Dataiku 10 empowers everyone from IT operations and domain experts to data scientists and risk managers

Dataiku has introduced new AI governance and oversight as part of the company’s unified AI platform to allow organizations to scale analytics and AI initiatives under one centralized control tower.

  • Tuesday, 7th December 2021 Posted 2 years ago in by Phil Alsop

Dataiku 10 features a built-in suite of tools to help IT operators and data scientists automatically evaluate, monitor, and compare models under development or in production. In addition, also added in Dataiku 10, organizations can deliver value faster with packaged industry solutions, dedicated workspaces for business users, and accelerators for exploratory data analysis, geospatial analytics, and computer vision. Examples of newly available Dataiku solutions across industries include market basket analysis, product recommendations, plant electricity and CO2 emission forecasting, and real estate pricing, with more in the works.

 

With the rapid adoption of AI, AI governance has become a major pain point for every industry, as it was never designed to manage the copious amounts of data we see today or the democratization of machine learning (ML). From CMOs to CFOs, data is crucial to any organization’s ability to allow employees in all departments to make impactful business decisions. However, without a modern data governance strategy, today’s organizations are opening themselves up to unnecessary risks.

 

“The new risk and value assessment capabilities in AI governance plans allow businesses to assess and manage projects across multiple divisions, which now brings together IT leaders, product managers, operations and more. These departments have historically worked in silos, but data has made it imperative to break down these silos,” said Ritu Jyoti, group vice president, AI and Automation Research at IDC. “As all industries have scaled usage of AI and ML, it’s no longer the sole domain of data scientists and IT managers, but a requirement of the C-suite and more to be able to access and understand. The only way to make this happen is to have one simple language for how departments manage MLOps, creating a consistent framework and approach for every project.”

 

“We’ve always believed that to scale AI, organizations need to enlist non-experts to the cause and bring more people into the fold to ensure project success. Dataiku 10 helps make that a reality,” said Clément Stenac, CTO and co-founder at Dataiku. “This latest version is focused on governance, MLOps, and industry solutions that increase involvement from AI-adjacent roles such as IT operators, risk managers, project managers, and domain experts.” 

 

Today’s announcement is the most significant update to the Everyday AI platform, focused on three core functionalities and themes:

 

1.         MLOps - Safely scale with oversight for analytics and AI initiatives

Dataiku 10 introduces enhancements to its MLOps suite of tools to help IT operators and data scientists evaluate, monitor, and compare ML models, whether under development or in production. Automatic drift analysis and enhanced what-if simulations give teams better visibility into the behavior and performance of live models.

 

2.         Governance and oversight

AI Governance in Dataiku provides a dedicated watchtower in Dataiku where AI program and project leads, risk managers, and key stakeholders can systematically govern projects and models and oversee progress across the entire AI portfolio. Customers will be able to see all their models in a central model registry — regardless of whether they were developed natively in Dataiku or externally using tools like MLflow. Structured frameworks for project workflows, approvals, and project qualification provide superior AI oversight for enterprise customers. 

 

3.         Deliver value faster with business solutions and accelerators 

Organizations can accelerate speed to value by leveraging our off-the-shelf projects for various vertical use cases that customers can adapt and apply for their own purposes. Examples of advanced tools include new geospatial analytics, native deep learning capabilities, assisted data exploration, and enhanced visual interactive insights.