CFOs Get Schooled on Taking Enterprise AI From Curiosity to Business Impact
Artificial intelligence (AI) is hurtling forward at an unprecedented pace, changing everything along its way.
For chief financial officers (CFOs) and other business leaders, this advance is met with hesitation. The reason? A lack of clear standards, concerns about data privacy and the risks of unreliable AI outputs.
“Lack of operating procedures when it comes to AI, particularly GenAI, is like putting a speed bump on a highway,” Jeff Stangle, director of product (AI and platform) at Coupa, told PYMNTS. “Now, the size and shape of that speed bump is really comparable to what your practical frameworks look like.”
Without governance, these speed bumps could slow AI adoption in procure-to-pay operations — unless businesses take the right approach.
CFOs are navigating a delicate balance. They’re accountable for financial outcomes while also ensuring AI-driven decisions are unbiased, secure and reliable. Without clear governance, hesitation sets in. And while some companies push AI adoption forward at breakneck speed, others are stuck waiting for clearer guidelines.
So, what’s the solution? Stangle offered a three-pronged approach: practical, transparent and purpose-built frameworks.
Overcoming Lack of Standards
Without a holistic perspective, companies risk deploying AI solutions that create inefficiencies, fail to align with business objectives, or introduce unforeseen vulnerabilities. Above all, the holistic perspective needs to be practical.
“Your framework needs to be practical,” Stangle said. “You don’t want to miss things. You don’t want things to fall through the cracks.”
By establishing a structured and practical governance model, businesses can ensure AI functions as a reliable and value-driven asset rather than a source of operational ambiguity.
At the same time, without transparency, businesses risk relying on “black box” AI systems — models that produce outputs without clear visibility into their inner workings. This lack of insight can lead to unintended biases, regulatory challenges and difficulty troubleshooting errors.
“All AI is built on algorithms, but how many degrees of separation do you have from the truth?” Stangle said, adding that transparency means ensuring AI models remain interpretable and explainable, so businesses can understand how decisions are made and adjust accordingly.
And ultimately, AI should never be implemented for the sake of novelty — organizations must first identify the specific business challenges they aim to address before developing or deploying AI solutions.
“AI is a tool. GenAI [generative AI] is a tool to solve problems. We don’t build code that doesn’t have a defined problem,” Stangle said.
From Skepticism to Adoption
Stangle recommends a crawl-walk-run strategy to AI adoption. Start with automation, he advises, noting that many enterprises are already using AI-driven automation in basic finance and procurement functions. This is where businesses see the biggest cost savings, he said.
That’s crawling. Walking is experimenting with cross-functional AI applications and beginning to pilot GenAI solutions with select teams.
Time to run? That would entail rolling out GenAI across the organization once it has been tested, optimized and integrated into a governance framework.
Still, two of the biggest AI concerns raised by CFOs in the most recent CAIO report by PYMNTS Intelligence and Coupa are around data privacy and misinformation.
“Data privacy isn’t just about anonymization and encryption. It’s also about aggregation. One of the key things we do at Coupa is anonymize, encrypt and aggregate data to generate insights while keeping information secure,” Stangle said.
Without strong data governance, AI models risk being trained on biased or incomplete data, leading to unreliable outputs. Misinformation is another challenge.
“If your dataset is perfect on day one, it erodes at about 2% a month. So, you need to have data quality processes in place as part of your framework — and you need to keep doing them,” Stangle cautioned. “The right time to clean your data? All the time.
“If data is not clean, then it’s not private,” Stangle emphasized, adding that should be a guiding principle for every CFO looking to make GenAI a trusted, transformative force in their organization.
One of the biggest missed opportunities in AI adoption, he said, is failing to treat vendors as strategic partners. Stangle advises CFOs to rethink the traditional request for information (RFI) process. By fostering collaboration between enterprises and AI providers, businesses can develop shared best practices and industrywide standards, ultimately accelerating adoption.
“GenAI isn’t a light switch you just flip on,” he said. “If you try to implement it all at once, you’re going to get some pretty bad outcomes. But if you build confidence step by step, AI can become a powerful tool.”
The path to confident AI requires practical governance, strong privacy measures, clean data and a thoughtful, step-by-step approach. But for those who embrace it, the rewards — efficiency, insights and competitive advantage — are well worth the effort.
The post CFOs Get Schooled on Taking Enterprise AI From Curiosity to Business Impact appeared first on PYMNTS.com.