OMNIA architecture

OMNIA plays in the customer experience / contact center arena. Clearly, it is only a component of the whole solution that companies use to ensure the best experience to their customers and it has to play nice with many other systems: contact center suites, IVRs, CRMs, corporate directories and authentication systems. Integrations are thus an essential part of OMNIA and we have worked hard to ensure that they are easy to implement.

For starter, OMNIA comes pre-integrated with several of the most common contact centers technologies in the market: as we encountered contact center software suites from different vendors being used by our customers, we integrated with them and optimized the integration in OMNIA. It also supports the protocols to connect with IVR systems and with several CRM platforms.

Interactive Media decided long ago that it did not make sense to implement speech services or our own. Text-to-speech and speech-to-text (TTS and STT) used by OMNIA for its Virtual Agents are becoming completely commoditized, with quality rising rapidly and more and more offers on the market. So, OMNIA integrates with several of the main players, with the ability to use different services for multiple use cases. In this way, it is possible to understand the answer to an open question to the customer (eg How can I help you?) and a more specific answer (such as telephone number), because of the ability to switch to the most appropriate service task during a call. Consequently, we have a higher percentage of speech recognition and less need for referrals to human agents, resulting in an impressive ROI in operation.

In addition to the Media Server and MIND, OMNIA comes with several useful tools and modules. These are shown in the figure below as part of the overall architecture, together with the main integrations with Channels, Speech services and Contact Centers. New integrations are being added all the time.

The Media Server is OMNIA’s front-end service. It streams all media content and implements all the integrations to manage the services of third-party modules. It is also a web server and provides dynamic HTTP pages for multimodal interactions.

MIND acts as an application server to the Media Server, controlling both the call setup and the media that is played in the calls. The MIND environment is where the Virtual Agents live: the MIND AI engine evaluates the utterances from the caller (that have been transcribed into text) and decides what the Virtual Agents says next, identifies intents and provides the self-service answers.

Developers use the MIND Studio and MIND Skill modules to create and train Virtual Agents. The MIND Studio is a web GUI to create the Virtual Agents flow and manage all aspects of their deployment. The MIND Skill focuses on the Virtual Agents domain knowledge: the semantic elements that allow the AI engine to understand the users’ utterances.

OMNIA also provides a Business Intelligence module, which allows non-technical personnel to monitor the Virtual Agents KPIs and change some aspects of its service – for instance if a product has been discontinued and the Virtual Agent should change the way it talks about it. Finally, the OMNIA OAM module is for administrators to monitor and control low-level aspects of the services, receive and react to alarms and reconfigure the service if needed.

OMNIA projects: reliability, performance and high intents recognition

At Interactive Media, we have a long experience in using OMNIA (and its predecessors) to build the conversational experience that our customers want for their customers. Many factors come into play: the customer’s organization, the service they provide, the number of intents that have to be recognized, the figure of the customers calling in, and the lingo that the company uses.

We have become experts in analyzing all these factors and building the correct structures for a quick and effective implementation and we have codified the project strategy in a master plan, that we use over and over. It is not an automatic pilot for reaching perfect Virtual Agents, but it makes for a fast and predictable deployment – in a matter of weeks instead of months, and with well-defined milestones and activities.

We realize that we are not the only company offering a platform and services to implement, deploy and run Conversational AI Virtual Agents. In fact, the field is rather crowded: in the past few years many have joined us. But we believe that OMNIA gives us an advantage, in terms of ease of use, reliability, performance in recognizing the callers’ intent and in scaling up to millions of calls successfully served per month.

We can’t wait to train OMNIA to provide a delightful experience to your customers too: give us a call!