Words - emotions - machines
what is TSA?
Techmo Speech Automation is a toolkit for creating complex voice interfaces. TSA makes the creation process faster and more efficient, adjusts the models to the speaker’s needs and improves both the quality and speed of data processing.
Where to use TSA?
The software is intended to be used wherever verbal communication is possible. Examples include a contact center or an industrial environment, as well as applications such as voice assistants or contactless machine operation.
We created Techmo Speech Automation based on our previous experience with voice technology implementations and seasoned it with a couple of revolutionary ideas. We believe that this will lead to a new standard of quality in voice interface implementations – which makes us happy, because we enjoy working in the field and want it to evolve continuously.
Szymon Pałka
Chief Scientific Officer
Features
data
extraction
extraction
Techmo Speech Automation enables statistical analysis of large speech-related datasets to examine the usage of a given vocabulary or to group utterances thematically.
specialized
models
models
TSA offers a wide host of implementation possibilities and facilitates quick creation of voice assistants. We work closely with our clients to produce a model tailored to the given domain, which ensures proper recognition of specialized terminology.
Diarization
Speaker identification, or diarization, is the process of distinguishing individual voices from a group of people talking simultaneously or from people talking in the background. Techmo Speech Automation offers speaker identification and can split the recording into utterances belonging to different persons.
Emotion
detection
detection
Techmo Speech Automation recognizes emotions present in the speaker’s voice. As a result, users get live information about how the conversation is going, which increases the percentage of successful interactions.
Sentiment
tracking
tracking
TSA tracks the speaker’s sentiment to identify their attitude. Results of this analysis can be used to improve conversation quality, plan marketing strategies, research customer expectations or monitor the brand’s public image.
Transcription
TSA provides software for creating and analyzing transcripts of audio coming from phone calls, meeting recordings, dictation and many more.
Contextual
analysis
analysis
TSA provides enhanced contextual analysis for AI systems (e.g. chatbots or virtual assistants) beyond text transcription. It offers semantic processing of the full recognition lattice, which allows the software to better understand the utterance and, as a result, improve service quality.
Noise
robustness
robustness
Techmo Speech Automation provides a speech synthesis system that can work in various environments. It is robust against noise, allowing the user to understand the utterance regardless of surrounding acoustic conditions.
on-premises
installation
installation
TSA modules can be installed on-premises, which means they don’t need Internet access and operate entirely in the local IT infrastructure. This gives our customers full control over their data.
Emotional
synthesis
synthesis
TSA provides next generation speech synthesis tools – naturally sounding voices that can express various emotions and be adapted to the speaker’s mood on the fly.
voice
branding
branding
TSA offers customers the option of implementing a unique voice, which helps with building brand identity and ensures consistent verbal communication with the customer’s own clients.
Language
dynamics
dynamics
TSA quickly reacts to changes occurring in language and incorporates them into the model for proper speech recognition. Such changes may include appearance of new vocabulary, adding new meanings to words already in use, occasional insertion of phrases from other languages, using slang, colloquial language or abbreviations and changes in grammar.
BENEFITS
DYNAMIC ADAPTATION TO CHANGES IN LANGUAGE
REDUCED NEED FOR TRAINING DATA
SHORTER RECORDING ANNOTATION PROCESS
HIGHER QUALITY OF SPEECH SYNTHESIS
FASTER BUILDING OF SPECIALIZED MODELS
EFFICIENT MECHANISMS FOR SPEECH ANALYSIS
REACTING TO THE interlocutor’s emotions
LOWER COST OF CREATING VOICE INTERFACES
Techmo Speech Automation was created as a response to a dire market need. It can speed up and improve a variety of complicated processes. This is mainly a gift to our Partners and Customers. It will provide them with the kind of quality that is a game changer for the voice interface market.
Piotr Stankiewicz
Head of Business
BUSINESS
PROFITS
TSA is a modular system that facilitates the automation of speech processing and analysis. It lowers the cost and shortens the time needed for creating specialized models. It automates the creation of speech recognition modules and streamlines the production of voice assistants by employing semantic analysis tailored to the nature of the task. TSA can also be used to conduct sentiment analysis and emotional speech synthesis.
BUSINESS BENEFITS
Implementing the TSA software reduces the cost of producing intelligent voice interfaces, which in turn makes them more accessible to a wider consumer base. Creating more intelligent models also stimulates market demand for such products. Moreover, TSA significantly improves the quality of voice assistants – more data ensures better understanding of the speaker’s intentions.
USERS
TSA is mainly aimed at integrators and developers of software that incorporates voice interfaces. It will enable them to create more advanced solutions by implementing emotion recognition, sentiment tracking, diarization, NLP analysis of obtained transcripts, chatbots, knowledge bases and many more.
COLLABORATION
We provide technology for building voice assistants that fulfills all business, marketing and information requirements. Our software simplifies the multi-aspect analysis of voice recordings and the process of transcription. We also offer quality control and evaluation of voice assistants. Our Partners can benefit from additional support.
Techmo provides business solutions in the fields of speech recognition, speech synthesis and natural language processing. Since 2013, the company has been conducting research and development activities based on the latest IT achievements and, as a result, co-creating technologies of the future. Techmo offers software for automation and optimization processes, such as:
SPEECH SYNTHESIS
SPEECH RECOGNITION
Our customers include major telecommunication providers, banks, insurance companies and leaders in the utility market.
CONTACT
TECHMO Sp. z o.o.
ul. Torfowa 1/5, 30-384 Kraków
KRS 0000455399;
NIP 6762464155;
REGON 12282371500000;
Share capital 336 500,00 zł
Sales Team
Michał Wątor
mob.: +48 737 189 747
e-mail: michal.wator@techmo.pl
Media Relations
Magdalena Foltyn
mob.: +48 727 705 373
e-mail: magdalena.foltyn@techmo.pl
POIR.01.01.01-00-0884/19
Subsidies: 7 840 584.58 PLN (budget: 10 222 194.26 PLN)
The project is partly funded by the National Centre for Research and Development under the POIR program. The project is partly funded by the European Union through the European Regional Development Fund under the Smart Growth Operation Programme 2014-2020. The project is being implemented as part of the National Centre for Research and Development: Fast Track competition.