Network Speech Solutions; the art of empowering a telephone network with embedded and integrated Advanced Speech Recognition (ASR), sometimes referred to as "Automated Speech Recognition".  In network speech solutions technology is applied directly in the middle of the core carrier network as opposed to on the edge, or even in more cases on the overlay network.

Instead of the speech technology remaining isolated within the control and domain of a single and very specific Interactive Voice Response (IVR) platform it is deployed as a common, or shared-asset across the entire operating network, enabling not only enhanced revenue generating applications directly as the dial-tone for a specific product but as an asset which can be shared by the entire carriers operating mandate.

The same shared asset may be enjoyed by Customer Care, Operator Services, Enhanced or Premium Rate departments, Enterprise or B2B departments and the core carrier network itself.  It may be used to satisfy rudimentary and rather basic customer interactions, for example, to play a prompt notifying the caller of the status of a particular transaction, or can be used to provide very rich and powerful Natural Language Speech Recognition (NLSR) to all facets of the consumers experience of the network.

Instead of "Press one for top-up or Press two for Billing or Accounts" the callers can enjoy a single more human intuitive greeting along the lines of "Welcome to TelcoOne, what would you like to do?" This more natural user-interface lends itself particularly to increased user adoption rates, improves containment within the self-service application and makes speech self-service platform elements like "Voice Portals" that have many branches of functionality extremely easy to access and use.

A common limitation of DTMF based user menu's is the breadth and depth that can be employed at any given time. The traditional 'rule-of-thumb' is no more than 3 options wide, by 3 options deep. Any more than this quickly results in user confusion, or simply forgetting what options to use to get to the target application quickly (so called dial-ahead behaviour).  Flexible yet flat Network Speech Solutions allow the caller to effectively say anything they want to get to the desired application by keyword. It is Anana responsibility to ensure that all potential grammars are encapsulated at run-time to cover all options available in the portal options tree.

Additional benefits of a correctly deployed Network Speech Solution is the ability to share the pool of Speech Resources as a common asset across the Telecoms Providers range of business and products.  A single service node of network speech can be freely used by many departments concurrently, levering the asset base to its fullest, whilst reducing the total number of ports or instances needed to provide resilience and service levels.  Typically speech solutions are scaled on an application by application basis, and run within their own domain, making it very difficult to share the assets meaningfully.

A well designed and architected network speech solution, using standards based interfaces for VoiceXML, Media Resource Control Protocol (MRCP) and the flexibility of web based applications servers will and does significantly reduce op-ex, cap-ex and offer the maximum capability to 'sweat an asset' across all lines of business.