When Bristol Myers Squibb set out to overhaul its procurement function using artificial intelligence, it made a deliberate choice: move forward without waiting for perfect data.
“We’re getting our data right as we go along,” Rhonda Griscti, executive director of digital strategy and global process lead at the pharmaceutical company, said in an interview.
The gamble has paid off, resulting in a much faster process for soliciting and evaluating bids from suppliers, among other improvements, according to Griscti.
“We’ve gone from six to nine months on average to less than 30 days for our RFP [request-for-proposal] process,” she said.
Bristol Myers has moved quickly with its AI-driven procurement transformation, even as data challenges have complicated deployments of the technology more broadly.
Just 10% of CFOs fully trust their data quality, and more than a third cite data trust as their top barrier to AI return on investment, according to a December 2025 report from global professional services firm RGP.
On the question of data readiness, Scott Rottmann, president of consulting services at RGP, takes a more cautious approach compared with Griscti.
“If the data is clean and of a quality that allows you to make good decisions, I think that's fine,” Rottmann said in an interview. “There’s an 80–20 rule here. You’re never going to be perfect at anything you do. So, you get 80% of the way there and sort of move forward.”
Data readiness debate
The differing perspectives highlight a broader debate in enterprise AI adoption: how much data preparation is enough amid pressure to deliver AI results quickly.
Skeptics of faster deployment caution that moving too quickly can introduce operational and governance risks if inaccurate, incomplete or inconsistent data sets are used to train or feed automated decision systems at scale. Proponents argue that earlier deployment can accelerate time to value, delivering measurable operational improvements sooner.
In an April Forbes article, AI strategist John Sviokla argued that organizations often over-invest in data cleaning and risk delaying value creation by prioritizing perfect data quality over extracting insights from imperfect data.
“You don’t sequence the cleaning before the processing,” wrote Sviokla, co-founder of AI advisory firm GAI Insights and an executive fellow at Harvard Business School. “Instead, you build the processing capability and let it reveal what is worth cleaning.”
Griscti said a key step is having a data lake — a centralized repository designed to store and process large volumes of raw data. “You don’t need your data to be perfect to start,” she said. “You just need to know where it is. Have it in a data lake. Be able to tap into it, and you can clean as you go.”
Griscti started at Bristol Myers Squibb in November 2021 as a senior director in R&D procurement and was promoted to her current role four years later. The procurement function was already under pressure to speed up its processes when she joined. Momentum for change accelerated after a new chief procurement officer arrived with a mandate to modernize the function.
The company launched a transformation effort in June 2024 focused on standardizing processes, improving capabilities and eliminating fragmented, email-based workflows.
Rapid AI rollout
The next phase involved AI adoption. Griscti and her team conducted a “soft launch” in November 2024 with Palo Alto-based Globality, whose platform leverages AI to automate key parts of the process for finding and selecting suppliers. The technology was later scaled organization-wide in February 2025.
As a result, sourcing volumes increased sharply, with more than $1 billion flowing through the platform in the first year — far exceeding initial projections, Griscti said. The transformation also expanded procurement activity significantly, enabling roughly 10 times more RFPs while bringing previously outsourced work back in-house with about 50% fewer resources, she said.
In-sourcing freed up cash previously spent on external providers, which Griscti redirected toward technology investments.
Griscti said her advice to other executives leading similar transformations is to avoid getting bogged down by data concerns.
“People will argue with me about this, and I’ve had this discussion at many of the conferences,” she said. “But if you wait for perfect data, you’ll never get started.”
