From Virtual to Feasible: How Chief Financial Officers Replan the Application of Artificial Intelligence
The estimated reading time for the full text is about 6 minutes.
IMD Professor José Parra Moyano points out that many AI projects fail to achieve the expected results. He suggests that Chief Financial Officers (CFOs) should refocus on three core elements: business value, the completeness of the data foundation, and employee engagement and recognition.
The Chief Financial Officer (CFO) plays a central role in an enterprise's response to the industrial transformation led by Artificial Intelligence (AI). With their overall view of the enterprise and leadership in financial strategic planning, CFOs have unique advantages in promoting the implementation of AI technology and can precisely balance innovation investment with quantifiable business results.
Generative Artificial Intelligence (Generative AI, or GenAI for short), including the upcoming Artificial Intelligence Agents (AI agents for short), can optimize operational processes, improve decision-making efficiency, and drive revenue growth. Analysts at Barclays Investment Bank predict that AI agents can independently complete approximately 7 billion tasks, helping enterprises achieve a system-level productivity leap. However, to truly unleash the potential of this technology, technical deployment alone is far from enough. CFOs need to ensure that AI projects are deeply coordinated with the enterprise's strategic goals, avoid resource dispersion and "island-style" development, and thus build sustainable competitive advantages. If financial leaders plan the implementation path of AI from a strategic perspective, this technology is expected to bring significant return on investment to the enterprise.
However, many enterprises still face challenges in transforming the potential of AI into actual business results, and a large number of AI projects fail to meet the expected goals. This phenomenon exists not only in traditional AI projects but also in GenAI projects. Even the AI agents that will be widely used this year may fall into the same dilemma.
To effectively address the complex challenges in the implementation process of AI projects, enterprises should focus on three key dimensions: business value, data foundation, and personnel collaboration. These three dimensions together form a "value - data - personnel" framework, providing decision-makers with a structured thinking path to answer the following key questions: What kind of value does the enterprise expect to achieve through AI? Do we have the key data resources to support the implementation of AI? How will employees and stakeholders understand and respond to the AI-driven transformation?
By focusing on the above core dimensions, CFOs will be able to optimize resource allocation more precisely, systematically reduce risks, and significantly improve the long-term success rate of AI projects.
I. Value Dimension: Define the Business Value Brought by AI
The first dimension of the framework requires enterprises to clearly define the value they expect to create through AI. Although this requirement seems reasonable, when asked "what specific business problems does the enterprise hope to solve through AI", many organizations often struggle to give a clear and direct answer.
Truly successful enterprises do not blindly pursue AI technology itself but focus on solving quantifiable and measurable actual business challenges - for example, optimizing sales performance through algorithms. Take a salesperson as an example. By using AI to predict high-potential customer groups and adjusting marketing strategies accordingly, the salesperson ultimately increased the annual revenue from $1 million to $1.3 million. This specific case not only demonstrates the practical application scenarios of AI but also highlights the quantifiable business value it brings to the enterprise.
Enterprises that excel in the field of AI application usually adopt a focused and practical strategy. They prioritize solving specific and achievable business problems, accumulate phased results quickly, and avoid the high-cost risk of trial and error by not blindly pursuing disruptive changes. This gradual success can not only optimize business performance but also create internal development momentum for more ambitious plans. The paradox hidden in this strategy is that when CFOs concentrate resources on AI projects oriented towards solving practical problems, they are actually indirectly promoting the precipitation of the organization's knowledge assets. This strategy may even trigger a profound change in the organizational culture, enabling the enterprise to gradually develop the ability to use AI to solve larger-scale and more complex business problems in the future.
Data collaboration platforms enable organizations to train AI models while ensuring privacy.
II. Data Dimension: Ensure Accessibility and High Quality
The second key dimension for successfully promoting the implementation of AI is data. The industry often uses a proverb to summarize this point: "garbage in, garbage out". The effectiveness of AI models highly depends on the quality and accessibility of data. However, many enterprises do not meet the requirements for supporting AI training in terms of data volume, data diversity, or data structuring.
The key lies in examining data from the perspective of "data accessibility" rather than just "data ownership". An enterprise may "own" certain data but cannot legally use it without user authorization. At the same time, there is also data that an enterprise does not own the ownership of but can access through compliant channels.
The core value of Data Collaboration Platforms is that they enable organizations to train AI models while ensuring privacy. The technical principle is to deploy algorithms at the data storage location instead of migrating the data to the organization for model training. This design ensures the secure storage of personal information at the data source while not hindering the analysis and utilization of the data.
