"We were wasting too much time debating the accuracy of numbers as opposed to using that time to make decisions."
That's how Satty Saha, Group CEO of CreditInfo-a credit bureau operating across 30+ countries-described the moment he realized his organization had a data problem. Not the kind of data problem you'd expect from a company whose business is data and analytics. An internal data problem.
CreditInfo provides credit data and risk insights to lenders worldwide, helping them make better decisions about who to lend to and at what terms. But when Saha joined as CEO, he discovered that despite selling data solutions to customers, the company's own financial data was "all over the place."
Different markets using different accounting systems. Local teams reporting according to local standards while headquarters needed global GAAP. Month-end close meetings consumed by debates about whether the numbers were even correct before anyone could discuss what the numbers meant.
Sound familiar?
"It was really old school," Saha recalls. "People were collating information on Excel sheets, sending it to someone at the center who would manually reconcile everything. Lots of back-and-forth correcting information. Prone to human errors, things getting missed. Real loss of productivity."
The cost wasn't just inefficiency-it was strategic paralysis. When leadership can't trust the numbers, they can't make confident decisions. When finance teams spend 60-70% of their time gathering and reconciling data, they're not analyzing trends, modeling scenarios, or advising on strategy.
Any CFO or COO who's tried to implement enterprise-wide data transformation knows the hardest part isn't the technology-it's getting 30+ local finance teams across different countries to change how they work.
"People don't like change, even though the change can make their work life easier," Saha notes. "We needed to win hearts and minds. Making connections for people-how this change improves their work life-is really important."
The breakthrough came when local CFOs realized: once we do this integration work, headquarters stops asking us for data manually every month.
No more urgent Excel requests. No more building custom dashboards for every leadership presentation. No more explaining variances in hastily assembled PowerPoint decks. The data flows automatically; reports generate on demand.
"That was one of the big wins," Saha explains. "The time they would spend answering queries from head office-they could now reclaim. People started seeing the benefit."
Jiri Manas, COO of Keboola (and former CTO/CIO across banking, retail, and real estate), emphasizes this human dimension of data transformation: "People need to understand why we do things. Then they need to understand: what's in it for me? Is the impact positive or negative? People have fears of the unknown."
His approach: find the champions. "Some people jump into change right away. Some are laggards. The rest are in between. You find those naturally curious people who want to be part of the initiative and make them champions in their regions."
When you're staring at a global data mess, where do you start?
For CreditInfo, the answer was clear: establish a single source of truth for financial data first.
"We looked at the portfolio of countries we had," Saha explains. "Many companies were using different accounting systems, recording information according to local standards, while at group level we needed global accounting standards. There was change required at both group and local level to harmonize data reporting standards."
The implementation sequence:
"Once we had the finance data in and data was coming in automated fashion, we moved to the next metric I check every morning-pipeline and sales information from CRM systems," Saha says. "Then HR systems, production systems. Over time, you have everything available at a single global data warehouse."
Manas calls this approach "proof of value" rather than big-bang transformation: "Companies have decades of technology vendors promising 'this one will finally fix all your problems.' We don't do massive projects that deliver late and over budget. We start by demonstrating the technology can deliver value very quickly, then continuously deliver small parts of value to the business every week."
Time Spent on Data Collection vs. Analysis
Monthly Close Reporting Cycle Time
Data Accuracy Dispute Rate
Report Generation Time
Track finance team time allocation through activity logs or time studies. Measure close cycle from data request date to final report distribution. Monitor meeting effectiveness by tracking time spent on data validation versus decision discussion. Calculate report generation time from request to delivery for ad-hoc analyses.
When does a skeptical organization start believing in data transformation?
"When we started generating reports automatically," Saha recalls. "That's when people saw the benefit-this has taken work I'd otherwise spend answering questions from head office, and now they can self-serve."
The second inflection point: "When we integrated the first market, got the data in, matched it, ironed out differences between historical reporting and what the data was showing, understood where differences came from and reconciled them-that smoothed the journey. The learning from the first pilot was really important. Then connecting the rest of the markets became a lot smoother."
Manas adds a technical dimension: "With the power of technology, you can surface problems that were always there but hidden. We've had customers discover they're missing significant square meters that aren't being rented-someone put in a typo in the system. It's not a technology problem, it's a data entry problem. But technology surfaces it so you can clean it and trust the numbers to make decisions."
Here's where it gets interesting.
"Ten or fifteen years ago, decision-making was far more subjective," Saha reflects. "Now it's based on data. And the way we're generating insights from data-that pace has accelerated a whole lot with automation and AI."
CreditInfo is now experimenting with natural language querying of their unified financial database. "We're testing being able to query our database using natural language. If we do that, we democratize analysis. People not confident with Excel or BI skills can query the data and get meaningful insights."
The implications for decision velocity are profound: "Generation of insights and summarizing those insights and acting on them-that cycle time which earlier took days after we collected month-end data will now become instant. You get data really quickly, draw insights really quickly, make decisions much faster."
Manas sees the trend line clearly: "It's the year 2050 and we naturally speak to computers. Underlying technology gives us answers, similar to working with assistants, but vastly connected to the right information in your organization. You're literally having a discussion about your business with an AI assistant. Now the question is: is it 2050 or is it 2028? This is where we see it coming."
When asked to complete the sentence "AI is useful when...", their answers reveal the pragmatist's perspective:
Saha: "When it can become a force multiplier and enhance productivity for us."
Manas: "When it gives you value."
Not hype. Not revolution for revolution's sake. Value. Productivity. Force multiplication.
But here's the critical point both leaders emphasize: AI is only useful when you have clean, unified data to feed it.
"Once you have all the data and clean data in the repository, that's when you can start querying with natural language," Saha explains. "You can't skip the foundational work."
Before we let Saha go, we asked: what's the first metric you check every morning?
"Two things," he answers immediately. "Pipeline quality-not just volume, but conversion rates, cycle times, things that help us predict how the quarter will be, better than just optimism from salespeople. And cash. For me, that really tells you the truth about the health of the business. If cash is off, you know something else is off-whether it's sales, delivery, or collections."
Both metrics require clean, unified data from multiple systems to calculate reliably. Both drive strategic decisions that can't wait for month-end close.
This is why finance leaders are investing in data infrastructure now. Not because it's trendy. Because the alternative-making decisions on stale, disputed, manually-compiled numbers-is no longer tenable when markets move this fast.
"We are in for a change in the way we operate and work," Saha reflects. "It's going to make us really, really more efficient as managers."
But efficiency isn't the end goal-it's the means. The real goal is reclaiming finance teams' time and cognitive capacity from manual data drudgery so they can do what they were hired to do: analyze, advise, and drive business value.
Your finance team didn't join the profession to reconcile Excel files at 9 PM. They joined to be strategic partners to the business.
The technology exists to make that happen. The question is: how many more month-end closes will they suffer through before you build the foundation?
About This Conversation
This article is based on a Keboola Talk with Satty Saha, Group CEO of CreditInfo (credit bureau operating in 30+ countries), and Jiri Manas, COO of Keboola (former CTO/CIO across banking, retail, and real estate). The conversation explored practical approaches to financial data transformation, change management, and AI readiness.
.png)
.png)
.png)