John Edwards, Technology Journalist & AuthorFebruary 18, 20257 Min ReadTithi Luadthong via Alamy StockBusinesses of all types and sizes are launching AI projects, fearing that failing to embrace the powerful new technology will place them at a competitive disadvantage. Yet in their haste to jump on the AI bandwagon, many enterprises fail to consider one critical point: Will the project meet its expected efficiency or profitability goal?Enterprises should consider several criteria to assess the ROI of individual AI projects, including alignment with strategic business goals, potential cost savings, revenue generation, and improvements in operational efficiencies, says Munir Hafez, senior vice president and CIO with credit monitoring firm TransUnion, in an email interview.Besides relying on the standard criteria used for typical software projects -- such as scalability, technology sustainability, and talent -- AI projects must also account for the costs associated with maintaining accuracy and handling model drift over time, says Narendra Narukulla, vice president, Quant analytics, at JPMorganChase.In an online interview, Narukulla points to the example of a retailer deploying a forecasting model designed to predict sales for a specific clothing brand. "After three months, the retailer notices that sales haven't increased and has launched a new sub-brand targeting Gen Z customers instead of millennials," he says. To improve the AI model's performance, an extra variable could be added to account for the new generation of customers purchasing at the store.Related:Effective ApproachesAssessing an AI project's ROI should start by ensuring that the initiative aligns with core business objectives. "Whether the goal is operational efficiency, enhanced customer engagement, or new revenue streams, the project must clearly tie into the organizations strategic priorities," says Beena Ammanath, head of technology trust and ethics at business advisory firm Deloitte, in an online interview.David Lindenbaum, head of Accenture Federal Services' GenAI center of excellence, recommends starting with a business assessment to identify and understand the AI project's end-user as well as the initiative's desired effect. "This will help refocus from a pure technical implementation into business impact," he says via email. Lindenbaum also advises continued AI project evaluation, focusing on a custom test case that will allow developers to accurately measure success and quantitively understand how well the system is operating at any given time.Ammanath believes that a comprehensive cost-benefit analysis is also essential, balancing tangible outcomes such as increased productivity with intangible ones, like improved customer satisfaction or brand perception. "Scalability and sustainability should be central considerations to ensure that AI initiatives deliver long-term value and can grow with organizational needs," she says. "Additionally, a robust risk management framework is vital to address challenges related to data quality, privacy, and ethical concerns, ensuring that projects are both resilient and adaptable."Related:Metrics MatterPotential project ROI can be measured with metrics, including projected cost savings, expected revenue increases, hours of productivity saved, and anticipated improvements in key performance indicators (KPIs) such as customer satisfaction scores, Hafez says. Additionally, metrics such as time-to-market for new products or services, as well as any expected reduction in bugs or vulnerabilities revealed by a tool such as Amazon Q Developer, can provide insights into an AI project's potential benefits.Leaders need to look past the technology to determine how investing in generative AI aligns with their overall strategy, Ammanath says. She notes that the metrics required to measure AI project ROI vary, depending on the implementation stage. For example, to measure the potential ROI, organizations should evaluate projected efficiency gains, estimated revenue growth, and strategic benefits, like improved customer loyalty or reduced downtime. "These forward-looking metrics offer insights into the initiatives promise and help leaders determine if they align with the business goals." Additionally, for current ROI, leaders should consider using metrics that look at realized outcomes, such as actual cost savings, revenue increases tied directly to AI initiatives, and improvements in key performance indicators like customer satisfaction or throughput.Related:Pulling the PlugIf an AI project consistently fails to meet expectations, terminate it in a calculated manner, Hafez recommends. "Document the lessons learned and the reasons for failure, reallocate resources to more promising initiatives, and leverage the knowledge gained to improve future projects."Once a decision has been made to end a project, yet prior to officially announcing the ventures termination, Narukulla advises identifying alternative projects or roles for the now-idled AI team talent. "In light of the ongoing shortage of skilled professionals, ensuring a smooth transition for the team to new initiatives should be a priority," he says.Narukulla adds that capturing key learnings from the terminated project should be a priority. "A thorough post-mortem analysis should be conducted to assess which strategies were successful, which aspects fell short, and what improvements can be made for future endeavors."Narukulla believes that thoroughly documenting post-mortem insights can be invaluable for future reference. "By the time a similar issue arises, new models and additional data sources may offer innovative solutions," he explains. At that point, the project may be revived in a new and useful form.Parting ThoughtsEstablishing a strong governance framework for all ongoing AI projects is essential, Hafez says. "Further, a strong partnership with legal, compliance, and privacy teams can enhance success, particularly in regulated industries." He also suggests collaborating with external partners. "Leveraging their expertise can provide valuable insights and accelerate the AI journey."When implemented and scaled properly, AI is far more than a technological tool; it's a strategic enabler of innovation and competitive advantage, Ammanath says. However, long-term success requires more than sophisticated algorithms -- it demands cultural transformation, emphasizing human collaboration, agility, and ethical foresight, she warns. "Organizations that thrive with AI establish clear governance frameworks, align business and technical teams, and prioritize long-term value creation over short-term gains."As AI continues to advance and evolve, IT leaders have an unprecedented opportunity to align investments with enterprise-wide goals, Ammanath says. "By approaching AI as a strategic lever rather than a standalone solution, organizations can position themselves at the forefront of innovation and value creation."Read more about:Cost of AIAbout the AuthorJohn EdwardsTechnology Journalist & AuthorJohn Edwards is a veteran business technology journalist. His work has appeared in The New York Times, The Washington Post, and numerous business and technology publications, including Computerworld, CFO Magazine, IBM Data Management Magazine, RFID Journal, and Electronic Design. He has also written columns for The Economist's Business Intelligence Unit and PricewaterhouseCoopers' Communications Direct. John has authored several books on business technology topics. His work began appearing online as early as 1983. Throughout the 1980s and 90s, he wrote daily news and feature articles for both the CompuServe and Prodigy online services. His "Behind the Screens" commentaries made him the world's first known professional blogger.See more from John EdwardsNever Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.SIGN-UPYou May Also LikeWebinarsMore WebinarsReportsMore Reports