
How to Create a Winning AI Strategy
www.informationweek.com
Lisa Morgan, Freelance WriterMarch 3, 20258 Min ReadBrain light via Alamy StockArtificial intelligence continues to become more pervasive as organizations adopt it to gain a competitive advantage, reduce costs and deliver better customer experiences. All organizations have an AI strategy, whether by design or default. The former helps ensure the company is realizing greater value, simply because its leaders are putting more thought into it and working cross-functionally to make it happen, both strategically and tactically.Its very much back to the business, so what are the business objectives? And then within that, how can AI best help me achieve those objectives? says Anand Rao, distinguished service professor, applied data science and artificial intelligence atCarnegie Mellon University. From there, [it] pretty much breaks down into two things: AI automates tasks so that you can be more efficient, and it helps you make better decisions and with that comes a better customer experience, more revenue, or more consistent quality.Elements of a Winning AI StrategyKevin Surace, CEO at autonomous testing platform Appvance, says the three elements of an effective AI strategy are clarity, alignment, and agility.A winning AI strategy starts with a clear vision of what problems youre solving and why, says Surace. It aligns AI initiatives with business goals, ensuring every project delivers measurable value. And it builds in agility, allowing the organization to adapt as technology and market conditions evolve.Related:Will Rowlands-Rees, chief AI officer, at eLearning, AI services, and translation and localization solution provider Lionbridge agrees.It is critical to align your AI strategy and investments with your overall business strategy -- they cannot be divorced from each other, says Rowlands-Rees. When applied correctly, AI is a powerful tool that can accelerate your organizations ability to solve customer problems and streamline operations and therefore drive revenue growth. This offensive approach will organically lead to cost optimization as efficiencies emerge from streamlined processes and improved outcomes.Brad O'Brien, partner at global consultancy Baringa's US Financial Services practice, advocates having a clear governance framework including the definition of roles and responsibilities, setting guiding principles, and ensuring accountability at all levels.Comprehensive risk management practices are essential to identify, assess, and mitigate AI-related risks, including regular audits, bias assessments and robust data governance, says OBrien. Staying informed about, and compliant with, evolving AI regulations, such as the EU AI Act and emerging US regulations, is vital. Maintaining transparency and thorough documentation of the entire AI lifecycle builds trust with stakeholders. Engaging key stakeholders, including board members, employees and external partners, ensures alignment and support for AI initiatives. Continuous improvement, based on feedback, new data and technological advancements, is also a critical component.Related:Ashwin Rajeeva, co-founder and CTO at enterprise data observability company Acceldata, believes a successful AI strategy blends a clear business vision with technical excellence.It starts with a strong data foundation; reliable, high-quality data is non-negotiable. Scalability and adaptability are also critical as AI technologies evolve rapidly, says Rajeeva. Ethical considerations must be embedded early, ensuring transparency and fairness in AI outcomes. Most importantly, it should create tangible business value while maintaining the flexibility to adapt to future innovations.How to Avoid Common MistakesOne mistake is assuming that generative AI replaces other forms of AI. Thats incorrect because traditional types of AI -- such as computer vision, predictions, and recommendations -- use different types of models.Related:You still need to look at your use cases and standard methods. Look across the organization, look at the value chain elements, and then look at where traditional AI works and where generative AI would work, and what some of the more agent kind of stuff would work, says CMUs Rao. Then, essentially start pulling all of the use cases together and have some method of prioritizing.The accelerating rate at which AI technology is advancing is also having an effect because companies cant keep up, so organizations are questioning whether they should buy, build or wait.Change with respect to AI, and especially Gen AI, is moving very fast. Its moving so much faster that even the technology companies can keep pace, says Rao.AI is also not a solution to all problems. Like any other technology, its simply a tool that needs to be understood and managed.Proper AI strategy adoption will require iteration, experimentation, and, inevitably, failure to end up at real solutions that move the needle. This is a process that will require a lot of patience, says Lionbridges Rowlands-Rees. [E]veryone in the organization needs to understand and buy in to the fact that AI is not just a passing fad -- its the modern approach to running a business. The companies that dont embrace AI in some capacity will not be around in the future to prove everyone else wrong.Organizations face several challenges when implementing AI strategies. For example, regulatory uncertainty is a significant hurdle and navigating the complex and evolving landscape of AI regulations across different jurisdictions can be daunting.Ensuring data privacy and security is another major challenge, as organizations must protect sensitive data used by AI systems and comply with privacy laws. Mitigating biases in AI models to prevent unfair treatment and ensure compliance with anti-discrimination laws is also critical, says Baringa's OBrien. Additionally, the 'black box' nature of AI systems poses challenges in providing clear explanations