By using website you agree to our use of cookies as described in our cookie policy. Learn More

TDWI Upside - Where Data Means Business

AI Readiness with Fern Halper

TDWI’s Fern Halper explains what enterprises need to do to prepare for AI and introduces initial results from TDWI’s AI Readiness Assessment.

Fern Halper, vice president and senior research director for advanced analytics at TDWI, joined the latest Speaking of Data podcast to discuss AI readiness, specifically recent trends seen in enterprises using an assessment tool that is freely available from TDWI. [Editor’s note: Speaker quotations have been edited for length and clarity.]

For Further Reading:

Lessons Learned from Facebook’s Poor AI Implementation

Tackling Information Overload in the Age of AI

How to Find the Right AI Solution: 3 Innovative Techniques

“AI is on everyone's mind, and with the advent of generative AI, there's a lot of pressure on organizations to get ready for AI,” she began. In TDWI surveys, generative AI has recently been ranking higher than machine learning as a priority for organizations. Self-service BI and other projects are still priorities, but a top priority in 2024 is generative AI.

“It's important to put a data foundation in place and get experience with machine learning before you start building generative AI applications. Many organizations are starting to see that it's not just sort of easy peasy -- you build a prompt interface and you’ve built a generative AI application. Organizations want to use it against their own data, which means their data foundation needs to be in place.”

Halper noted that generative AI has triggered people to think about AI in general, along with more traditional machine learning, natural language processing, and related technologies, and that’s driving a need to determine if your organization is ready for AI. Thus was born the TDWI AI Readiness Assessment

“The way we thought about AI readiness wasn't just around whether you have the necessary algorithms. We were thinking about different dimensions for AI readiness. Is your organization ready for AI? Do you have executives who support AI? Do you have the funding, resources, and culture in place?

“Then there’s the issue of your data -- is it ready for AI? Do you have the right platforms to deal with diverse data that's going to be part of AI -- not just structured but unstructured data? What does that data foundation look like? Are you operationally ready for AI? Once you start building these models, they're going to go into production, so you need to be ready operationally and have teams in place. Do enterprises have the skills needed? Are they familiar with the algorithms and can they make use of them? If you're going to be developing applications, do you have the skills for that?”

Halper explained that many organizations want to build apps with AI, but they’ll need to make sure they have effective governance in place. “Are you prepared from an AI governance perspective or are you even thinking about issues that are important for AI, such as ethics and the responsible use of AI? What about bias? We wrapped the whole thing together and created this assessment that consists of about 75 questions that people can answer to get a score that will help them see how ready they actually are.”

Of all these issues, Halper thinks the most important are data readiness and skills readiness. “The scores for organizational readiness were pretty high [meaning organizations had addressed that issue]. Executives are on board; they’re excited about AI. They want to push a culture of innovation. What executives don’t necessarily understand is that you need a data foundation and the skills, though that simply may not be what executives necessarily think about.”

Data Quality Management

Strong data quality is also necessary, but it’s hard to achieve, and now they’re dealing with new kinds of data such as vectorized images. Organizations have made the move to cloud platforms, which are useful for data readiness, and often enterprises have multiplatform environments, but they haven't necessarily achieved the maturity to be completely AI ready, in part because of the data quality issue. That’s no surprise. In other surveys, TDWI has found that organizations aren't very satisfied with their data quality.

“When I was at Bell Labs a long time ago, I was using machine learning algorithms to analyze data. I wanted to see if I could use those algorithms to find out if there were problems with the data, and I was able to do that, much to the surprise of the people who were running the data center at the time. Now, so many years later, vendors are infusing their tools with AI and with machine learning algorithms and other algorithms to automatically detect if there are problems with data -- more so with structured data than unstructured data at this point. That's actually the biggest area we see organizations using automated tools for -- data quality management, just because it's such a big issue, and the ‘garbage in, garbage out’ adage still rings true.”

Halper explained that the assessment, which asks questions across five categories, has been available for over a month, and some of the results have already surprised her. For example, the overall median score for all enterprises was 62 out of 100. Although 62 may not sound like a high score, Halper noted, “We do many assessments, and the scores are always about 53 out of 100. Now, the score is 62 out of 100, which is higher than I thought it would be.” TDWI divided the readiness scale into five stages, and 62 is within the “Standardizing” stage, which is when enterprises are putting their strategy in place. They may have some preliminary use cases; leadership is on board and understands the impact AI can have. Enterprises typically have a data platform, such as a data warehouse or data lake, and many of these are in the cloud. Enterprises are starting to get a handle on managing their data, and they may be starting to build a predictive model.

“These results actually jibe with what we've seen in our research where around half of organizations are actually on the cusp of predictive analytics and starting to make use of these technologies. We routinely see in our surveys that getting data ready for advanced analytics is a top priority for organizations -- it's really what they're trying to do for data and analytics. The biggest area of investment for data management this year is cloud data warehouses. That could explain why enterprise scores in the organizational and data readiness components of the assessment have been high -- and lower in skills readiness. They have their data platforms in place and the data professionals they need, but those professionals still need additional skills.”

Finding workers with the skills enterprises need can be difficult. “Some places will attract data scientists and data engineers more easily than others; in some parts of the country, it's easier than in other parts. There's a move to upskill data analysts to do more, which certainly could work. They understand the business and the data. Is this upskilling working? Organizations are saying it’s too soon to tell. Of course, it's not just data scientists -- data engineers and Ops people are needed, and those skills are expensive.”

Enterprises can take the assessment at no cost at You’ll find other assessments at

TDWI Membership

Accelerate Your Projects,
and Your Career

TDWI Members have access to exclusive research reports, publications, communities and training.

Individual, Student, and Team memberships available.