Smart Matches to Reduce Input Errors

Analysis of past usage data has revealed a pattern that majority of errors are resulting from an activity phrase which is different from what is documented and expected. This has increased as more agents are using the APIs.

We have built a fuzzy logic which has started to reduce these errors.

For example,

• the entered phrase of "ecodiesel wrangler 4dr, 4wd" will pick the expected "Wrangler 4dr EcoDiesel 4WD"
• "combustion of fuel" will pick the expected "Fuel Combustion"
• "product usage" picks up the expected "Industrial Processes And Product Use"