Set your portco up for success with more informed decision making and AI success.
(This post is part one in a two-part series on achieving optimized results with emerging AI technologies.)
Most companies today are sitting on a gold mine. It’s the nuggets of data they’ve gathered over time. These include their operational activities, client behaviors, employee productivity, market performance, and more. But having data is not the same as having useful information, intelligence, and insights. With data, what counts is its quality.
AI and machine learning algorithms are designed with one purpose — to extract deep insights from data inputs. They are machines. Data is their raw ingredient. No matter how intelligent the machines are, if the raw ingredients are flawed, the output will be too.
Every company is responsible for the quality of its data inputs. Maintaining high-quality requires data literacy. With the rapid expansion of AI underway, data literacy is a critical skill set.
According to Gartner, data literacy is the underlying component of a company’s ability to use existing and emerging technologies. Poor data literacy is ranked as the second-biggest internal roadblock to chief data officers’ success.
Without data literacy, companies are at risk of experiencing several common challenges from AI technologies, all of which create bad outcomes, including:
- Incorrect data
- Scarce data
- Bad data
- Biased data
How to Expand Your Portco’s Data Literacy
Cortado Group’s Revenue Operations Practice Leader, Robert Gammon, works with multiple portfolio company leaders every year. Robert helps companies extract greater intelligence and revenue potential from their data. There are nine things he wishes every company understood about AI and data literacy — from prioritizing your data house to encouraging your entire team to uplevel their data literacy.
Understand the fundamentals of how to boost your company’s data literacy — so you can extract the best possible insights from AI-driven technologies. Below, we highlight the nine key insights Robert Gammon emphasizes for achieving superior data literacy and leveraging AI effectively:
1—Prioritize getting your data house in order
Among the findings in its AI Priorities Study 2023, Foundry discovered that 56% of IT decision-makers are eager to learn more about generative AI. They want to deploy AI in a variety of ways, including chatbots and virtual assistants (56%), content generation (55%), industry-specific applications (48%), data augmentation (46%), and personalized recommendations (39%).
For companies that don’t yet have their data houses set up to optimize AI, they should follow these data transformation steps to begin getting their data house in order:
- Define your data strategy: Crystalize the goals of your data-driven objectives, including how your organization will collect, store, manage, and use data to achieve your business goals.
- Design your target state and roadmap: Assess your existing information and architecture. Find gaps that need to be filled to achieve your future-state data design, quality, security, and governance.
- Mobilize your data transformation: Test your data model in a series of pilot projects to refine your use cases, address any issues, and instill new ways of working.
2—Understand the difference between good and bad data
According to multiple studies, including those published by the Harvard Business Journal, the cost of bad data can be staggering:
Data quality refers to how accurate, complete, consistent, and relevant your data is for meeting your AI goals. High-quality data provides AI systems with the nuggets of gold that produce accurate predictions and deliver meaningful recommendations. Without high-quality data, even the most advanced AI algorithms can’t provide valid insights. With quality data for AI initiatives, portcos can gain valuable, actionable insights for smarter decision-making.
Common sources of bad data that can derail AI initiatives include:
- Inaccurate data
- Mislabeled data
- Data from unknown sources
- Incomplete data sets
- Inadequate data collection methods
- Biased methods for data collect
3—Find gaps in your data and fill them in with trusted sources
Too often, companies have gaps in their data. The result can mean either missing out on the full scope of available information or risking the possibility of receiving poor results. The bottom line result can be significant financial losses. The most recent research from Gartner finds that poor data costs organizations an average of $15 million annually.
Knowing what you don’t know is essential to finding your data gaps. Finding gaps is the hard part. Filling gaps is easier. Companies can fill their gaps with data from trusted sources, such as:
- Surveys
- Transactional data
- Customer data
- Online tracking websites
- Online marketing analytics
- Software tools
4—Know the sources of your commercial data in great detail
According to a Forrester study, the most important factor in whether companies will do business with you is trust. Buyers are nearly twice as likely to recommend a business to colleagues if they trust the organization than if they don’t (85% versus 48%).
One way to build trust is to trust the sources of your data. Knowing where your data came from is critical for gaining confidence in the output of AI-driven technologies.
The following definitions help explain this step on the road to data literacy:
5—Understand the age, history, and relevance of your data
In a study on AI and data, the Enterprise Strategy Group found that, at 31%, the number one challenge business leaders say they are facing in implementing AI-driven solutions is a limited availability of quality data to feed into AI models.
Among the factors that make data right for AI models is age. Historical data provides context and a wealth of knowledge on which to train a model. Understanding your data’s age, history, and relevance can greatly improve your ability to extract meaningful patterns, such as performance trends, and their shifting nature over time.
