Organisations of all sizes are considering and preparing for AI adoption to streamline operations, reduce human error, and take over easy tasks. But, AI introduction is a time-consuming and resource-dependant process.
Big Tech players – including Alphabet, Apple, Amazon, and Meta – are making big investments in AI and have more time, people, and money to introduce, refine, and establish AI processes within their infrastructure. Most businesses don’t have these same resources to support AI introduction and enhance their informational ecosystem.
While we see Big Tech rapidly investing in and implementing AI, among the everyday enterprises, it’s the smaller companies who have an easier time getting started. The bigger the company, the more time, money, and personnel are needed to audit and prepare for new technology adoption.
Another limiting factor for larger enterprises in deploying AI is their vast amounts of data. Data is key to the AI implementation process, and larger enterprises have far more of it to juggle. Larger data warehouses mean higher costs and increased complexity when introducing new AI solutions.
Companies Are Considering AI, But Many Are Struggling to Deploy It
AI is bound to be a transformative technology, and its initial introduction to work shows how it can eliminate mundane tasks and reduce human error in simple processes. But, the technology still has a way to go before it’s adopted and introduced across organisations of all sizes.
AI readiness spending is rising across industries, as companies look to innovative technologies to increase their competitiveness. According to an MIT Tech Review report (A Playbook for Crafting AI Strategy), overall AI spending in 2022 and 2023 was modest or flat for most companies, with only one in four increasing their spending by more than a quarter. That changed in 2024, with nine in ten respondents expecting to increase AI spending on data readiness and in adjacent areas like strategy, cultural change, and business models.
Despite investing more in AI readiness, companies are still struggling to fully deploy the technology – only 76% have fully deployed AI in just one to three use cases. Nearly all organisations (98%) say they are willing to forgo being the first to use AI if that ensures they can deliver it safely and securely. Governance, security, and privacy are the biggest brake on the speed of AI deployment, cited by 45% of respondents (and a full 65% of respondents from the largest companies).
Data quality is another highly limiting data issue in deployment. This is especially true for larger firms with more data and substantial investments in legacy IT infrastructure. Companies with revenues of over $10 billion are more likely to cite both data quality and infrastructure as limiters, which suggests organisations presiding over larger data repositories find the problem substantially harder.
Data Audits and Availability Are Key Attributes To Successful AI Implementation
To prepare their infrastructure for AI technologies, organisations must first audit their informational ecosystem to identify gaps and weak spots. The more clearly an organisation understands how their information is passed, protected, and shared, the better prepared they’ll be and the fewer mistakes they’ll make when introducing AI.
Data quality audits should happen routinely. But, for AI technology, it’s essential to conduct a pre-audit and continual audits to ensure AI models are developing effectively and producing accurate, complete, and reliable results.
These audits provide an understanding of how an organisation and customer data is shared, stored, and protected. This becomes even more important for sectors with strict compliance and regulatory requirements.
In tandem with audits, organisations should introduce a data integration and governance strategy that includes insight into data stewardship and policy management. When developing a data integration strategy, it’s critical to identify and catalogue data sources, establish secure and systemic transmission methods, and develop a centralised data lake with human oversight of data quality and stewardship to help consolidate and manage information. Utilising these approaches will help reduce disruptions in the digitisation process, enhance overall accessibility to informed decision-making, and increase information analysis within an organisation.
Data shareability is also required for successful adoption, so any team within an organisation can access and analyse data for their specific business scenarios. According to IDC, only 12% of organisations connect customer data between departments. However, AI is proving to be a catalyst for connection and collaboration to improve and accelerate operational efficiency at scale.
Essentially AI can make sense of unstructured data in ways that simply weren’t possible before. This is exactly the reason why now is the time for enterprises to take data governance and stewardship seriously.
Introducing AI At Your Company
Large and small enterprises can both take steps to ensure their data is setting them up for success in AI deployment. For those who have begun to introduce AI:
- Clearly define your objectives and goals for AI initiatives.
- Develop and execute a clear data strategy. Data quality, governance, management, and stewardship are critical to get practical use for and greater returns from AI-infused uses. We know garbage-in/garbage-out. However, the scale of garbage processing with AI could be insurmountable in the near future for enterprises that do not implement and adhere to rigorous data management principles now.
- Emphasise the importance of gathering feedback from established AI models. This includes auditing inputs, monitoring produced metrics, and analysing produced results to gauge if/where refinement is needed. This feedback loop supports ongoing performance updates and ensures results are accurate.
- Involve stakeholders and data stewards in the AI digitalisation process to help foster deeper understanding and collaboration internally, and help identify potential roadblocks earlier in the implementation process for a smoother transition and increased success potential.
- Highlight how AI is currently used to automate/streamline data management to reduce human error. For example, AI agents for social media can proactively create posts but rely on input data to do so, making it crucial to follow data management best practises and exercise proactive compliance to minimise errors.
- Evaluate the shelf-life of data to help inform action-based decisions. Without considering the shelf-life of data, decision-making can become outdated and ineffective. AI is maximised when the technology can access a wealth of real-time data. The importance of a data’s shelf-life significantly increases when paired with AI technology.
Overall, one thing is clear for AI adoption: company size matters. Larger enterprises are finding that their wealth of resources is actually inhibiting them, and smaller enterprises are finding that less may actually be more when it comes to AI deployment. For both, a strong data foundation will take them to the next level. By proactively monitoring datasets to spot outdated and inaccurate information, disorganisation, and areas for improvement, enterprises will find greater success in AI implementation – no matter their size.
