Three Mindset Shifts to Achieve Autonomous Finance

By Dennis Gannon
October 1, 2022

CFOs need to adjust their thinking in three key ways before they’ll be ready to fully embrace the set of technologies needed for autonomous finance.


Can you picture an autonomous finance function in which the majority of operation—and even a significant amount of management—is conducted by machines with no need for human intervention? This is a concept that goes far beyond current conceptions of finance automation, which often are really “interrupted automations” that are still heavily supervised and reliant upon human judgment to run and maintain the process.


Nearly two-thirds of the CFOs Gartner surveyed on this topic can indeed picture autonomous finance and believe it will be a reality within the next six years. This is partly driven by their own experiences as they and their teams get increasingly comfortable with some of the underlying technologies that support it, particularly those commonly referred to as AI.


Gartner already sees a sufficient number of organizations willing to experiment and solve business problems with AI that we predict half of all large finance departments will be using AI by next year to create their short-term financial forecasts. The technologies themselves are also maturing at an astonishing rate.


Despite CFOs’ vision of a fully autonomous finance department within this decade, few are embracing the full set of technologies that support an autonomous finance function that will deliver the very future most are predicting (see Figure 1 for examples of use cases). For instance, only 21% are currently using machine learning, 19% prescriptive analytics, 12% natural language processing, and 8% blockchain.


Click to enlarge.


The risk is that much like in the months and years following the outbreak of the COVID-19 pandemic, we’ll witness yet another new digital divide—this one between finance organizations that are opening up entirely new pathways to revenue generation and cost savings with autonomous capabilities and those that aren’t. The very nature of these technologies, with their ability to self-learn and rapidly improve, means that the organizations that are early adopters are more likely to compound their lead on the following pack.


Why have some companies become AI leaders on the pathway to an autonomous finance function within the decade while others remain frozen in hesitancy? The answer has less to do with the complexity of technology or any external factor and more to do with CFOs’ own mindsets about technology.


When asked about what was preventing them from taking the first steps toward embracing autonomous finance, CFOs indicated that their own mindsets were among the top barriers. While at first it may seem surprising that “mindset” is ranked as a bigger barrier than obstacles like cost and expertise, this just speaks to how fundamentally different an autonomous finance function really is. These technologies are conceptually challenging in a way that traditional finance technology never was.


A CFO’s mindset—what he or she believes to be true and why—will underpin every decision about where to invest, where to pilot a new approach, and what is worth pursuing. The reason mindset is so foundational is that all the processes, approaches, and strategies that CFOs have thus far built out, that are used to manage and lead, get wired into their brains based on certain beliefs they have about how the world works. And a lot of that wiring is unhelpful as they try to move toward an autonomous function.


Because of the fundamental differences that go into creating an autonomous (as opposed to automated) finance function, the actions and strategies that feel right to CFOs end up being the wrong things to do. Automation isn’t hard conceptually; most CFOs today using robotic process automation (RPA) in some form can understand the steps that the technology will be taking to get from point A to point B. But autonomizing a function is difficult in a different way—it’s difficult to know exactly what’s going on in (for example) a private blockchain or to understand intellectually how machine learning works.


Autonomous finance isn’t just difficult to start on because it’s new. It’s difficult because these concepts are inherently difficult. As CFOs face decisions around autonomous finance, they must (1) recognize how their current mindsets hold back the levels of experimentation necessary to make needed progress and (2) embrace the mindset shifts that may feel unnatural, or even wrong, to them today.


After working with CFOs extensively on their autonomous finance strategies, Gartner has identified the three most important mindset shifts needed to embrace the future of autonomous finance (see Figure 2). The three ways in which CFOs’ brains are wired against the technologies that will enable autonomous finance:


  1. Believing finance should start small to avoid costly failures with technology investments
  2. Believing finance should use technology as a tool but rely on people to make decisions
  3. Believing their teams will embrace technology only when they see its benefits



Click to enlarge.


These mindsets served as illuminating starting points to help stuck CFOs look more deeply at his or her beliefs and create a set of contrary guiding principles that will guide meaningful action toward autonomous finance.




When it comes to making what’s perceived to be risky technology bets with the budget, CFOs are conscious that they may be held to different standards than other parts of the organization. Costly failures are considered unacceptable for a function such as finance, which acts as the gatekeeper for responsible spending. CFOs believe responsible spending means narrow experimentation with technology and waiting to scale until the application is proven.


