The RITA Project
(B2C / Demo application for Bay Area Rapid Transit (BART))
RITA — Real-Time Information for Transit Agencies — was conceived and delivered as a demo application for the design technologies and services of Interactive Intelligence (later acquired by Genesys). It became the company's most successful demo platform — the only one to draw productization interest from multiple prospects, leading to active negotiations with the BART system itself before the Genesys acquisition shut the project down.
I originated the concept, defined the scope, secured funding, designed the entire interaction, and oversaw all subsequent versions. While not a one-person production, RITA was very much one person's vision — and that vision resonated widely enough to attract substantial commercial interest.
What they asked for
Interactive Intelligence needed a demo application that could showcase three capabilities at once: sophisticated conversational design, a lightweight and platform-portable voice architecture, and robust real-time API integration. The brief was unusual — they wanted something that would function not just as sales theater, but as a genuinely useful real-world application that prospects could pick up and use themselves.
Living in the Bay Area at the time, I had become acutely aware that no single transit application offered hands-free access to real-time information across the region's many agencies. Every existing app required touch input, relied on published schedules rather than live data, and confined itself to a single agency. The opportunity was clear: design a voice-forward system that could pull live data from multiple transit agencies and deliver it through a single hands-free conversation.
What we envisioned
RITA was designed as a voice-forward application — appropriate to the BART rider's typical context (mobile, hands occupied, often in motion). But it was architected from the start for synchronized hybrid modalities: voice as the primary input and output, with visual and aural content synchronized for the planned GPS station-locator features. RITA's API calls reached not only into the BART system itself, but into the operating system of the user's phone, accessing real-time GPS data and rendering directions in the native maps application.
The technical foundation was as significant as the user-facing design. RITA was built on a lightweight, declarative voice markup standard that is completely platform-agnostic and fully portable — a standards-based approach that allows the same application to run across any compatible platform without rewriting. That kind of portability is an elusive but important goal, and RITA's forward-looking and standards-based foundation produced a noteworthy solution — a system that could be pitched to clients across platform boundaries, dramatically broadening Interactive Intelligence's potential client base and turning architectural elegance into commercial reach.
What we built
The team felt it was important to always focus on how the application would be used, not just by the end-user callers but by the salespeople using it for demo purposes, as well as by the transit agencies we hoped to sell to, and the broader-based prospective client base who could see for themselves how powerful technology could be put to use to make their own customers' lives easier.
RITA's features were design and developed in a phased approach that gradually added not just more features, but more complex and proactive AI-powered features, as follows:
Schedules and "Next Trains" departure information
This would also provide good showcasing for our vXML rapid coding and deployment, as well as the robust API calling features, and real-time information updates
Fares (but not yet with simulated "Add Value" fare card functionality)
Transfers
This also enabled us to show RITA's ability to adapt according to time of day. The BART system's routes change, depending on the day of the week and the time of day. RITA took this all in stride, and with Schedules and Fares already firmly in place, these on-the-fly adaptations and modifications to routes could be seamlessly accommodated.
"Easter Eggs"and directions to destinations not directly served by BART
Station information, including nearby points of interest
Administrative (non-travel-related) information such as Lost & Found or BART Police
GPS "closest station" information and other feature enhancements
For each phase, both the design documentation (including the call flow diagrams, branding and persona definitions, and sample dialogues) and the corresponding ancillary documentation (including a brochure and a handbook for the sales team, ) was updated to include all the information salespeople would need to know, to demo the new functionality, as well as fun "Try it for yourself!" usage scenarios for users to walk through -- without even needing to know anything at all about the San Francisco Bay Area!
What it achieved
RITA generated an estimated $1.2 million in new business for Interactive Intelligence as a demo platform, but its impact extended well beyond sales theater. The Bay Area Rapid Transit agency itself opened productization negotiations to deploy RITA commercially, and additional transit agencies in Boston, Atlanta, Houston, and Miami entered preliminary discussions. The Genesys acquisition closed the project before those discussions reached contract.
RITA remains a flagship example of what voice-forward conversational design can accomplish when paired with lightweight, standards-based architecture, robust real-time data integration, and a coherent vision of who the actual user is. The portability that gave RITA its commercial reach was a function of architectural choices the field made deliberately — choices about standards-based design, vendor independence, and the freedom to migrate platforms as needs evolved. The principles haven't lost their value, and one of the most interesting questions in the field today is how they'll take new shape within emerging AI technologies..