Ajith Sankaran, Senior Vice President, Course5 Intelligence.
Perhaps no other technology in recent times has garnered so much interest and generated so much hype as ChatGPT and generative artificial intelligence. Amid all this hype, what I believe is being missed is the potential generative AI has in terms of delivering value to business processes and providing a significant impetus to faster technology-led growth of businesses.
I believe one such area where generative AI could have a significant impact is in the space of data analytics. I am a senior vice president at a data analytics company that has started leveraging generative AI in our analytics solutions. This experience has shown me that while it is still early days, and many of the possibilities of generative AI in data analytics are yet to emerge, there are already clear indications of the possible areas of impact.
The Data Analytics Value Chain And Potential Areas of Impact
1. Getting The Data: Despite all the advances in automation and AI, data collection and data engineering continue to be among the most time-consuming elements in the data analytics value chain. Specific aspects such as data classification, tagging, data cleaning, etc. can be resource-intensive and require significant human effort and intervention.
This is an area where I believe generative AI could bring efficiencies by leveraging large language models. Generative AI can be used to append external industry classification data to customer relationship management data sets, which would help cut down processing times. Data infrastructure providers would have to reorient their solutions and strategies to align with the possible disruptions from generative AI.
2. Analyzing The Data: There have been a lot of discussions already around the ability of generative AI and tools like ChatGPT to create software code to build analytic models. According to GitHub, developers who use GitHub Copilot (built on OpenAI’s Codex) are able to gain time savings, with 88% of respondents indicating they are more productive and 96% indicating that they are “faster with repetitive tasks.”
Going beyond time savings, I believe the ability of generative AI to bring in a broader business context while automating coding could be a real game changer in analytics. There are other use cases as well, such as the ability of generative AI to create training and synthetic data to build supervised learning data sets for training AI and machine learning models.
3. Generating Insights: Even today, generating insights that drive business decisions from data analytics often results in a largely manual activity, despite all the talk about “automated insights.” I’ve noticed existing analytics and applied AI techniques can, at best, deliver basic levels of customized insights from data at the level of descriptive analytics and, in some cases, diagnostic analysis. Generating predictive and prescriptive analytics insights is still largely driven by humans.
From my perspective, generative AI could potentially leverage contextual data and then mimic human inferencing processes to contextualize analytics results and bring out actionable insights. Moreover, generative AI could drive persona-based contextualization of insights, which could uplevel the impact of data analytics.
4. Delivering Insights And Driving Decisions: Creating analytics reports and impactful business intelligence outputs could also be positively impacted by generative AI. Specifically, when it comes to automation of reporting with contextualization, generative AI can make a big impact. Another area of impact would be delivering near-to-real-time sights by cutting off the need for human intervention to add contextual oversight.
The Way Ahead
At a broader level, a Harvard Business Review article highlighted that the adoption of generative AI would require a fundamental restructuring of economic systems. This is also highlighted by Accenture, which said generative AI will herald “total enterprise reinvention.”
While the future is exciting and holds many promises, organizations also need to be wary of the limitations as well as the challenges of generative AI, including:
• Data security and privacy: This is the biggest challenge. In a recent study by Salesforce, more than 70% of IT leaders said “generative AI will introduce new security risks to our data.”
• Bias and ethics: Bias has been one of the key challenges with AI, and it will be the same for generative AI. As indicated in a Wall Street Journal article, industry leaders are taking this very seriously, and some are even choosing not to use AI for certain applications due to the chance of bias.
• IP risks: I’ve already seen concerns being raised about critical IP being put out in the public domain by employees trying to leverage large language models.
• Accuracy and “black-box” issues: The results of any AI system, including analytical outputs from generative AI applications, might not always be accurate. Moreover, in systems such as ChatGPT, there is no way to examine what data was used and what models were applied.
How To Begin Using Generative AI For Data Analytics.
1.If you use generative AI, make it part of enterprise data strategy frameworks and plans. Even if the adoption would be in smaller steps and not immediate, if you decide to use generative AI, keep in mind that it should be an integral part of the data strategy.
2. Address the generative AI challenges proactively. The key challenges such as security, bias and accuracy are real and have the potential to derail analytics solutions driven by generative AI. Organizations should assess these challenges as it applies to them and proactively address them.
3. Tackle components of the analytics cycle rather than end-to-end applications. With multiple issues such as accuracy and explainability, at least initially, organizations would be better off leveraging generative AI for specific use cases within the analytics value chain.
4. Choose programs that will drive business impact. While many interesting analytics projects can be launched that can use generative AI, organizations should be very choosy in selecting those that would help drive business metrics to prevent disillusionment.
As organizations plan their data analytics strategy, they must consider the potential use case of leveraging generative AI. Depending on the analytics maturity and business priorities, organizations can decide on a road map of whether and where they would leverage generative AI.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
Read the full article here
