Ender Turing Sọfitiwia Ìtúpalẹ̀ Àárín Ìpè láti ọwọ́ Lidiia Suslova

Ender Turing Sọfitiwia Ìtúpalẹ̀ Àárín Ìpè láti ọwọ́ Lidiia Suslova jẹ́ aṣẹ́gun olókìkí púpọ̀ ti 2023 Àmì-ẹ̀yẹ Idẹ A' Design Award nínú Ẹ̀ka Ìdíje Àpẹẹrẹ Ojú-ìbáṣepọ̀, Ìbáṣepọ̀ àti Ìrírí Olùlò.

IdẹOjú-ìbáṣepọ̀· 2023

Ender Turing Sọfitiwia Ìtúpalẹ̀ Àárín Ìpè

látọ̀wọ́ Lidiia Suslova

Oníbàárà: Ender Turing

Aṣẹ́gun Àmì-ẹ̀yẹ Idẹ A' Design Award nínú Ẹ̀ka Ojú-ìbáṣepọ̀

22,531

Aṣẹ́gun Àmì Ẹ̀yẹ Tí a Jẹ́rìísí

Jẹ́rìísí Ipò Àmì Ẹ̀yẹ fún #149673

Ṣàbẹ̀wò Ojú Òpó Wo ní A' Design Award Ìwé Ìròyìn Àmì Ẹ̀yẹ

Àwọ̀

Àpẹrẹ Náà

Nípa Ender Turing Sọfitiwia Ìtúpalẹ̀ Àárín Ìpè

Ender Turing jẹ́ pẹpẹ sọfitiwia tí ó dá lórí ìkùukùu (cloud) fún ìṣàkóso àbójútó iṣẹ́ àárín ìpè (call center quality management). Pátákò ìdarí (dashboard) tí ń ṣiṣẹ́ ní àkókò gidi (real-time) ń fún àwọn alábojútó àti alákòóso láàyè láti ṣàyẹ̀wò iṣẹ́ ṣíṣe pẹ̀lú ìsọdọ̀tun ààyè lọ́wọ́lọ́wọ́. Àwọn ìwọ̀n tí a ń fihàn pẹ̀lú iye àwọn ìpè, àpapọ̀ àkókò tí a fi ń dáhùn ìpè, yíyanjú ìpè lẹ́ẹ̀kan náà, àwọn àmì ìtẹ́lọ́rùn oníbàárà àti wíwà àwọn aṣojú níṣẹ́. Ìṣàyẹ̀wò-ṣáájú tí a ṣe láifọwọ́yí fún àwọn ìpè àti ìfọ̀rọ̀wérọ̀ ń dín àkókò àyẹ̀wò dídára kù, tí àwọn àlẹmọ́ tí ó ṣeé yí padà sì ń jẹ́ kí a yan àpẹẹrẹ kékeré dípò gbogbo ìfọ̀rọ̀wérọ̀. A lè ṣẹ̀dá tàbí ṣàtúnṣe àwọn káàdì ìdíyelé (scorecards) láìní ìrànlọ́wọ́ àwọn onímọ̀-ẹ̀rọ. A ń pèsè ojú-ìwé ìbáṣepọ̀ náà ní èdè tí ó lé ní mẹ́wàá.

Ìdánimọ̀ Náà

Nípa Àmì-ẹ̀yẹ Idẹ A' Design Award

Ender Turing Sọfitiwia Ìtúpalẹ̀ Àárín Ìpè jẹ́ olùborí Àmì-ẹ̀yẹ Idẹ A' Design Award, ìdánimọ̀ olókìkí gan-an tí a fún àwọn àwòṣe tó tayọ tí wọ́n yàtọ̀ nípasẹ̀ ìṣẹ̀dá, àrékérekè àti ọgbọ́n inú. Àwọn olùborí Idẹ jẹ́ àwọn ohun èlò tó ní ìjáfáfá gan-an tí wọ́n sì ní ìṣẹ̀dá tó yani lẹ́nu tí wọ́n ń ṣe ẹ̀rí ìrírí gbígbòòrò, ìrònú àti ọgbọ́n àwọn olùṣàwòṣe wọn. A retí pé Ender Turing Sọfitiwia Ìtúpalẹ̀ Àárín Ìpè yóò jẹ́ ohun tí a kà sí gidigidi, tí a gbóríyìn fún tí a sì nífẹ̀ẹ́, tó jẹ́ ti ọ̀jáfáfá lọ́nà tó ṣe pàtàkì tó sì ní ìmúṣẹ̀dá tuntun. Àwòṣe tó borí Àmì-ẹ̀yẹ Idẹ A' Design Award yìí ń kó àwọn ọ̀nà tó dára jùlọ nínú iṣẹ́ ọnà, sáyẹ́ǹsì, àwòṣe àti ìmọ̀-ẹ̀rọ jọ, ó sì tayọ nítorí àfiyèsí rẹ̀ sí àwọn àlàyé kékeré. Ender Turing Sọfitiwia Ìtúpalẹ̀ Àárín Ìpè ń tẹ́ àwọn ohun tí a nílò lọ́rùn lọ́nà àgbàyanu, ó ń fọwọ́ kan àwọn ìfẹ́ inú, ó ń gbà ọkàn ó sì ń fa ìfẹ́ láti mọ̀ mọ́ra.