of AI decisions to stakeholders and regulators. Allocating sufficient resources, including skilled personnel and financial investment, is necessary to support AI initiatives.In his view, common mistakes in AI strategy implementation include:A lack of clear governance frameworks and accountability structures.Insufficient risk management practices, such as overlooking comprehensive risk assessments and bias mitigation.Poor data management, including neglecting data privacy and security that can lead to potential breaches and regulatory non-compliance.Inadequate transparency in documenting and explaining AI processes results in a lack of trust among stakeholders.Underestimating resource needs, such as not allocating sufficient skilled personnel and financial investment, can hinder AI initiatives.Encountering resistance from employees and stakeholders who hesitate to embrace AI technologies is a common challenge.[P]rioritize governance by establishing clear frameworks and ensuring accountability at all levels. Stay informed about evolving AI regulations and ensure compliance with all relevant standards, says OBrien. Focus on transparency by maintaining thorough documentation of AI processes and decisions to build trust with stakeholders. Invest in regular training for employees on AI policies, risk management, and ethical considerations. Engage key stakeholders in the design and implementation of AI initiatives to ensure alignment and support. Finally, embrace continuous improvement by regularly updating and refining AI models and strategies based on feedback, new data and technological advancements.One of the biggest mistakes Shobhit Varshney, VP and senior partner, Americas AI leader, IBM Consulting has observed organizations selecting AI use cases based on speed of implementation rather than properly articulated business impact.Many organizations adopt AI because they want to stay competitive, but they fail to realize that they aren't focusing on the use cases that will create significant long-term value. It's common to start with simple, easy-to-automate tasks, but this approach can be limiting, says Varshney. Instead, organizations should focus on areas where AI can have the greatest impact and have enough instrumentation to capture metrics and continuously iterate and evolve the solution. The best starting point for AI use cases is unique to each business and its important to identify areas within the organization that could benefit from improvement.He also says an all-too-common mistake is automating an existing process.We need to rethink workflows to truly unlock the power of these exponential technologies. As we evolve to agentic AI, we need to ensure that we rethink the optimal way to delegate specific tasks to agents and play to the strengths of humans and AI, says Varshney.Jim Palmer, chief AI officer at AI-native business and customer communications platform Dialpad, says a common challenge is ensuring AI models have access to accurate, up-to-date data and can seamlessly integrate with existing workflows.Theres a gap between AIs theoretical potential and its practical business application. Companies invest millions in AI initiatives that prioritize speed to market over actual utility, Palmer says.Bhadresh Patel, COO of global professional services firm RGP thinks one of the biggest challenges organizations is the significant gap between ideation and execution.We often see organizations set up an AI function and expect miracles, but this approach simply doesn't work. This is why it's important to prioritize the pockets of use cases where AI can have the biggest impact on the business, says Patel. Another challenge organizations often face is when functional people do not take the time to understand the capabilities and limitations of the tools they have at their disposal. Leaders must understand why theyre making new AI investments and what the overlap is in terms of existing capabilities, training and user knowledge.Acceldatas Rajeeva says organizations often grapple with fragmented or poor-quality data, which undermines AI outcomes.Scaling AI initiatives from proof of concept to enterprise-wide deployment can be daunting, especially without robust operational frameworks. Additionally, balancing innovation with regulatory and ethical standards is challenging. A lack of skilled talent and clear success metrics further complicates these efforts, says Rajeeva. One significant misstep is treating AI as a technology-first initiative, ignoring the importance of data quality and infrastructure. Organizations sometimes over-invest in sophisticated models without aligning them with practical business goals. Another common mistake is failing to plan for scaling AI, leading to operational bottlenecks. Finally, insufficient monitoring often results in biased or unreliable AI systems.And remember, foresight and agility are more valuable than 20-20 hindsight.Start with the end in mind. Define success metrics before you write a single line of code. Build cross-functional teams that can bridge the gap between business and technology, says Appvances Surace. And remember, an AI strategy isnt static -- its a living, evolving framework that should grow with your organization and its goals.About the AuthorLisa MorganFreelance WriterLisa Morgan is a freelance writer who covers business and IT strategy and emergingtechnology for InformationWeek. She has contributed articles, reports, and other types of content to many technology, business, and mainstream publications and sites including tech pubs, The Washington Post and The Economist Intelligence Unit. Frequent areas of coverage include AI, analytics, cloud, cybersecurity, mobility, software development, and emerging cultural issues affecting the C-suite.See more from Lisa MorganWebinarsMore WebinarsReportsMore ReportsNever Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.SIGN-UPYou May Also Like
0 Comments
·0 Shares
·32 Views