Swampfox is a company founded on the desire to bring exceptional Avaya Voice Portal and Call Center based products and services to market. The team, recently an Avaya engineering team based in Columbia S.C., is the group that created UCC Speech Access (now one-X® Speech) for Avaya, as well as much of the Voice Portal platform. The team led by Bob Cooper, previously the Chief Architect of Voice Portal and Avaya Intelligent Customer Routing (ICR), and Rick Ulmer, has worked one on one with developing and deploying solutions at some of Avaya's largest and most innovative customers.

Swampfox has extensive experience in VoiceXML, CCXML (outbound, complex call control, call recording, conferencing, call redirection, etc.), SIP, Video/IVVR (Voice Portal now contains an advanced video server and VoiceXML 3 capabilities) and knows the internals of Voice Portal better than any other company in the market.

Though the team knows Voice Portal extensively, it is a top-notch engineering group with a great deal of Linux and Windows internals skills, and has been building large fault resilient products for fortune 500 companies for over 15 years. Swampfox welcomes the opportunity to work with Avaya customers, Business Partners or other ISVs to help them expand their markets or just ensure that their next Avaya deployment is a successful one.

Swampfox is an Avaya Platinum Dev Connect member.

Swampfox, as a company, has extremely strong ties with Avaya. We are Platinum DevConnect member which is unusual for a newly-formed ISV. No ISV has a stronger relationship with Avaya than Swampfox.

Why is this?

Swampfox has a unique set of skills and knows Voice Portal better than any other Avaya partner because we conceived, designed and built much of it. We then worked with some of Avaya's largest customers throughout its lifecycle.

Prior to the founding of Swampfox, the team built Voice Portal's Media Processing Platform, the MPP. This is the portion of the architecture that deals with the following:

  • VoiceXML - the code is typically generated by an application server, but the MPP interprets it and requests services (prompt play, DTMF detection, speech recognition, etc.) of various MPP subsystems.
  • CCXML - every application, even VXML ones, run as CCXML applications
  • Telephony - SIP and H.323 (control and media)
  • Video - VoiceXML 3 tags, SMIL interpreter and media
  • Call Detail and Session Detail records - the MPP creates these on the fly
  • Transcription and performance trace data - including utterance capture
  • ASR and TTS logic - MRCPv1/MRCPv2 protocols and load balancing/failover logic
  • Service menu
  • UUI/UCID detection/generation

The Swampfox team envisioned, designed and implemented the MPP (every release through Voice Portal 5.0), as well as components of Intelligent Customer Routing (ICR), and a sundry of customer proof of concepts. During our tenure, the members of Swampfox had the opportunity to work directly with Avaya's largest customers, the sales team, professional services and tier 4 support. The Swampfox team:

  • Answered and prototyped unique customer requirements, such as, can I build a pure CCXML/SIP call router with Voice Portal, can Voice Portal work with SS7 and trigger delayed billing using SIP signaling, etc.
  • Created, executed, monitored, and investigated the stress tests to see how far the MPP could be pushed before it started to degrade to understand the limits of the system.
  • Triaged and explained behavior when Tier 4 would hit an elusive latency, call disconnect issue or unexpected interaction between VoiceXML based telephony and CTI.
  • Educated customers on passing call data via SIP signaling and bypass the need for CTI
  • Served as a Voice Portal expert when the sales team needed help closing a deal or answering technical questions

Does our leaving jeopardize the future of Voice Portal in any way?

Not at all. The Swampfox team worked with Avaya over the past year to transition all of our knowledge and expertise to other very capable engineers within Avaya. In fact, we delayed our departure to help ensure a smooth transition and knowledge transfer. Most of these individuals were engineers that we had worked with for many years, and Avaya has made sure that Voice Portal is staffed by some of its best and brightest.

So, if you are looking for a team that understands Voice Portal, we know it. Inside and out.

The Columbia team organically grew as a result of startup known as Conita Technologies. Conita was founded with the vision that a 'personal virtual assistant' (PVA) could be of great assistance to a user in the same way that a nurse or administrative assistant could be of great assistance to a surgeon or business executive. In both cases, the assistant's value is directly proportional to smoothness of the working relationship and in the real world, this relationship evolves and adapts over time, thus ultimately becoming more beneficial. In the same way, the PVA's role is one of convenience.

Those with the greatest need for a PVA spent time away from their desks, but still needed to have instant access to data and people, and be aware of evolving events. As such, the initial market for such a product was in the mobile productivity space.

Conita believed that the capabilities of a PVA would vary by industry and by company, and thus did not try and build a one size fits all service. Instead, a more ambitious task of building a platform (PVAServer) and service creation (PVAStudio) environment was tailored for the targeted channel. The goal was to abstract away the complexities of speech recognition and telephony and let the designer focus on user interaction and data access. VoiceXML, CCXML and SCXML were not around at the time, and, thus, the team created a language suited for complex interactions, yet allowed existing scripting languages to be used to access backend data. It allowed the PVA to interact with the user via speech, DTMF, and wireless devices. For example, when a user was in a meeting, he could set the PVA to alert him for further instructions via a handheld device instead of calling him.

