The 2023
State of MLOps

As AI moves into production, the business case for MLOps is growing. Here’s what data scientists, industry leaders, and professionals across industries are saying about AI Commercialization.

Download Report

Contact us




What does it take to achieve the speed and scale necessary to realize the value of AI operations? The answer is MLOps, the happy marriage of AI and the traditional DevOps model. Our researchers tapped ML, AI, and data practitioners and leaders across industries; the top five were B2B & B2C technology, manufacturing, retail, consumer packaged goods (CPG), and healthcare to produce this year’s State of MLOps Report. Discover in this report:

  • Does successful AI have a secret recipe?
  • How do large companies commercialize AI?
  • What are the challenges of data annotation? What is causing model failures?

Successful AI requires high-quality data.

The success of enterprise-grade AI projects depends not only on the data volume but also on data quality.

3 out of 5 professionals surveyed for our report consider higher quality training data to be more important than higher volumes of training data for the best outcomes from investments.

Data quality tops the list.

Our report revealed that some of the biggest challenges with ML production include data quality, expertise, and volume. Nearly 50% of the participants believed that lack of data quality or precision is the #1 reason ML projects fail. Lack of expertise was next in line.

Human labeling is critical.

95% of our survey respondents agreed that human intelligence is the key to their efforts, with 96% saying human labeling is critical for the success of their AI/ML data models. 86% of the respondents currently leverage human labeling at scale within their existing data pipelines.

Data annotation tools are NOT mature.

78% of the respondents indicated that finding the right data annotation solution is challenging, which is why 45% of the companies in the last 12 months have used four or more tools. It implies that the data labeling tooling industry is not yet mature enough to offer a robust solution.

Outsource data labeling for scalability.

91% of the respondents in the survey said that they rely on outsourcing because human labeling expertise and annotation are scalable.

They say that outsourcing saves time (53%), adds flexibility (50%), saves costs (48%), and boosts data accuracy (47%). 

Edge case resolution is foundational to Commercial AI.

Our survey reveals that proprietary data from edge cases is foundational to deploying commercial AI applications. 96% of the respondents said solving data edge cases is critical, and human intelligence is central to identifying, mining, and resolving those cases.


Human intelligence is a critical factor in commercializing AI. Organizations must find ways to seamlessly integrate human expertise with data labeling tools to audit, monitor, and handle edge cases.

A combination of technology, talent, and techniques to achieve high-quality data for AI/ML solutions will be the key to success moving forward.

Decoding AI Commercialization

Download Report

We have curated the experiences and expectations of AI, ML, and data leaders and practitioners across industries into a crisp report. Check it out.

About The 2023 State of MLOps Survey by iMerit

To arm enterprises for the future of AI production at scale, iMerit partnered with VentureBeat to gather insights from data scientists, industry leaders, and professionals across industries who are on the ground, shepherding AI products into the market.

The results make a clear case for the impact of MLOps on ML/AI projects. iMerit is a leading AI data solutions company providing high-quality data that power AI/ML applications for enterprises.

With end-to-end data labeling solutions across industries, including autonomous vehicles, agri-tech, geospatial, financial services, government, medical AI, and technology, iMerit employs more than 5,500 full-time data annotation experts in Bhutan, Europe, India, and the United States.