Dash the Bot

by Kevin Fasusi |

The world is going Bot crazy; our AI overlords are just around the corner. To ingratiate ourselves and be on the right side of history we almost avoided pressing a Bot into our service, but we couldn’t resist. Zero1, HAL, Skynet don’t judge us… C’mon everyone else is doing it.

Meet Dash (data, analysis, service, handler… just kidding it is an abbreviation of the word dashboard) our reporting Bot as of release 0.0.4. Okay, “Bot” is a bit of a stretch, it’s not a full-on Bot. Dash does not have an opinion on your colleagues, the football league or the latest Netflix Original. Dash can, however, use natural language to query the analysed data, provide insights and serve up hyperlinks as answers to inquiries. Below is a Gif of Dash in action:

Interface and Lexicon

Dash can perform many tasks competently if spoken to correctly. While more development for less constrained interaction is in progress, currently Dash understands several commands and questions about the analysis of which Dash is the custodian. For example, ask Dash:

  • “Which SKU has the highest revenue?”
  • “Which SKU generates the lowest revenue?”
  • “Which SKU has the largest excess?”
  • “Which SKU has the largest shortage?”

Hopefully, a pattern is emerging. The basic premise is that Dash can be asked to retrieve data for any keyword found in the reporting suite at SKU level and as an aggregate across the inventory profile. Other commands include:

  • “Show {replace with SKU ID}”: retrieves a hyperlink to the analysis of the SKU requested.
  • “Describe {replace with SKU ID}”: retrieves a breakdown of the SKUs main statistics (in development).

The link to the ‘Dash Lexicon’ page provides a useful guide to interacting with Dash. The aim in future releases is to make this guide redundant.

Conclusion

The reporting suite has a digital custodian called Dash. Dash is a Bot of sorts and aims to satisfy your need to query the reports quickly using natural language. Although Dash cannot yet gossip with you about the weather, the interface does provide a method of querying the analysis using natural language which is easier to remember than SQL or MDX. Obviously, the solution provided is not as robust as those domain specific languages (DSL) but serves a more constrained purpose in a simple to understand way. Dash is very much in the early stages of development. Dash doesn’t know military strategy or who Sarah Conor is. However, Dash can tell you which SKUs are overstocked; so there’s that.

Read More: