On January 23, 2023, McKinsey announced its acquisition of Iguazio, a provider of AI & machine learning operationalization (MLOps) software. With over 70 experts in data and AI, the addition of Iguazio's technology will enable leading consultancy firm McKinsey to accelerate and scale AI deployments for clients. Thus, the play is not (just) about putting more software under McKinsey’s belt.
McKinsey's AI subsidiary, QuantumBlack, has been assisting clients in embedding AI into decision making for over a decade but has struggled to deploy and scale AI effectively. The acquisition of Iguazio will provide McKinsey's clients with AI software solutions and a team of data scientists to supercharge their AI efforts. According to McKinsey, the Iguazio and QuantumBlack teams will be integrated, creating a single product that allows real-time integration of AI in decision making. This acquisition is McKinsey's first in Israel, with plans to expand the QuantumBlack location in Tel Aviv.
AI software + AI people = AI success
McKinsey research shows that only 10% of AI projects succeed in real business environments, despite a global investment of $490 billion from 2012 to 2021.
But have faith and do not despair.
Other studies, such as a recent one from ClearML, indicate that businesses are beginning to put more thought and money into operationalizing their machine-learning models. For example, in ClearML’s survey, 85% of respondents said they had a dedicated MLOps budget in 2022.
This acquisition speaks to an important truism when it comes to deploying AI: it isn’t easy. As the above data shows, companies struggle to put their machine-learning models into production despite their initial enthusiasm.
When it comes to AI, a winning combination is bringing powerful software together with data professionals. In this case, McKinsey bought a software company to aid in their deployment of AI for their clients. Sometimes, like in the case of ServiceNow’s acquisition of Element AI, it can be the reverse. In both instances, the truism remains: bringing together the right software with the right people can help businesses gain insight into their data and turn it into (data) gold.
What G2 data shows regarding MLOps software adoption
At G2, we have seen that adoption of MLOps software, while low, is still on the rise, as seen in the chart below. With more people on the job and more budget commitment to getting models into production, we predict that adoption will continue to rise.
Figure 1: Adoption of MLOps Software, based on G2 reviewers being asked: “What percent of your users have fully adopted [the software]?”
McKinsey’s acquisition is a microcosm of a more significant trend. At G2, we have seen the importance of having the right people and software when deploying ML models. McKinsey recognizes this and is showing the world how to make it a reality.
Edited by Shanti S Nair