Could AI-Boosted Workflow Help Distill Actionable ESG Insights From Dark Data?
This year alone, some 120 billion terabytes of content will be created, captured and consumed by businesses – growing by half as much again by 2025. Locked somewhere within all of that potential business intelligence is an untapped opportunity to understand how well those enterprises are performing against their social and corporate governance (ESG) targets.
Today, the vast majority of enterprise knowledge is impenetrable: as much as 80% of it exists in static documents, emails or some other unstructured form, while 54% takes the form of “dark data” –completely undiscoverable digitally.
Until now, that is. Next-generation AI tools now promise to surface and make sense of these previously impenetrable or uncombinable knowledge resources, helping employers to better target their ESG efforts.
The holy grail
To excel at ESG, employers need to be able to preempt critical scenarios such as underrepresented members of the workforce or high-potential people becoming stressed, demotivated and leaving the company. Causes might have to do with repeated customer dissatisfaction and the inability to complete tasks effectively. But those clues to employee burnout could also be down to developmental neglect or unfair treatment, pointers to which may exist within appraisal notes, calendar schedules or sick leave records.
Another ESG goal might be to discover and address environmentally inefficient use of resources – from new insights into order duplications to excessive energy consumption (hidden across purchase orders, invoices and delivery notes).
The challenge is not only to capture and structure all of this intelligence digitally and assign to it rich metadata (to aid its discovery), but also to link it in a meaningful way with associated data, and to then harness the latest AI techniques and tools to monitor, cross-analyze and distill meaningful insights from all of those inter-related knowledge assets.
AI’s expanding capacity
AI technology is very effective in pattern matching, especially now – thanks to a wide range of deep learning capabilities, from visual analysis/image recognition to natural language processing (NLP). These can help to precisely identify and capture what the content is, through a process of continuous scanning. They can also build up metadata (assigned wisdom) about the subject (e.g. an employee) through “understanding” each document.
The application of “contextual AI” adds further value, helping companies to understand how each item of content adds to the overall intelligence around a topic. This is about joining the dots between content with related metadata to capture the context of content and compare/contrast related information over time. This builds the ability to understand correlations, trends and outliers/red flags – or untapped opportunities – on demand. It is through this application of AI that a company might determine the link between a particular manager and colleagues feeling held back or underdeveloped, for example.
Then there are intelligent content assistants, boosting AI’s role in search and discovery – a kind of ChatGPT equivalent for the workplace; in other words, a bot that can query an enterprise’s metadata-enabled content to distill insights such as “Show me high-potential individuals in our employment who are not satisfied/showing signs of restlessness” or “Where are teams fielding the most complaints from customers, when are resolutions taking longest, and what is causing this?”.
Ending the chase for knowledge
Even just transforming the everyday lives of knowledge workers, who typically spend over a third of their day hunting for information to complete a task, can boost employees’ well-being by enabling them to complete their work and meet targets more effectively.
The key to maximizing AI’s potential here is to keep content infrastructures and platforms as flexible as possible so that the latest capabilities can be added on an ongoing basis.
Dr John Bates is the CEO of SER, which specializes in intelligent content automation (ICA) - the convergence of content management (capturing, storing, searching, archiving, and managing enterprise content), business process automation and AI-powered content understanding.
