Elevate CX with AI-driven experiences
Expertflow is your connector to the latest AI technologies. We bring it all together - AI engines of your choice for real time data enrichment in self service and agent assistance, for post-call analysis, and for all other applications.
We are not limited to working with certain commercial engines and will integrate your preferred solution on request. We have many commercial solutions available such as Azure AI Speech and Language, Google Dialogflow, and Amazon Lex. We also offer Open Source based solutions such as Nvidia NeMo, Rasa, Spacy, or models found in Huggingface, covering even less commonly used languages.
Beyond AI connectors, our experts can help you select and create your AI assets and language models for your organisation, industry and language.
For most of the topics described here, we provide both cloud and on premise integrations.
Most of our AI integrations work with Cisco, Genesys and Expertflow CX.
Our AI integrations working with voice stream are based on call forking - learn more about it here.
General Info
Integration for Expertflow CX, Cisco UCCE/UCCX/PCCE. Contact us for other Contact Center integrations.
Documentation
Automated Speech Recognition (ASR) and Transcription
Automated Speech Recognition both using cloud based external services (such as Amazon Lex) and also on prem. We support many languages including less commonly spoken languages.
ASR is available for post-call analysis, such as transcription and translation, and also for real-time applications, such as the selection of self-service options by the customer's voice during the IVR.
See documentation on CX SIP Proxy and Conversational IVR for technical details.
Translation
We provide translation solutions for post-call analysis and real-time translation in transcription. We integrate with both commercial translation engines such as Google Translate, Microsoft Azure, and also with open source such as Nvidia Nemo. We support many languages including less commonly spoken languages.
Natural Language Understanding (NLU)
We provide Natural Language Understanding integrations such as conversational chat bots or question-answering using LLM. We support many languages including less commonly spoken languages.
Biometric authentication
We offer integrations for AI-based biometric authentication (voice and face baed) such as Microsoft face API, Amazon Rekognition, Google Cloud Vision API, Sestek, Nice and more.
Chatbots
Expertflow CX / Hybrid Chat contains rasa.ai out-of-box but wit its Chatbot API it allows integration with any AI. Besides integration options we also offer support in choosing the right solution for you - open source or not, LLM-based or not - depending on your needs.
We support many languages including less commonly spoken languages.
Integrations for
- Google Dialogflow
- Rasa
- Teneo
- Amazon Lex, …
- Mistral
- .. or your favorite AI solution. Please contact us for questions and to get a quote.
Knowledge Database integration with Large Language Models for Question Answering
Combine LLM (Large Language Model) engines such as LLama, Mistral or OpenAI's ChatGPT with your internal authoritative knowledge sources/ and documents to provide answers to both customers and agents. These "Retrieval-Augmented Generation" LLM solutions can be deployed on-prem in a controlled and PII-compliant manner. They can be used both for chat- and voice bots, but also knowledge base retrieval on websites or internal knowledge bases.
Information extraction (automated tagging, feelings, clustering, summaries)
We integrate AI-based tools to extract information from conversations such as for automatic generation of wrap-up codes, generating summaries, sentiment analysis or extracting frequently asked questions.
We integrate with commercial solutions such as Microsoft Cognitive Services or Amazon Lex - and also with Open Source.
Toxic language detection / filter
We provide configurable AI-based engines that allow detection and muting out of toxic language such as hate speech, insults, threat or harassment. User feedback and large Language Models (LLM) are sources that can bring along inappropriate language which must be filtered out before it can be sent to customers.