To understand your data’s age, these steps are a good place to start:
- Know when your data is too old to be relevant for understanding your current business trends.
- Work with your IT team to understand your data retention policies. Help craft a policy that ensures your business decisions are made based on relevant data.
- Ensure your AI models are being trained using accurate data to support better decision-making.
6—Ensure you are interpreting your data accurately
The Enterprise Strategy Group also found that leaders have high expectations for generative AI to provide valuable data insights. Achieving these expectations requires proper interpretation of AI results.
Without knowing why you’re seeing certain patterns in your data, you might try to solve the wrong problems. Or just as bad, solve the right problems incorrectly. Data literacy means knowing how to interpret your AI systems’ output accurately, according to factors, such as these:
- Data context and domain knowledge: Data context refers to the background information and relevant details of a dataset. Domain knowledge is comprehending how things function in a specific field.
- Statistical competence: The foundation is formed by having knowledge of related concepts and critical thinking, including understanding these three variables:
- Correlation: It’s commonly said that correlation is not causation. Just because variables seem to be related, does not mean they are impacted by one another.
- Interpreting visualization: Visual displays of information to communicate complex data relationships and data-driven insights in unique and meaningful ways. These may include charts, graphs, animations, and more.
- Recognizing patterns: Finding patterns in data is the heart of AI. Data literacy means being able to interpret AI’s pattern identification.
7—Establish a data governance system
Data governance is a framework and process for managing data assets throughout their lifecycle. It’s not about locking down data, but making it more available for secure, authorized users to access.
To achieve this outcome, data governance defines factors for success. These include management roles and responsibilities, data standards, and processes for data collection, storage, and use. A holistic data governance system includes activities, processes, and technologies that ensure data security, privacy, cleanliness, accuracy, availability, recency, and usability.
In a Snowflake data study, it was found that company leaders are deploying 70% to 100% more governance measures around data and a more refined approach including tagging standards and features.
To establish data governance requires implementing tools and platforms for data strategy, stewardship, cataloging, and lineage. The result is critical for effective AI, because it helps:
- Demonstrate accountability
- Maintain quality assurance
- Ensure transparency
- Mitigate model bias
- Assist in scaling and adoption
- Support compliance regulation and trust
- Support ethical AI use ethical
Components of a data governance program, include:
- Data cleansing: Clean and filter your data to ensure it’s not skewing your analysis due to duplicates, errors, unwanted outliers, and missing data.
- Data normalization: The process of transforming data into a standard format so that it can be compared and analyzed more easily.
- Data validation: Checking data against a set of rules to identify accuracy and errors.
- Data consistency: Identifying and removing inconsistencies from data, like duplicate records, misspellings, and incorrect data types.
- Data augmentation: Enhancing your data’s diversity and representativeness.
8—Continually measure and adjust data for AI
To optimize the potential of AI requires adopting an attitude of continuous improvement. Organizations that adopt a data-first AI strategy, rooted in their culture and operations, can make accurate, timely, and valuable decisions faster. This starts culture from the top and includes changing mindsets. Every department can achieve data excellence and consistently make data-centric decisions with confidence.
In a recent study by Varicent, it was found that high-performing teams prioritize data accuracy. Accurate and current data is essential for successful Go-to-Market planning and improving strategic quality:
9—Encourage your entire team to uplevel their data literacy
A successful AI launch requires not just the technology, but also a human touch. For optimal performance, companies must balance human and technology resources. This requires a widespread culture of data literacy, which many organizations are already cultivating.
In a study by Tableau, it was found that 82% of decision-makers expect basic data literacy from employees in every department, including product, IT, HR, and operations. By 2025, nearly 70% of employees are expected to use data heavily in their jobs, up from 40% in 2018.
Elevate your entire teams’ data literacy, so everyone understands the value of data and AI and uses data at a high level. Two ways to achieve this outcome are:
- Bring in data expertise: Have a lead person take charge of your data literacy initiative, such as a data engineer or scientist, or a machine learning engineer.
- Develop a data literacy strategy: Get employee buy-in and trust for your portco’s data strategy. Create systems for learning and training how to work with data and AI outputs to achieve optimal results.
Summary
Turning data into knowledge and knowledge into insights gives tremendous positive momentum to profitable growth. Inspire your portco to commit to transforming into an insights-driven enterprise backed by robust data literacy. You’ll empower them to consistently outperform your competitors — and meet your revenue growth targets tomorrow.
This article is part one of a two-part series on achieving optimized results with emerging AI technologies. Don’t miss part two, where we’ll explore more advanced strategies and practical applications.
For personalized guidance from Robert Gammon on how to enhance your portco’s data literacy and AI capabilities, schedule a coffee chat with Cortado. Set up a time to discuss your needs with Robert and start your journey towards data-driven success.
