Have summaries of our latest blogs delivered to your inbox, so you can stay up to date on the topics and current events that matter to your business.
From machine learning to generative AI, the latest developments in AI technology offer a major competitive advantage to companies who can best apply them to their business’ challenges. But those...
Companies are increasingly moving to an API-first approach in which they start any AI or digital project by focusing on the API they will use to integrate the data required. In this blog, we look at...
With the start of Giving Season just around the corner, we’ve pulled together some tips on trends and prospect research tricks to help you make the most of this year’s giving season and move...
Many people get started in nonprofits because they are very passionate about the cause. This means that the work is incredibly meaningful and can be very fulfilling. However, with funding and staffing...
Storytelling is a powerful tool: even just going to see a movie or turning on the television can provide a sense of escapism, boost our empathy for fellow humans and change the way we see the world. Apply...
Nearly 9/10 executives consider investing in AI and data to be a top priority for their company, yet 8/10 of those initiatives are likely to end in failure.
In this post, we explore the key reasons behind why AI projects do not achieve their aims. We then draw lessons from those failures to help companies exploit the vast opportunities of AI, with help from credible data and technology from LexisNexis®.
“Move fast and break things” was Facebook’s early motto. As CEOs watch their competitors bringing in emerging AI technologies and large data sets to power them, their instinct may be to move as quickly as possible to catch up or pull ahead. But this excitement should be tempered by the unfortunate fact that as many as four in five AI projects end in failure, according to a report in the Harvard Business Review.
Each project is different, but there are eight common reasons why an AI initiative fails to achieve its objectives:
Too many companies focus their energy (and resources) on acquiring new technology, with data as an afterthought. But data that is inaccurate, unprovenanced, biased, outdated, or partial will replicate all these problems in AI’s outputs.
AI projects often bring in data from various sources using a confusing mix of bulk data deliveries and APIs. Moreover, each data set may be structured (or unstructured) in a different way and require significant work to clean and ensure it can be used in the company’s chosen analytics software. An API-first model, which starts by looking at how data will be ingested into technology via an external API, can help to overcome this issue.
MORE: How third-party data helps your organization gain a market advantage
Companies often fail to use AI in support of their overall strategy. Each project should have clear objectives and attempt to solve a relevant and achievable problem and recognize that AI may not always be the best method for tackling some business needs. Author Bernard Marr studied failed AI projects and wrote in Forbes that “one thing they have in common is they are all caused by a lack of adequate planning”.
The MIT Sloan Management Review reviewed AI and big data projects in three large banks in India and found most problems “invariably occurred at the interfaces between the data science function and the business at large”. In other words, people with the expertise to implement the technology were not integrated into the day-to-day activities of the business and silos developed between the two sides.
MORE: Top 5 ways professional services teams are using generative AI
Companies using AI and big data analytics must ensure the data sets they use conform to regulatory standards of compliance and ethics. A report by IBM found the average cost of a data breach was $4.45 million in 2023. Misuse of data is also likely to deter customers, investors, and employees from continuing their relationship with a company.
Companies often overlook the need adequately train staff in how to use new technology. If the staff on the ground cannot see why or how an AI tool improves their existing way of working, they will simply not change their processes. Time and resource must also be freed up to give staff time to learn and ask questions.
MORE: From start to finish: Your checklist for responsible AI
Sometimes companies bring in AI initiatives but, without transparent communications about what they are doing, customers are reluctant to engage with because of perceived risks. For example, a company developed an app to offer customers an instant recommendation for a service and the option to sign up immediately. While this appeared to offer the firm a competitive advantage by reducing onboarding time, the uptake from customers was low. Surveys later found this was because many users doubted the credibility of the service and its use of AI.
Companies need expertise to identify and mitigate bias in technology. For example, a bank developed an algorithm for loan decisions, but it was trained on data from pre-screened applications, which were more likely to result in successful repayments. This bias in the underlying data led to a higher-than-expected failure rate for loans approved by the algorithm.
MORE: The future of company research: Leveraging AI for comprehensive analysis
Many of the 8 reasons for failure come back to one issue: the data powering AI projects, and the way that data is delivered. A company’s best chance of embracing the opportunities of AI is to use only the highest-quality, enriched data which is sourced in a legally and ethically compliant manner and delivered through a flexible API.
Firms should seek a wide range of comprehensive data sets, including data on news, legal, company and financial, biographical, IP, social media, and more. But more data is not always best, if that includes data which is inaccurate or outdated. Companies must understand the provenance of their data sources and the rights and risks around its use. It is best to engage a trusted provider which has forged partnerships with many content providers and pulls on original sources.
The tone and strategy set by a company’s leadership is also important. 93% of tech and data executives say a data strategy is critical to getting value from generative AI, according to a 2024 survey by Amazon Web Services. Before diving into using AI and big data analytics, consider which of your firm’s core strategic objectives would best be targeted through technology, or which of your challenges (or your customers’ challenges) could be overcome.
MORE: AI for business research: Unlocking new insights and opportunities
The success of any AI initiative hinges on access to high-quality, credible data from trusted sources. Data and technology from LexisNexis® offers key advantages to help you realize AI’s opportunities and overcome its risks:
Our powerful combination of credible, licensed content and sophisticated technology can transform your AI and generative AI initiatives and set you up for success. Contact us today to learn more about how our trusted and ethical data and AI solutions can drive your business forward.
Or download our free ebook, Harnessing Data for AI Innovation, to learn more about the how your company can exploit AI’s opportunities and manage its risks with high-quality data.