Journey Toward an Artificial Intelligence Data Strategy – part 1
Artificial Intelligence (or AI) is gathering a lot of attention recently as it represents tremendous opportunities for automation, operational insight and market intelligence. AI, while still in its infancy, delivers in days what simply cannot be achieved by individuals in weeks, even months.
An example recently is one where Infinite engaged with a financial investment firm to profile data anomalies in its ERP system against its Salesforce CRM system in preparation for a large NetSuite ERP migration. Consider the following:
- 400 SQL tables from the source ERP system for data accuracy and cardinality;
- SQL scripts were used to surface data anomalies;
- Team included a SQL DBA and two Business Analysts;
- After roughly 3-months, an initial data profile was created.
- Deeper discovery of initial 400 SQL tables was completed by Infinite, including:
- Duplication analysis;
- Variation profile;
- Completeness analysis; and
- ERD validation.
- SQL results were cross-analyzed against highly customized Salesforce Classic database.
- Cross-table results were absorbed into Tableau for Business Unit validation.
- Final rules for data quality were presented to business stakeholders.
- Alteryx was used for its proclivity toward AI and data profiling.
- Engagement was completed in 3-weeks, using 1 resource.
Could this be considered a full-fledged “AI engagement”? Probably not, although aspects of the automated analysis were 100% machine learning based. More importantly, this represents that an AI strategy doesn’t have to begin with bots and RPA-based digital agents. The journey to a complete AI Data Strategy can begin with something as simple as data profiling.
In part 2, we will talk about spectrum and applications of AI and its importance in creating a laser-focused AI Data Strategy.