(who-is-Danny/Shutterstock)
Chief knowledge officers have lots of challenges on their plates nowadays: knowledge integration, safety, privateness, compliance, cloud migrations, and IT employees and sources, to call a couple of. However a whopping 68% of CDOs in a latest report recognized knowledge high quality as their primary downside. The driving drive behind the surge within the consciousness of knowledge high quality falls squarely on the emergence of generative AI, says Mike McKee, the CEO of Ataccama.
“AI has been that catalyst to look extra intently at knowledge high quality,” McKee mentioned. “The nearer that folks have seemed, the extra involved they get, and consequently, it’s gone means up the precedence listing for folks to handle.”
Ataccama commissioned Hannover Analysis to survey about 300 senior knowledge professionals within the US, UK, and Canada in late 2024. The outcomes of that survey, which Ataccama printed in its newly launched Information Belief Report 2025, present the numerous enterprises are dealing with challenges of their plans to maneuver ahead with GenAI immediately due to points with knowledge high quality.
“AI fashions are solely as efficient as the information they depend on,” Ataccama says in its report. “And when that knowledge is unhealthy, the implications are far-reaching.” Impacts of unhealthy knowledge, the corporate says, embrace: inaccurate insights, operational slow-downs, waste of sources, jeopardized compliance initiatives, and lowered ROI.
The report additionally discovered that legacy methods share lots of the blame. They’re ill-quipped to deal with growing knowledge volumes, and lots of have been designed to supply interval knowledge updates, not the continual, real-time streams demanded by AI, Ataccama says. Sustaining knowledge high quality throughout the group is a problem for 41% of respondents, the corporate says. These are all the explanation why simply 33% of organizations report significant progress in AI adoption, the corporate says within the report.
Information development exacerbates knowledge high quality points (Yurchanka Siarhei/Shutterstock)
Information high quality has been an issue because the first byte was written. There are untold ways in which knowledge can go unhealthy (with human error main the best way), and knowledge professionals have spent numerous hours making an attempt to handle them.
As an illustration, throughout the knowledge warehousing period, firms embarked upon heavy-handed, top-down initiatives in an try and dictate knowledge high quality requirements for the group. When knowledge volumes have been decrease, firms might get away with brute-force approaches, corresponding to grasp knowledge administration (MDM) initiatives that outlined a centralized “golden report” that could possibly be relied upon for decision-making.
However these boil-the-ocean approaches aren’t efficient in immediately’s massive knowledge setting. For a wide range of causes–together with the proliferation of knowledge silos (each on-prem and within the cloud), the fast growth of use instances, and the emergence of unstructured knowledge as a valued useful resource–knowledge high quality has gotten worse as volumes go up.
The very best that firms can do is attempt to apply accessible sources to essentially the most urgent knowledge high quality problem at hand, McKee says.
“Making an attempt to grasp all these totally different knowledge sources can rapidly develop into a idiot’s errand,” he says. “Surely, the piece of recommendation is begin with the enterprise initiative. Don’t begin with a theoretical knowledge challenge. [Ask yourself] what space of the enterprise are you making an attempt to enhance? What space of the of the enterprise has the least belief within the knowledge and wishes to handle knowledge high quality points first?
Mike McKee joined Ataccama as CEO in August, 2023
“To run this advertising and marketing marketing campaign, I can simply use these 5 knowledge sources,” he continues. “I do know there’s 25 knowledge sources, but when I truly take info from these 5 knowledge sources, match and merge, convey that collectively, know what the right info is or the grasp info from a subset of the sources, then I’m driving that enterprise initiative higher.”
Whereas we’ll by no means have good knowledge, there’s huge room for enchancment on what we’ve immediately. Many firms battle to enact significant analytics due to misspelled names, faulty addresses, clean fields, and deliberately incorrect knowledge entered into types. Each firm struggles to purge these errors from their databases and file methods, however the problem has develop into much more urgent since we’ve tried to make use of this knowledge for GenAI.
“I believe an enormous turning level was ChatGPT a few years in the past,” McKee tells BigDATAwire in an interview. “Unexpectedly, persons are speaking about AI. Unexpectedly, the boards and the enterprise leaders are like, hey, are you able to begin utilizing AI? And swiftly, the CIOs of the world who’ve been engaged on these knowledge initiatives for thus lengthy, they’re like, ‘Hey, we’ve been working these knowledge initiatives. You didn’t care about them then. Now you care a lot.’”
The persevering with explosion of knowledge is placing the onus on knowledge stewards and knowledge engineers to trace down and repair knowledge high quality issues, McKee says.
“You’ve gotten a hard and fast quantity of knowledge professionals making an attempt to deal with an exploding quantity of knowledge, which, as soon as once more, goes to have a detrimental influence on knowledge high quality and the necessity to have automated knowledge high quality instruments to handle that concern,” he says.
Ataccama is searching for to handle the information high quality downside by way of automation. It employs machine studying and AI to assist with the matching and merging capabilities in its knowledge high quality product, and in addition to automate a lot of the rule creation and rule documentation work, McKee says. It additionally makes use of GenAI methods to assist enhance knowledge high quality to bolster different downstream GenAI initiatives, a terrific instance of the virtuous cycle of knowledge and AI.
However higher knowledge high quality instruments can solely get you thus far. Of their 2025 AI & Information Management Government Benchmark Survey, Randy Bean and Tom Davenport discovered that 92% of respondents “consider that the first barrier to establishing data- and AI-driven cultures is folks and group change-based, and solely 8% thought expertise was the perpetrator.”
Information investments are growing (Supply: “2025 AI & Information Management Government Benchmark Survey” by Randy Bean and Tom Davenport)
With regards to the significance of knowledge high quality, nevertheless, Davenport and Bean, who’s an advisor to Ataccama, are in full settlement with Ataccama and McKee: GenAI is exposing a knowledge high quality as the huge downside that it’s.
“…[A]s a consequence of the fast enhance in curiosity and dedication to AI funding, a rising share of organizations are actually specializing in their knowledge initiatives as effectively,” Bean and Davenport write. “It’s more and more understood that the standard of AI is basically dependent upon the standard of the information that’s accessible.”
The excellent news is that the popularity of the issue of knowledge high quality is resulting in extra sources being dedicated to it, each when it comes to increasing the human and organizational heft wanted to assault it, in addition to shopping for higher instruments. Of their 2025 AI & Information Management Government Benchmark Survey, Bean and Davenport word that investments in knowledge and GenAI are growing. So is the p.c of decision-makers who say that knowledge and AI are a prime precedence.
Ataccama can also be seeing this pattern influence its income. Whereas the Boston-based firm doesn’t focus solely on knowledge high quality, it’s the heritage of the corporate, which has its roots within the Czech Republic. In line with McKee, bookings elevated 100% within the final yr, whereas top-line income jumped 30%. That signifies firms are recognizing and responding to the issue, he says.
“I’d say that extra organizations are prioritizing it…and I believe we’ll see main enhancements within the subsequent two to a few years,” McKee predicts. “I believe we’re type of within the ‘admit you’ve an issue’ stage. When you admit you’ve an issue, search for options. After which as you search for options, then you definitely’ll begin to see an enchancment general.”
Associated Objects:
Ataccama Introduces AI Agent For Enhanced Information Administration
Overcoming the Monetary Implications of Poor Information High quality
Information High quality Getting Worse, Report Says