production
- VM (Virtual machine) on Azure makes the data scientist (developer) independent of their computer and setup
- Docker technology integrates naturally in Azure
Data
- The platform integrates with both traditional and modern data warehousing (BI)
- The solution can integrate with your current BI-setup or run on the side as the first step into Azure
- Data never has to leave the tenant of the organization
State of art security
Interested in how SMAI Consulting can help your business
Email
Phone
LinkedIn
Website
Address
CVR