Scaling Analytics & AI: Breaking Free From IT Roadblocks

Scaling AI Catie Grasso

In June 2023, Dataiku and Databricks surveyed 400 senior AI professionals in large companies around the world and — given the critical role that IT leaders play in helping organizations scale AI — we wanted to share a few of the insights specific to the respondents in IT and information security departments. 

Please note that while IT leaders made up a small sample of the total number of respondents (n=87), we feel like these takeaways are still important to share. So, what are the top barriers preventing organizations from delivering more value from analytics and AI, from the lens of IT leaders? How can they rise above the chaos and deliver data and AI products without sacrificing control? Find out here.→ Read Now: AI, Today: A Survey Report of 400 Senior AI Professionals


1. 61% of IT Leaders Struggle to Operationalize and Iterate on Analytics & AI Projects

Operationalizing and iterating on analytics and AI projects is critical for successful scaling, but many IT leaders struggle in this area. One primary reason fueling this is the lack of alignment between IT and business teams. Often, data experts (alongside IT leaders for oversight) develop analytics and AI solutions that do not meet the specific needs of business users. This disconnect can lead to a lack of adoption and trust in the generated insights.

From a methodology standpoint, operationalization requires a consistent, efficient process for integration, testing, deployment, and then measuring impact and monitoring performance (followed by, of course, making necessary modifications and integrating, testing, deploying, etc., those modifications). Inconsistent packaging and release can lead to a subtle degradation of a model’s performance between development and production. 

Traditionally, it’s the data engineering or IT team that is responsible for refactoring of the data product to match target IT ecosystems requirements (including performance and security). However, this handoff between data team and IT or data engineering teams can be significantly eased when the two are working with the same tools and are aligned on project goals — so again, communication (even between technical teams) is key. 

⚡Check out this blog for more on how IT can help prevent AI project failures.

2. 55% of IT Leaders Lack Quality Data or the Ability to Easily Grant Access to the Right (Quality) Data

Quality data is the lifeblood of successful analytics and AI projects. Unfortunately, many organizations struggle with data-related challenges that hinder their scaling efforts. Some common issues include:

Data Silos: Disparate data sources scattered across various departments can make it difficult for IT leaders to access and integrate data efficiently. Siloed data inhibits the ability to gain a comprehensive view of the organization, leading to incomplete or inaccurate insights.

Data Governance and Security: Poor data governance and security practices can compromise data integrity and raise concerns about data privacy. Ensuring data is handled in a compliant and secure manner is essential to building trust in analytics and AI projects.

Data Accuracy and Completeness: Low-quality data, such as duplicate records, missing values, or outdated information, can lead to erroneous conclusions and decision-making.

To address the lack of quality data, IT leaders should focus on implementing robust data governance frameworks. This involves collaborating with data executives (i.e., CDOs) to create data policies, define data ownership, and establish data quality standards. 

Furthermore, data cleansing and enrichment processes should be put in place to enhance the quality of existing data. Regular audits and monitoring of data pipelines are necessary to ensure that data remains accurate and up-to-date. After all, data pipelines are what keep analytics and AI projects afloat — you can have the best AI strategy, but without the right data, projects won’t succeed. 

Interestingly, our survey revealed that over half (57%) of senior AI professionals in the IT or information security departments agree that they have clear owner(s) responsible for data quality. Even if IT stakeholders aren’t in charge of data quality and it falls under the remit of a data executive, they can orchestrate data quality during complex ingestion tasks and delegate data quality tasks to the ones who know the data best (the business departments) for correction in a decentralized approach without losing control

⚡Check out this interactive site on how to prevent data quality issues from fully halting analytics and AI initiatives with Dataiku.

3. 45% of IT Leaders Lack the Right Data Talent or Data-Literate Workforce

To bridge the talent gap and foster a data-literate workforce, IT leaders must prioritize upskilling initiatives along with collaboration with the business. According to our friends at Capgemini, IT organizations can foster collaboration and un-siloed environments by:

  • Encouraging cross-functional collaboration: IT leaders should facilitate collaboration between data scientists, engineers, domain experts, and business stakeholders, encouraging regular communication, establishing interdisciplinary teams, and promoting knowledge sharing. This aligns with our survey findings: 76% of IT leaders agree that their advanced analytics teams are interdisciplinary (i.e., business people working with data people).
  • Establishing agile workflows: Agile methodologies, such as Scrum or Kanban, enable IT teams to break down projects into manageable increments, allowing for iterative development and quicker time-to-market. By adopting agile workflows, organizations can adapt to changing requirements and deliver AI solutions that align with business needs.
  • Investing in AIOps skill development: Upskilling employees is crucial for scaling AI initiatives. IT organizations should provide training programs and resources that enable employees to acquire the necessary AIOps skills. This investment in skill development ensures that teams can effectively contribute to AI projects, fostering a culture of continuous learning and innovation to deliver the value.

⚡Check out this interactive site on what has worked best from Dataiku customers around the world when it comes to upskilling. 

From Challenge to Triumph: Overcoming Challenges to Scale

Scaling analytics and AI projects is a complex endeavor that demands careful attention to operationalization, data quality, and workforce capabilities (among other challenges). IT leaders play a pivotal role in overcoming these barriers to ensure the successful deployment and adoption of data-driven solutions.

By fostering collaboration between IT and business teams, adopting agile methodologies, and promoting a data-driven culture, organizations can enhance their ability to operationalize and iterate on analytics and AI projects. Additionally, prioritizing data governance and implementing data cleansing processes can address data-related challenges.

Lastly, developing a data-literate workforce through targeted upskilling and cross-functional collaboration will empower organizations to fully leverage the potential of analytics and AI, driving innovation and competitive advantage in today's data-driven landscape.

You May Also Like

Democratizing Access to AI: SLB and Deloitte

Read More

Secure and Scalable Enterprise AI: TitanML & the Dataiku LLM Mesh

Read More

Revolutionizing Renault: AI's Impact on Supply Chain Efficiency

Read More

Slalom & Dataiku: Building the LLM Factory

Read More