Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

How to deploy AI without breaking trust

The challenges facing finance today can offer a roadmap for how organizations can responsibly scale AI.

[Photo: Getty Images]

Profile photo of Rob Israch

President · Tipalti

Record-setting investments and trillions in projected economic impact have placed AI on virtually every company’s agenda. Firms need speed for competitive advantage, but their futures are also determined by how confidently they can foster the adoption, experimentation, and scaling of AI.

Industry forecasts suggest most companies will be AI-driven either directly or through their customers and partners in 2026.

AI’s transformational impact will determine the future of many organizations. But success in this transformation depends on a critical factor: organizational and customer trust. As organizations’ confidence rises with the practical use of AI, the trust gap widens with scale.

Tipalti’s latest research of finance professionals proves this point. Our survey found 60% of those currently using AI were “extremely optimistic” versus 37% of those merely experimenting with AI. Finance leaders tell us trust is an important element in AI adoption: If you can’t trace how an answer was made, you can’t trust it.

Guardrails facilitate AI experimentation

AI’s breathtaking pace may feel familiar because we’ve been here before. The cloud computing boom of the 2000s and 2010s promised speed, scale, and reinvention, but delivered unevenly. As firms rushed to migrate applications and data, many prioritized speed over discipline. The result was a wave of trust-breaking experiences, from security missteps to customer alienation. The most successful companies maintained trust through discipline and execution. When customers don’t understand how technology drives outcomes, or if they cannot audit those outcomes, adoption fails.

We find a dual-sided risk in today’s AI approaches. Unbridled experimentation can expose companies to unacceptable operational, security, and reputational risk but, overly rigid controls prevent AI initiatives from succeeding.

Our research suggests the strongest opportunities for firms to build trust in AI involve governance, training, transparency, and better data quality. While trust does not demand perfection, it requires a system with foundational priorities of governance, auditability, and implementation readiness.

Audit priorities should focus on three areas: explainable models, high-quality inputs, and clear accountability for AI-driven decisions. Your teams should document the data and product trails, including prompts, data sourcing validations, and approvals across the product lifecycle.

Responsible AI requires compliance, risk management, and incident disclosure. According to Deloitte’s 2024 State of Generative AI in Enterprise survey, organizations that take trust-building actions when implementing gen AI processes “were more likely to report that nearly two-thirds (66%) or more of their organization’s expected gen AI benefits had been achieved.”

Organizations also need implementation readiness. AI should be integrated into existing systems with human oversight built in from the start. The focus must be on data accuracy, not just data volume.

Finance is a test case for AI adoption

We find finance is a proverbial canary in the coal mine for AI adoption because many organizations look to the finance function as a proving ground. This is due to the data intensity, repeatability, and strict audit standards of key finance workflows, which are ripe for AI implementation and measurable wins. Our research shows organizations are using AI in their finance departments for analysis (63%), reporting (62%), forecasting (58%), and fraud detection (57%). As a result, regular users of AI report tangible improvements in the quality of work (98%) and decision-making (97%).

Finance also provides a stress test for AI. The use of AI in finance shows experimentation is possible in highly regulated environments, even as institutional trust remains a primary limiting factor for scaling AI. The National Institute for Standards and Technology frames AI adoption as inseparable from risk management in its December 2025 draft Cyber AI Profile. Scale falters when processes aren’t accountable. The challenges facing finance today can offer a roadmap for how organizations can responsibly scale AI.

Customer trust is essential

Businesses cannot afford to be inward looking. Security and compliance set standards but customers define acceptable AI uses.

AI trust benchmarks are often siloed within businesses. With security, legal, and compliance teams defining acceptability, trust historically involved a checking-the-box exercise. The better alternative involves high security and compliance standards, followed by a customer-validated, market reality check.

While security and compliance set the foundation, customer expectations, usage, and outcomes are signals for whether AI is perceived as harmful or helpful, transparent or opaque. Experimentation with AI should factor in customer impact, with compliance guardrails, accountability processes, and governance standards articulated externally. Successful companies will earn customer confidence and treat trust as a shared, non-negotiable priority.

Trust is an AI accelerant

Businesses often struggle to move AI from the pilot phase to delivering measurable business value. This transition requires space to learn, test, and iterate, without the anxiety of requiring production-grade status for AI models on day one.

Few organizations move from AI experimentation to scale, according to 2025 research from McKinsey & Company. Almost two-thirds say they haven’t begun scaling across the organization.

To get to the scaling phase, companies must prioritize trust. Consistent with our research, Gartner emphasizes governance, policy enforcement, and monitoring as requirements for AI deployments. This demands balance. The question is not how do we rigidly control AI? Instead, companies should ask: What standards do we need for our teams to safely experiment?

Final thoughts

Finance operation teams are proving to be leaders in the transition from experimenting with AI to deploying AI. Clear standards for control, auditability, and explainability provide a roadmap for responsible AI. In a marketplace defined by speed, organizations that build trust are poised to lead the AI transformation.

Market forces reward speed and experimentation. To accomplish this, companies must build trust in their AI use from the beginning. Effective innovation means using trust as the factor to empower and facilitate AI experimentation. 

Rob Israch is president of Tipalti.