To help seed the market, Conita built an application to run on the platform, PVAExchange. This not only gave us something to show what could be done, but also helped us refine what services should be provided natively by the platform versus had to be built by the application writer. PVAExchange was a product that allowed you to fully exercise Microsoft Exchange with just your voice. Naturally, you could access your inbox to:

  • read, reply, and forward messages, read attachments
  • schedule appointments, accept and reply to meeting requests
  • manage tasks

In addition to these standard functions, you could:

  • place by saying a number or saying a contact's name
  • conference calls together
  • screen callers

If someone called while you were already on the phone with someone else or otherwise engaged in a PVA session, the PVA would ask the caller for his/her name and then whisper it to you. You could then ignore it, which sent the caller to voice mail, or have the caller join your session. Though not part of the standard PVAExchange product, we also showed partners how this could be extended to carry out other actions on that same inbound call, such as directing the caller to someone else or sending him/her a specific message.

The user interface challenges were substantial, as it had to adapt to new users who knew little, as well as extremely advanced users, who knew every interaction by heart. The grammar was almost entirely flat and did not enforce a treelike mental model. The product had built in feedback mechanisms as well as scheduled newsletters that helped implicitly train the user community. These newsletters were delivered to the user based on when he/she was added to the system. At the time, Nuance recognized that it was probably the most aggressive use of their technology that they had seen. Even to this day, we have yet to see a product compare, and it has not substantially changed in over eight years.

The Voice User Interface (VUI) team had a genuine passion for the difference between public facing versus private-facing applications, the subtle use of sound and the need for adaptability. Studies were performed and a fair number of patents were submitted and granted. Even now, members of the team speak at industry tradeshows on VUI concepts and platform architecture choices. For example, see the list of Conference Sessions presented by Bob Cooper at SpeechTEK.

To get a feel for the user experience of PVAExchange, play the Avaya one-X® Speech promotional video (Dustin Donaldson is pictured here using the product). Avaya has changed the product since the Contia acquisition but the majority of the user interface is as it was.

The product was engineered to live behind the corporate firewall and be managed by the buyer's IT staff. It was designed with enterprise level scalability and failover in mind, and consisted of a management server and a collection of media servers, all acting as a unit. The architecture for Voice Portal stemmed from what we learned building PVA.

Members of Conita came from companies like NCR and Avtec. NCR, at the time, built SMP Unix servers (NCR Columbia built the OS, Compilers, firmware, hardware, etc.) for enterprises so the requirements of the typical Fortune 500 company were well understood and a requirement from the start. Avtec built real-time communications equipment for mission critical installations like railways, airlines and police departments.

The founding of Conterra/Conita

In 1996, a Columbia, S.C. based small innovative software company, Conterra Software, was acquired by the high-flying Silicon Valley VC backed company General Magic. Conterra, led by Jeff McElroy, had the concept of a 'wildfire-like' virtual assistant that General Magic believed could play a big role in their Internet like commerce service. The Columbia team built a prototype of the personal virtual assistant using technology from Nuance and Dialogic. It took a team of C++ developers and a handful of telephony/DSP cards months to build the prototype. General Magic decided to build Portico, a consumer service; however, the Columbia team had its site set on building a platform so that others, such as an IT group within a fortune 500 company, could build their own unique flavor of PVA without having to know the details of speech recognition or telephony protocols.

In 1998, with some initial seed money from General Magic, Conita Technologies was born. From there, Conita built the products mentioned above (PVAExchange, PVAServer, and PVAStudio) and received venture capital and private funding from companies like SAIC/Telcordia. Avaya became an OEM customer of Conita's and eventually purchased the company. The core engineering team that was Conita is now Swampfox.

The team has a broad range of skills, and the members have worked on many different projects throughout their career. Each member has well over 10 years experience and some have over 20 years. Unlike many teams, this group has a proven track record of developing a sustainable solution architecture, one that can be built and yet evolve to support an ever-changing marketplace, and then building it in a way that it can be successfully productized. Thus, everything the team takes on, it does so with a great deal of attention to scalability, reliability and supportability. The Voice Portal architecture and MPP platform are the team's latest examples of this, and the Swampfox team will bring this same attention to detail and level of experience to your project.

The team has extensive experience in complex user interface design, including voice, video and graphics, as well as Unix and Windows OS internals, Communication protocols, Business application integration, Real time audio and video processing, VoiceXML/CCXML, etc. However, rather than list a screen full of protocols and acronyms, please contact us. We would be happy to talk to you in detail about a project that interests you - one of yours or one of our own.

Avaya Voice Portal, in our opinion, is an exceptional platform. At its core, it is a CCXML platform, not just a VoiceXML platform. The difference is fundamental. VoiceXML was created for a very particular type of call flow. In this call flow all sessions start as a result of an inbound call, only the original caller can ever interact with the system (no way to do functional 'find me follow me' interactions), there is no way to support asynchronous events (interrupt a wait treatment to send the caller to an available agent) or do any kind of complex call or media control (take back and transfer, conferencing calling, call recording, answerless call routing, call progress/answering machine detection, etc.). If the platform supports the features, CCXML, together with VoiceXML, provide an elegant way to expose them to the application writer. Other platforms have had to resort to extending VoiceXML into a proprietary collection of platform-specific extensions.

Due to its flexibility, Voice Portal can be used to provide services that in the past had to be realized using custom dedicated solutions, if they could be realized at all. Swampfox is happy to help you with any and all of your application/deployment needs; however, here are a few projects that are of particular interest.

Swampfox is happy to work on any project that, at this stage of the company, is Avaya Voice Portal related; however, we have a keen interest in two markets in particular - healthcare and legal. If you would like to know our plans in either of these spaces, please contact us.