Highlights:

  • Capital One’s study shows that most data management decision-makers face key operational roadblocks that could slow down ML deployment.
  • The survey shows that people in charge of data management believe that AI and machine learning can help their businesses grow.

A new study by Forrester Consulting, commissioned by Forrester Consulting, has brought to the fore the biggest challenges, worries, and opportunities faced by organizations when using Machine Learning (ML) to enhance business performance across the enterprise.

In times when corporations are heavily investing in and giving priority to ML deployment, Capital One’s study shows that most data management decision-makers face critical operational roadblocks that could slow down ML deployment, including transparency, traceability, and explainability of data flow (73%) and breaking down data silos between internal departments (41%).

Dave Kang, SVP and head of data insights at Capital One, said, “Businesses see massive potential in applying machine learning but encounter headwinds in their data.” He added, “This can hinder businesses from seeing actionable insights and perversely shy away from adopting and operationalizing ML solutions in the first place.”

Data problems for machine learning

Breaking down data silos is yet another big problem for data managers. More than half (57%) say that internal silos between data scientists and practitioners slow down ML deployments, and 38% say that data silos within the organization and external data partners slow down ML maturity.

Other top challenges include:

  • Working with big, messy, and diverse data sets (36%)
  • Problems turning academic models into products that can be deployed (39%)
  • Minimizing the risks of artificial intelligence (AI) (38%)

Despite these issues, the data shows that ML adoption is on the rise, with almost 70% of executives planning to increase the use of ML across their organizations. In the next three years, top ML deployment priorities include automated detection of anomalies (40%), getting automatic transparent updates to applications and infrastructure (39%), and meeting new regulations and privacy requirements for ethical and responsible AI (39%).

Believing in the promise of ML

The survey shows that data management decision-makers believe that AI and machine learning can help their businesses grow. However, to keep improving their ML applications, these people need to break down barriers between people and processes.

They also need to find better ways to turn academic models into deployable products that can be used to illustrate a better return on investment to executives. By working with partners with first-hand experience and staying focused on the business promise of ML, decision-makers can show the key results of operationalizing ML, like efficiency, productivity, and a better customer experience (CX), to executive leadership.

Capital One hired Forrester Consulting to do the study. They asked 150 data management decision-makers in North America about their organizations’ goals, challenges, and plans for putting machine learning to work.