But what are the biggest obstacles for achieving productivity within Conversational AI?

According to Gartnerthe main challenges in achieving successful Conversational AI implementations revolve around: integrations, ongoing AI model training and language expansion, unrealistic expectations, and solution deployment.  

Let’s break these down into categories and explore further.  

  1. Integrations with enterprise systems and data sources are one of the most common challenges that organizations face to achieve business value.  The sophistication of solutions can vary significantly depending on the domain and the use case, and some factors add complexity to these solutions are frontend and backend integrations, omnichannel support, data analytics and BI integrations, among others.   

Backend connectors are a clear way to support easy integration that save valuable time and resources. With this in mind, Gartner recently ranked Artificial Solutions as a Tech Innovator in CAI because their “enterprise conversational AI platform, Teneo, uses Teneo connectors to support easy integration with any back-office technology and enterprise applications, including SAP, Salesforce, support systems and RPA platforms, such as Blue Prism and UiPath.  

conversational ai connectors

Gartner highlights that “this is a practical innovation that will reduce time to value around conversational agent investments for organizations. “As of today, Teneo offers over 35 pre-built frontend and backend connectors.  In addition to this, Teneo, with its Java-based platform and the integrated developers’ sandbox for advanced coding (the Integration Manager), has no restrictions whatsoever on external integrations. 

  1. Low scalability – organizations operating across multiple regions face a challenge in providing conversational experiences with new functionality, such as an added language. This means that you may need to start from scratch every time ongoing AI model training and language expansion is required. 

Going from an interaction with simple questions and answers to complex and contextual goal-driven conversations requires additional training data and dialogue design, which overlaps and supersedes the simple implementation.   

Furthermore, as per Gartner, “in organizations operating across multiple geographies, handling multilingual queries can still be a difficult issue.”  And this is precisely where Teneo excels, with proven, complex global multilingual rollouts around the world.  Teneo’s Localization capabilities (which are explored in detail in this Webinar) allow enterprises to handle multilingual implementations easily.  With Teneo it is possible to have an original implementation in one language (in any of the 86 languages currently supported by the platform) and be able to launch another solution in a different language, typically reusing about 80% of the original components, with no need to start from scratch.   

In Gartner’s own words: ”Teneo Master/Local Multilingual implementations … enable rapid rollout and scaling of multilingual solutions, as well as efficiency of maintenance and consistent quality experience for users.” 

  1. It simply takes too long. “While a case can be made for each, we contend that implementation lags are probably the biggest contributor to the paradox,” Brynjolfsson, Rock and Syverson state in their aforementioned paper.  Quick implementations for advanced VAs and CAI projects are a well-known issue.  There’s not only selecting the right expertise, training internal employees and developers, but also dealing with architectural and deployment challenges.  And yet, here we have Swisscom and Circle K overtly attesting to Teneo’s versatility, modularity and immense adaptability, supporting complex architectures, integrating with different backend systems and data sources, leveraging deep learning models for Natural Language Understanding purposes, helping customers through voice in several languages (can you imagine the challenges integrating with 5 different STT models and TTS languages?) and across different channels…  All in record-time implementations for both enterprises.  Simply put: “with Teneo, you can.” 

The AI and productivity paradox may not be solved soon, but you can choose the right tools to help Conversational AI bloom.  Our main recommendation is:  Be wise, choose wisely.  The success of your CAI project starts with selecting the right platform. You find extremely valuable insights on how to do that in this blogpost by Artificial Solutions. 

Furthermore, our dedicated forum can help you understand everything you need to know about our industry-leading CAI product, Teneo.

Learn more, here.

Carmen Del Solar is currently working as a Computational Linguist in the Information Technologies Sector. After +10 years of industry experience in Spain, she relocated to the USA in 2017. She is currently based in Chicago (Illinois), working as a Conversational AI Engineer at Artificial Solutions Inc. Carmen holds a B.S. in Linguistics and a M.A. in Applied Linguistics. She spent the first year of her Graduate studies as a visiting student at the Department of Linguistics at the University of Pennsylvania, USA. Her interests include NLP, Dialogue Systems, Human-Computer Interaction, ML and Data Science.