Currently, the forms of data collaboration platforms are diverse, including both proprietary services provided by private enterprises and open-source solutions adopted by institutions or alliances. The wide application of such platforms reflects the increasing awareness of enterprises about "mining the value of shared or sensitive data while ensuring data security". More importantly, these platforms are breaking the core bottleneck in AI development: the shortage of high-quality training data.
For example, in the medical field, hospitals and pharmaceutical companies can jointly train algorithms through the platform to improve the accuracy of diagnosis or treatment without sharing the original data. In more complex B2B scenarios, limited by regulatory requirements or privacy protection regulations, enterprises often have difficulty using customer data to train AI models. Data collaboration platforms provide a solution: enterprises can promote technological innovation while strictly complying with privacy compliance requirements.
By achieving data insights while ensuring privacy, data collaboration platforms are opening up new opportunities for multiple industries, including healthcare and autonomous driving, and helping enterprises break through in the increasingly complex data regulatory environment.
The core value of artificial intelligence lies in expanding human capabilities rather than replacing humans. The core essence is that even if AI is deeply integrated into business processes, human professional judgment and accumulated experience will still be irreplaceable, especially when dealing with complex and dynamically evolving challenges.
III. Personnel Dimension: Manage Cognition and Build Trust
The third dimension, personnel, often determines the success or failure of an AI project. Due to the public's concern about being replaced by AI, AI is often regarded as a threat, and this emotion also reflects a broader public anxiety. A survey by the Pew Research Center shows that 52% of American respondents are worried about the increasing widespread use of AI in public life, and the worry outweighs the excitement.
Enterprises must respond to these concerns directly and clearly convey the core message: the core value of AI is to expand human capabilities and empower employees to achieve higher-value work. Even if AI is deeply integrated into business processes, human professional knowledge and accumulated experience are still irreplaceable, especially when dealing with complex and dynamically evolving challenges. Managing employees' perception of the transformation is crucial. If they regard AI as a threat, they may have a negative attitude and even inadvertently weaken the effectiveness of the project implementation.
Successful AI projects often attach great importance to communication and change management, realizing that the misperception of AI will significantly increase the risk of project failure. Therefore, CFOs and other senior executives need to actively communicate with stakeholders at an early stage and maintain continuous interaction to gain support, reach a consensus, build trust, and ensure the smooth progress of the transformation process.
CFO's Action Guide
The "value - data - personnel" framework aims to provide clear guidance for CFOs and other decision-makers in dealing with the complex challenges of AI. Before approving any AI project, CFOs should focus on the following three core questions:
- What kind of value do we expect to create?
- Do we have the ability to obtain high-quality data?
- How will employees view this transformation?
If the above questions cannot be clearly answered, the enterprise may need to re-evaluate its implementation strategy. When the data resources are insufficient, the enterprise should first focus on solving the problem of data acquisition or access rights instead of blindly promoting the project. Similarly, if employees lack support for the project, the enterprise must be prepared for complex and costly change management.
By adopting the "value - data - personnel" framework to implement AI projects, enterprises can improve the success rate and effectively reduce risks. However, continuously measuring the project's effectiveness is also crucial. Successful enterprises often systematically track the project results. For CFOs, any major AI investment should include a clear follow-up evaluation plan and budget. With the right strategy and continuous attention to effectiveness, AI will become the core engine for driving enterprise productivity, innovation ability, and long-term growth.
This article is translated from I by IMD. The Chinese version is for reference only. Click "Read the full text" to get the original English version.
About the Author
About the International Institute for Management Development (IMD), Switzerland
The International Institute for Management Development (IMD), Switzerland, has a history of over 75 years and has always been committed to cultivating leaders and organizations that can promote a more prosperous, sustainable, and inclusive world. Led by a professional and diverse faculty, IMD has campuses in Lausanne, Switzerland, and Singapore, and a management development center in Shenzhen, China. It is committed to becoming the most trusted learning partner for individuals and organizations with global aspirations. IMD's executive education and degree programs have long ranked among the world's leading positions. This continuous leadership stems from IMD's unique educational concept of "Real Learning, Real Impact". Through executive education courses, Master of Business Administration (MBA), Executive Master of Business Administration (EMBA) programs, and professional consulting services, we help business leaders find better and updated solutions, challenge the status quo, and inspire the future.
This article is from the WeChat official account "IMD International Institute for Management Development, Shenzhen". Author: IbyIMD. Republished by 36Kr with permission.