As one CFO told us when discussing this topic: “Finance is not given the same permission to fail. We have to be purer. We’re giving everyone else a hard time about the numbers.” Unfortunately, this conservative mindset runs up against the reality of maximizing the benefits of autonomous technologies, which rewards greater experimentation and the creation of more learning opportunities.



Gartner analyzed survey data on finance functions’ use of AI and found that organizations that pilot AI broadly in the first 12 months identify twice as many applications of AI over the next several years than their peers (see Figure 3). And this is despite no meaningful difference in spending compared to organizations that pilot AI in a more limited fashion. There’s every reason to think that those organizations that open up such a lead in the number of AI pilot projects will continue to put space between themselves and competitors as the opportunities and learnings compound with the maturity of their autonomous finance systems.


Rather than starting with a small number of pilots, CFOs must encourage running several pilots simultaneously across finance subfunctions. This helps to reinforce a mentality across the team that this sort of experimentation with new technology isn’t confined to one part of finance, nor is it optional. Instead, it surrounds every finance employee and encourages further learning and adoption.


The fear of failure and a potential for wasting resources often hold finance functions back from broad experimentation in the early stages. But demystifying failure reduces this failure phobia. To demystify failure, CFOs can borrow an approach from author Tim Ferriss and complete a fear-setting exercise. Ferriss’s approach to overcoming the fears that hold back personal growth also maps well to how a team or individual can identify the obstacles that prevent them from pursuing a bold innovation (and the risks of inaction). Our slightly modified approach is:


  1. Start by defining the worst things that could happen if you were to pursue the innovation. Try to get beyond an intellectual analysis by visualizing what would happen if the worst occurred.
  2. Determine how to prevent the worst from happening.
  3. Identify how to lessen the consequences if the worst really did occur.
  4. Write down all the benefits of pursuing the innovation. (Presumably, the innovation is worth pursuing because of the possible future benefit.) Provide as much detail as you can about the benefits.
  5. Most importantly, write down the cost of inaction—the cost of not pursuing the innovation in the short term (such as after six months).
  6. Finally, expand on step 5 by writing the cost of inaction in the longer term (such as after 18 months).


Taken together, these steps will pave the way for the necessary mindset shift by helping you enable more experimentation with less worry about negative outcomes.




CFOs believe finance should use technology as a tool but rely on people to make decisions. Even when there’s evidence that technology is better or more accurate than humans, people are reluctant to trust it. This is a phenomenon referred to as “algorithm aversion.” When given a choice between human judgment and algorithms, humans prefer what they’re familiar with: other humans.


Algorithm aversion often manifests as holding technology to a higher standard than human counterparts. For example, in our survey, CFOs report the maximum acceptable variance for a traditional financial statement forecast generated by humans as 10%. In comparison, the maximum acceptable variance for a technology-generated forecast is 5%.


This discrepancy likely results from finance comparing algorithm performance to a specific target, often an overly lofty goal—or even perfection. Instead, finance should compare an algorithm’s performance to human judgment (see Figure 4).


Click to enlarge.


Going forward, CFOs and their teams need to give technologies for autonomous finance as much credit as they give people. In order to get to that point, however, leaders and their teams will need to build trust in AI and other technologies through autonomous decision-making guardrails. These guardrails include:


  • Fairness: The decisions must be made free of bias
    and discrimination and not manipulated to benefit insiders.
  • Clarity and transparency: Decisions made must be understood by humans and well-documented—no black boxes and no “just because.”
  • Accountability: Who is responsible if an autonomous decision results in harm?
  • Security and safety: Any autonomous process must respect data integrity and privacy.
  • Human-centricity: The more important a decision, the more it needs to be made in the name of benefiting humans.


It’s clear that getting comfortable with how autonomous technologies make decisions and developing appropriate guardrails are significant undertakings involving AI ethics; appointing and managing the right AI talent; and developing in-depth documentation, safety, and security practices.


