20 01732 759725 There’s an inescapable hype that surrounds artificial intelligence (AI). Around one-in-six UK organisations, or 432,000 firms, have embraced at least one AI technology. From automating repetitive tasks to powering medical diagnosis tools, the use cases are vast. But it’s not as simple as buying an off-the-shelf solution, powering it on and then reaping the benefits. AI is no silver bullet and systems are only as valuable as the data that underpins them. Poor data quality can create a range of issues for AI-driven tools, from misguided learning algorithms to exacerbation of social biases. If organisations are to achieve maximum benefit from their investment in AI, they will first need to prepare their data. Achieving AI-readiness AI and machine learning is the area of greatest investment for 26% of IT leaders, according to CIO’s State of the CIO 2023 report, and to ensure that investment doesn’t fall flat they must be AI-ready. That means gleaning value from their own data environment, rather than just snapshots of it. The problem is that too many firms have essentially locked their data down so that no member of staff can access it. This might help avert data breaches, but it also prevents people from opening a treasure chest of valuable Preparing data ahead of the AI revolution insights. According to IDC’s Data in Context 2022 report, by 2025 most organisations will use less than 5% of available data for decisionmaking. Building AI solutions on a mishmash of hidden data can have harmful consequences, with duplications and incomplete datasets creating biased or inaccurate insights and decisionmaking. For example, in 2018, Amazon was forced to scrap an AI-driven recruitment tool because it inadvertently favoured male candidates. The reason? The system had been trained on data submitted mostly by men over a 10-year period. In addition, if data isn’t labelled correctly, algorithms can’t learn from relevant information that will help them accurately automate tasks that humans don’t have the time or skill to complete. Data needs to be unleashed from its hiding place so that businesses know exactly what they have and where value can be derived. Achieving data fitness Achieving data fitness starts with a data assessment to find where every piece of data is hidden within the technology ecosystem. This process can also identify what type of storage every iteration of data sits on, how long it should be retained and when a staff member last accessed it. AI solutions need accurate and clean data to perform tasks effectively and an audit will enable fragmented data that is hidden in the shadows to be moved to a more appropriate location. Data that is most needed for AI can be relocated to more accessible storage, with lower priority data shifted to cheaper storage. In addition, any unnecessary duplicate files can be deleted. This process will give learning algorithms an accurate source of information to leverage and apply to automated actions, while the application of an effective governance structure with set rules at the final stage can help provide a reliable pipeline of useful data for those significant AI investments. The value of implementing a data fitness strategy is that it allows staff to adopt a selfservice approach in relation to their organisation’s data. Having connected datasets, rather than data sitting in siloes, gives users the ability to find answers to their questions via an AI-driven tool, with internal and external data (and metadata-based labels) providing relevant context. Another benefit of self-service is that it can help businesses avoid or reduce costs associated with highly specialised data scientists. Unlocking the potential While the excitement surrounding AI is palpable, its true potential can only be unlocked through robust data management. As AI becomes more mainstream and integral to business operations, organisations must ensure their data is clean, accessible and well governed. An effective level of data fitness can eradicate errors in AI deployments and also empower businesses to make smarter, faster decisions. Companies that invest in proper data management today will find themselves in a stronger competitive position in the coming years, while those that neglect it will struggle to keep pace in an increasingly AI-driven world. www.camwood.com Andrew Carr, Managing Director of Camwood, a leading IT consultancy with expertise in applications, modern workplace and managed services, highlights the importance of data fitness for effective AI DATA MANAGEMENT Andrew Carr
RkJQdWJsaXNoZXIy NDUxNDM=