20% efficiency enabled through optimal usage of knowledge articles and curating Next Best Actions through optimal usage of knowledge articles and curating Next Best Actions

Case Study

20% efficiency enabled through optimal usage of knowledge articles and curating Next Best Actions

These global organizations sought to enhance their frontline knowledge repository to enable optimized search, update their knowledge artifacts, optimize search and improve customer resolution times.

Challenge

With legacy knowledge articles and no optimized way to update, curate or search the relevant articles, the frontline support team was not able to resolve customer queries on time and with the required accuracy. The major challenges included:

Large Inventory: 10000+ knowledge articles in the repository with multiple versioning, edits
Relevance : The Next Best Actions triggered to support agents through the knowledge articles were many times not relevant. Hence the need to optimize the knowledge repository
Duplications: A primary analysis identified many duplicates/”similar” articles being created, hence the need to purge or merge these articles

Solution

Aligned Automation leveraged its GenAI experience go through these knowledge articles and identify opportunities to optimize through:

Article Tags: Articles were tagged based on the parameters defined, these were the unique identifiers/attributes for an article
Obsolete Articles (Quick Win): Outdated articles were identified and archived through user defined parameters after ensuring replacement articles are available
Duplicate/Merging: Compare potential duplicate articles and define a ‘Similarity’ score. For highly similar articles, the sections were mapped for further review by SMEs. The duplicates were then merged based on pre-defined rules.
Knowledge Inventory: Through attribute and tags, an inventory of the knowledge articles was created as foundation for the Next Best Action recommendations and support agent review
Next Best Actions: A recommendation engine was developed to give NBA recommendations to the support team based on historical resolution data and knowledge articles

Case Study