CFOs can start now by taking some basic steps and then building more complex processes as needed, including:


  1. Understanding and addressing common ingrained perceptions about technologies for autonomous finance: Ask stakeholders what they’ve heard to address their
    specific concerns.
  2. Showing a comparison of human and technology error rates in forecasting: Conduct an exercise in which the team compares its own forecast predictions to that of an algorithm and discusses the results.
  3. Diving deep into how the tool works: After a new tool is introduced, explain to the larger team the model’s logic and how it was built, and allow employees to pressure-test it before launch.
  4. Shrinking the risk space: Match the technology’s complexity to the operations it needs to perform. For example, choose the least sophisticated AI technology capable of achieving the objective.
  5. Assigning “grandmaster” designers who monitor technologies’ performance and take responsibility for it: Designers should anticipate what behaviors the technologies could develop after their initial launch, seek out unanticipated consequences, and address them.
  6. Investigating the feasibility of building a “conscience” for technology: Devote the budget to building a parallel system in sophisticated technologies that checks the behavior of the primary algorithms.




Most CFOs believe their teams will embrace technologies when and as they see the benefits. But while CFOs are usually excited to sponsor exciting new technology investments and tout their transformational potential, finance employees see those same leaders displaying little practical knowledge about the applications of these technologies and keeping an arm’s length from their use. Only 29% of CFOs say they invest significant personal time in learning about technologies for autonomous finance and their applications.


Changing their personal behavior toward technologies for autonomous finance is a superior way for CFOs to achieve employee behavior change. Employees don’t change because leaders ask them to; they change because of how leaders behave. It isn’t enough for leaders to talk up the potential of autonomous finance. They must also create an environment that enables everyone to live the culture of change.


This means CFOs must be strategic about role-modeling desired behaviors, rather than waiting for their teams’ behavior toward technologies for autonomous finance to change through a bottom-up approach.


CFOs should demonstrate new behaviors with “culture hacks.” A culture hack is a small change that exploits a single area where your culture is vulnerable to change. In this context, hacks are small, emotional, immediate changes that have big impacts.


Examples of culture hacks include:


  • Setting a minimum failure rate for innovators, signaling that if the team isn’t failing a minimum amount, it isn’t innovating properly.
  • Asking people to draw a concept (for example, “transformation”) in under one minute using no words and then having everyone describe what they drew.
  • Including an examination of what the team should
    stop doing (including behaviors, projects, assumptions, and habits) in every conversation about embarking on something new.
  • Asking the team, “What would this look like if it were easy?” when presented with a new initiative that appears time-consuming, complicated, or difficult.




Here is how one client company, a large multinational consumer company, approached “leading from the front” in technology.


CFOs know how a lack of digital skills throughout the finance team hinders the return on digital finance investments. To tackle this problem, the company set out to bring up the entire finance team—from junior analyst to the CFO—to a foundational level of digital skills in less than a year. It accomplishes its universal digital upskilling by mandating digital foundations training for the company’s finance leaders. This aligns the digital skills training with its finance competency model, allowing freedom in how staff members complete the training and fostering a digital community with new social media communication channels, digital learning champions, and progress monitoring.


The benefits of this approach include:


  • Executive leadership signals the criticality of digitally upskilling the finance team by mandating finance leaders lead by example in completing the digital foundations training program—even the CFO.
  • Finance collaborates with HR to curate a digital skills training curriculum explicitly targeted to specific skills in the digital finance competency model.
  • Finance leaders promote “freedom within a framework,” allowing staff to achieve digital foundations certification through multiple learning formats (such as individual or group learning, structured vs. unstructured curriculum), subject to the achievement of passing scores within a specified time frame.
  • Finance creates a learning network with dedicated learning leaders and social media tools to foster group learning and engagement.


As its digital foundations training nears 100% completion, the company’s finance team is already realizing higher returns on its digital investments. The team has seen a 50-fold increase in the number of active Microsoft Power BI users. And several users have moved on to obtain “superuser” status. Agile is now not only commonly used but commonly understood throughout the finance function, with its applications being considered in many digital projects relating to finance process improvements in ways big and small.




Confronting the way CFOs’ mindsets are wired against technology is the hardest barrier to overcome in the journey toward autonomous finance. Successful CFOs will embrace broad experimentation with technologies from the outset, build trust in working alongside new technologies, and recognize that personal behavior change is the best way to encourage employee behavior change toward technologies for autonomous finance.


Dennis Gannon is vice president, Research, at Gartner. He can be contacted at dennis.gannon@gartner.com.
0 No Comments
You may also like