I have designed, branded, developed & optimised Conversational Speech IVRs, Voicebots & Chatbots for Contact Center and mobile self-service applications working with multinational Tech giants (APPLE, SAMSUNG), Telecoms (VODAFONE, SKY, TALKTALK), Management Consultancies & Investment Banks, Insurance & Security companies (ALLSTATE, ADT, AVIVA), Banks (BARCLAYS, EMIRATES NBD), Retail, Utilities (TECO, PEOPLE'S GAS, CENTRICA), Logistics (DPD) & Governments (UK NHS & DWP, the European Commission, the Greek and Belgian Governments), as well as voice platform providers, CRM and Contact Centre technology providers and integrators.
Below are examples of my Success Stories:
Client: Fortune 100 Insurance company
Client had invested heavily in several voice AI projects and needed an experienced VUI Designer to design and oversee the implementation and testing of these speech IVRs. They also required a health-check on their on-hold strategy. The challenge was that their customers were abandoning the existing speech applications and often giving up even while waiting to speak to a Contact Centre advisor.
Maria carried out a diligent investigation of all existing VUI and application designs and also did extensive multi-channel user research to identify the core pain points for her client's end users, both while interacting with the voicebot and while in queue to an agent.
Maria delivered new VUI designs, made adjustments to any existing ones and also produced a detailed list of recommendations regarding the Contact Centre on-hold strategy and implementation. As a result, customer containment in the Speech IVR increased considerably with fewer requests to speak to a human advisor, but also abandonment in queue to an advisor dropped dramatically.
System Integrator's comments:
"With ridiculously deep subject-matter expertise in multiple facets of creating conversational speech applications, Maria brought an outstanding amount of professionalism, while delivering world-class design and strategy recommendations. She produced outstanding results - and impressed the heck out of our customer. It was a delightful respite to know that we were indeed in good hands".
Client: A multinational (hardware and software) tech company
The client reached out to me to monitor their existing Contact Centre automation Speech IVR and identify any issues with the Speech Recognition (ASR) and the Voice and Conversation Design. They knew that both ASR accuracy and concept accuracy (user intent) were relatively low (60-70%) and that many of their end users / customers were abandoning the system and opting to speak to an agent, usually already very frustrated in the IVR.
The chosen appoach was to listen weekly to thousands of live calls to the automated system, the Contact Centre agents and the agent queues in-between. Maria, thereby, identified certain problem system messages that were misleading customers to say unexpected things and subsequently tripping the ASR and generating error prompts (No-Matches and misrecognitions). This was causing customers a lot of confusion, frustration and anger towards the brand, which, in turn, meant that half of them would land in an agent queue for a long time, aggravating the problem and increasing the workload to the Contact Centre agents. Maria also discovered that several agents didn't know much about the Speech IVR and its functionality nor any of the information that customers had already successfully provided to the system. To remedy this, Maria carefully and purposefully reworded these system messages and recommended specific changes to the existing VUI Designs, mainly the order in which the system was asking customers questions. Finally, she recommended company-wide training on the Speech IVR for Contact Centre staff, so that the two CRM channels were tightly integrated and coordinated.
The recommended changes to the wording of the problem system messages, as well as the optimisation of the call flow, meant that it was clearer to the users exactly what was being asked of them. This significantly improved speech recognition (ASR) accuracy and as a result task completion (i.e. the customers were getting things done within the IVR). This, in turn, meant that the customers were staying longer and engaging more with the Speech IVR (higher containment), but also that fewer users had to resort to a human Call Centre agent (lower agent costs and attrition).
Maria's contact and process flow designs, dialogue scripts and Speech Analytics-based tuning recommendations contributed to "us saving up to $10 million annually".