Announcing AMAZON.AlphaNumeric for faster and easier skill development

On June 30, 2023, the new 3P grammar-based slot AMAZON.AlphaNumeric entered public beta. With this slot, skill developers can quickly and easily reduce the artifacts needed for alphanumerical sequences, such as tracking numbers, flight numbers, and zip codes. Developers will also save hours of work developing custom catalogs for letters and numbers.

Alphanumeric support was a highly requested built-in artifact. Skill developers were spending additional time building custom catalogs for alphanumeric codes while designing skill interaction models. For example, airline codes are several letters and numbers in sequence. To create an airline flight tracker, developers would need to build separate catalogs for each letter in addition to a number slot. By using AMAZON.AlphaNumeric, all those artifacts are reduced to just a single one. 

Developers can now gain access to a prebuilt internal alphanumeric grammar with just a few clicks of a button. The slot was written to support the widest possible range of linguistic variance, saving developers time and effort when designing skill interaction models. Use cases involving zip codes, confirmation numbers, tracking numbers, seat numbers, and other alphanumeric sequences are all covered by this slot type.

The Alexa Skills International Data (ASK ID) team worked with a group of language specialists so that the slot has consistent, comprehensive language coverage for the patterns used in different regions of the world. As a result, AMAZON.AlphaNumeric makes it easier to create a skill that can be used internationally. A developer building a skill for a shipping company, for example, will no longer need to design country-specific workarounds to support postal code patterns. Instead of creating new custom artifacts to support different locations, they can leverage the AMAZON.AlphaNumeric slot in each language.

AMAZON.AlphaNumeric creates a more consistent and reliable user experience across skills as well. Users can speak more freely outside of rigid patterns that might have been initially designed in a catalog. For instance, when presented with values like GU16 7HF, some people might say, “one six seven” when referring to the numbers, while another might say, “'one hundred sixty seven.” AMAZON.AlphaNumeric supports these variances and many more, so users are not restricted to one option. Support for this type of comprehensive language variation is nearly impossible to build through custom artifacts. Not only would it take a considerable amount of time in development, but the necessary catalogs could exceed the size limitations within a skill.

“We saw that alphanumeric patterns were difficult for developers to create well on their own,” says Andrew Jeske, manager of language research on the ASK ID team. “We knew that creating an internal grammar to provide support for developers would be the best way to resolve that issue.” Jeske is confident that the new slot will improve the Alexa skills experience for developers and users alike. 

Getting started with AMAZON.AlphaNumeric 

AMAZON.AlphaNumeric is available in all Alexa-supported languages except Arabic, which is still in development. It can be used in models today by simply accessing the slot through the developer console. Select AMAZON.AlphaNumeric from the built-in library on the “Slot Types” tab, then—with just one click—begin incorporating the slot into your interaction models. 

The AMAZON.AlphaNumeric public beta phase is scheduled to last 90 days. During this time, adoption rate and performance data will be collected. After the Alexa Skills Kit International Data team gathers and applies feedback from the previewers, AMAZON.AlphaNumeric will go into production with an official public launch. Based on developer feedback, improvements will continue to be made to the slot after launch. In addition, support for the Arabic language will be finalized and released as soon as it is available.

Developers can start using AMAZON.AlphaNumeric today to save hours of work on designing and internationalizing their interaction models.