“Make AI skills a job redesign program, not an additional training program”

Richard Paik, CFO at GitHub, believes finance leaders are asking the wrong question about AI skills. Knowing how to use AI tools is a good start, but it won't get you very far if the way your team works stays the same.

Rillion’s 2026 AI in Finance Report shows that while 60% of US finance leaders see AI skills becoming increasingly important, relatively few consider a lack of skills a major barrier today. Richard sees a risk in that disconnect. Early wins with AI can create a false sense of progress, while bigger gaps in processes, controls and ways of working go unnoticed.

So what does it take to build real AI competence? Richard shares lessons from GitHub's approach to AI adoption, why security and experimentation need to go hand in hand, and why CFOs should think about redesigning jobs rather than simply adding another training program.

You have a background spanning AI adoption, Digital Transformation and Cyber Security. When you look at how finance leaders think about AI competence, what do most of them get wrong?

Of course, this depends on where a company is in their AI adoption journey — as most of us are still driving early adoption today — but I see many finance leaders over-index on tool familiarity and under-index on operating model change. AI competence is not “who used a chatbot,” it is whether finance can redesign decisions, controls, workflows, and accountability around AI-enabled work.

Our data shows that security and data quality top the barrier list for US finance leaders, while lack of internal AI skills ranks well below security, data quality, and integration. As someone with deep experience in both AI adoption and Cyber Security, does that order surprise you, or does it make perfect sense?

That order makes sense. In finance, trust is a prerequisite: if security, privacy, and data quality are weak, skills investment won’t scale. So, teams often delay skills urgency because risk and governance gaps block adoption first.

That said, it’s unrealistic to expect to completely resolve data estate concerns before deploying AI solutions. There’s a balance to strike where customers are being smart about where and what data teams are targeting, running trials, and learning through AI adoption with data governance and controls. And of course, it’s important to buy enterprise-ready tools from trustworthy vendors. If a vendor doesn’t have clear documentation around data controls and can’t answer questions regarding the data they collect and process, that’s a red flag. 

Most finance leaders do not think they have a skills problem today. But 60% say understanding AI tools will become one of the most important skills in finance. In your experience, why is it so hard for organizations to see a competence gap before it becomes a crisis?

Competence gaps are hard to recognize because early AI use can look productive while masking fragility. Organizations mistake isolated wins for capability, until scale exposes missing foundations: data discipline, model governance, and process redesigns.

There are two skills that finance leaders need to invest in, beyond AI adoption. First, understanding how AI works, with a focus on what it does well today and what problems it can solve more efficiently. Second, system thinking to move beyond the initial prompt and output to consider how to effectively use AI systems from end to end, including which controls and processes should be standardized to drive long-term, sustainable improvements. 

You have worked closely with customers and partners on AI adoption across global organizations. What separates the finance functions that successfully build AI competence from the ones that get stuck? 

What separates leaders is execution discipline. They pick a few high-value finance use cases tied to measurable outcomes, they pair domain experts with AI practitioners from day one, and they institutionalize learning (playbooks, controls, reusable patterns), not one-off pilots.

Security concerns rank as the number one barrier to broader AI adoption in US finance. How do you build the kind of trust in AI systems that allows finance leaders to move past that barrier without compromising control?

You build trust through controlled enablement, not blanket restriction: strong data classification, guardrails, least-privilege access, model risk controls, human-in-the-loop approvals for material decisions, auditability, continuous monitoring, and feedback loops. Trust grows when leaders can see how decisions were produced and governed.

As CFO of GitHub, you sit inside one of the world’s most developer-driven organizations. How does the approach to AI competence look different in a tech-native company compared to more traditional finance functions?

In a tech-native company, AI competence is built as a system capability: shared platforms, rapid experimentation, telemetry, and cross-functional ownership. Traditional finance teams often treat AI as a project; tech-native companies treat it as an evolving core competency embedded in daily work.

At GitHub, we’re very intentional about helping teams across the entire organization become AI first. One of the most important things is creating room for experimentation and helping nontechnical teams understand where AI can really make a difference in their day-to-day. As a result, we now see teams from legal, marketing, social impact and more regularly using AI to accelerate workflows, reduce toil, and free up time for more creative and fulfilling work.

If you could give one piece of advice to a CFO who wants to build genuine AI competence in their finance team rather than just AI awareness, what would it be? 

Make AI competence a job redesign program, not an additional training program. Redefine finance roles, decision rights, controls, and KPIs around AIenabled workflows, then train to that new reality. That is how awareness becomes durable capability.

The interview was first published in Rillion's AI in Finance Report.

About Richard Paik

Richard Paik is CFO at GitHub, overseeing the company’s global financial strategy, planning, and operations. He previously spent more than 20 years at Microsoft in senior finance roles across several international markets.