Lẹ́yìn Àpẹrẹ Náà

Àkójọ Ìròyìn

Àwọn ẹ̀tọ́, àwọn ọmọ ẹgbẹ́ àti àlàyé ìmọ̀ ẹ̀rọ tí olùṣàpẹrẹ pèsè.

Àwọn Ohun Àkànṣe: A real-time dashboard allows supervisors and managers to monitor call center KPIs and agent progress with live updates. It displays metrics such as call volume, average handle time, first call resolution, customer satisfaction scores, and agent availability. This helps identify areas of improvement, detect issues, and optimize call center performance through proactive management and data-driven decision-making, leading to improved efficiency, customer satisfaction, and business outcomes.
Ìwọ̀n àti Àpò: The primary function of Ender Turing engine is to help users create and train AI models that can understand and generate human-like language. The platform provides a range of NLP capabilities, including text classification, sentiment analysis, entity recognition, language translation, and text summarization, among others. Users can build custom models using Ender Turing's drag-and-drop interface or integrate pre-trained models into their applications using the platform's API.
Àwọn Àkọlé: Artificial intelligence (AI), Natural language processing (NLP), Machine learning, Text classification, Sentiment analysis, Data processing, Text summarization API, Cloud-based platform, Language modeling, Text mining, Call center, Scoring
Ìmọ̀-ẹ̀rọ Ìṣelọ́pọ̀: We used a data-driven design approach to gather feedback for our design, which involved observing and interviewing customers of the product. All of our decisions were based on identifying user pain points and areas of frustration. As a result, we developed a product solution that eliminates most pain points for call center supervisors, increases their productivity, reduces stress and frustration, and features an intuitive UX interface covered with more than 10 language options.
Ìpèníjà Àpèsè: Experiments were done to test hypotheses and validate design decisions. We found that our users are non-technical and prefer a more prominent communication style. We avoid using complex logic constructions and heavy interface decisions. Our UX solutions are continuously simplified.
Ìwádìí: As we maintain ongoing communication with our end users, including data center supervisors and stakeholders, we have collected data through a range of research methods, such as interviews, surveys, and observations.

This has allowed us to gain valuable insights into usesr needs, preferences, and behaviors, which will inform the design of effective solutions that meet their specific requirements.
Ìmísí: Call centers handle up to 1000 calls per agent per day. Measuring and analyzing this data with metrics similar to those used by humans is crucial to improve client service and call center productivity. A productive call center saves businesses revenue and promotes polite and friendly communication with customers, enhancing a business's reputation, improving customer satisfaction, and increasing their overall quality of life.
Àkókò Iṣẹ́ Àkànṣe: The project started in June 2020 in Tallinn, Estonia. The Design for the Project started at January 2022 and still developing and enhancing. Product was nominated as winner at Latitude59 pitching competition in May 2022, Tallinn, Estonia and took 2nd place at Infoshare Startup Contest where been selected among 500 startups from all over the world in October 2022, Gdansk, Poland.
Ìlànà Ìṣiṣẹ́: Our solution provides a 20-fold increase in speed for scoring calls and chats. This is made possible by automated AI pre-scoring and call center quality assurance. Using flexible filters, you can randomly select only a small number of calls and chats for quality monitoring, rather than monitoring all of them.

Moreover, our system allows for flexible setup of scorecards with no limitations. This means that you will no longer need engineers or Excel sheets to modify or create new scorecards.
Àwọn Ìwé-ẹ̀rí Ohun-Ìmọ̀: Scientific publications:

1. Transferability Evaluation of Speech Emotion Recognition Between Different Languages
January 2022
DOI:10.1007/978-3-031-04812-8_35
In book: Advances in Computer Science for Engineering and Education (pp.413-426)

Ievgen Iosifov
Olena Iosifova
Oleh Romanovskyi
Vladimir Sokolov

2. Automated Pipeline for Training Dataset Creation from Unlabeled Audios for Automatic Speech Recognition

July 2021
DOI:10.1007/978-3-030-80472-5_3
In book: Advances in Computer Science for Engineering and Education IV (pp.25-36)

Oleh Romanovskyi
Ievgen Iosifov
Olena Iosifova
Vladimir Sokolov

3. Analysis of Automatic Speech Recognition Methods

January 2021
Conference: Cybersecurity Providing in Information and Telecommunication SystemsAt: Kyiv, Ukraine

Olena Iosifova
Ievgen Iosifov
Vladimir Sokolov
Oleh Romanovskyi
Igor Sukaylo

4. Sentence Segmentation from Unformatted Text using Language Modeling and Sequence Labeling Approaches

October 2020
DOI:10.1109/PICST51311.2020.9468084
Conference: 2020 IEEE International Conference on Problems of Infocommunications. Science and Technology (PIC S&T)

Ievgen Iosifov
Olena Iosifova
Vladimir Sokolov

5. Methods and components of natural language processing

August 2020 Adaptive automatic control systems 1(36):93-113
DOI:10.20535/1560-8956.36.2020.209780

Olena Iosifova
Ievgen Iosifov
Oleksandr Rolik

6. Techniques Comparison for Natural Language Processing

June 2020
DOI:10.5281/zenodo.3895814
Conference: 2nd International Workshop on Modern Machine Learning Technologies and Data ScienceAt: Lviv-Shatsk, Ukraine

Olena Iosifova
Ievgen Iosifov
Oleksandr Rolik
Vladimir Sokolov