Our Mission

AyGLOO's main mission is to provide a competitive advantage to companies by understanding and applying AI to their business processes.

Office2-AyGLOO - Explainable AI.png

Understand algorithms is

fundamental in:

I - critical processes to be able to optimize them and take advantage of all its power and

 

II - in those processes that need to know how they work for reasons of reliability, accountability,

transparency, fairness

and ethics.

 

We have focused on developing a product to facilitate non-technical users to understand AI. 

 

We also build the algorithms with embedded interpretability.

Automate business processes by applying NLP techniques to process large volumes of data, obtaining the result in a few minutes.

 

We have a large experience with innovative projects in NLP.

 

Our differential value lies in

the proven capacity in

projects of this type, which

also allows us to have a set

of algorithms already

successfully developed.

Finally, we have a long experience

of 25 years in machine learning,

deep learning and advanced

analytics projects that we are

glad to deploy to anyone who

wants a "data-driven" company

to make better decisions and

with an impactful final result

in finance (more revenue

and/or lower costs), productivity

and brand image.

old-mac-AyGLOO - Explainable AI.jpg

Our Story

In 2018, a group of professionals with extensive experience in Machine Learning, Deep Learning and advanced analytics projects got together to carry out a highly innovative process automation project applying NLP techniques.

 

The project was a resounding success and several more NLP projects followed, although each one a greater challenge, which has given us a differential capacity to face any project in this area, no matter how innovative it may seem.

In April 2021, the same team founded AyGLOO with the intention of consolidating ourselves as a company and incorporating a much-needed Explainable AI product in the market.

 

We have developed a powerful product aimed at users without technical knowledge so that, through a dashboard, they can understand in a simple and intuitive way how complex AI models work and why the output.

 

The current result seems spectacular and very different from any other Explainable AI product on the market. Simple, powerful, reliable and configurable to suit your precise needs.

We like to present ourselves as an AI start-up with a powerful and unique Explainable AI product and with a differential proposal for Process Automation and the ability to accompany a business that wants to be “data-driven” with guarantees.

25 years of experience in Machine Learning and Deep Learning projects endorse us.

escaleras-mosaicoAyGLOO - Explainable AI.jpg

Joint experiences of the AyGLOO team

Digital Services Companies (Urbanism)

  • Source of the document. Daily bulletins of the provincial councils and the CCAA.

  • Processing: Twice a day we connect to the 62 websites, and scrape to download the day's bulletins and apply NLP techniques to extract urban planning ads and identify data in each ad. The process is parallelized in the cloud and the result is immediate.

  • Final result: segmented urban planning ads and a segmentation table containing the fields extracted from each ad.

  • Initial configurator panel for business users and modify searches by historical dates, population, etc.

  • Final result (JSON files) integrated with client systems

 

Media Company

  • Source of the document: Daily publications of the written press in pdf

  • Processing: Different newspapers are processed daily to clip each news (press clipping) identifying each part of the news (headline, subtitle, summaries, body, signature, location, caption)

  • Final result: It is presented in a validation and visualization web front-end. A user validates the result and through a neural network the system is retrained to be more and more precise

  • There is a default system configuration but the business user can modify it or launch threads adapted to the needs of each moment.

 

Media Company

  • Source of the document: Radio and television news

  • Processing: Process in near real-time similar to the previous one, making speech to text at the beginning and summarizing each news

  • Final result: It is presented in a validation and visualization web front-end. A user validates the result and through a neural network the system is retrained to be more and more precise

  • There is a default system configuration but the business user can modify it or launch threads adapted to the needs of each moment.

 

Service Companies (Consulting and legal)

  • Source Data

    • Websites of official bodies of the European Union, national, regional and local governments

    • Social Network Accounts

  • Processing: Daily download information to search, filter, do semantic analysis, select and classify information by different themes.

  • Presentation to client of the result in a web front-end for validation

 

Service company

  • Source of the document. Internet and social networks and customer CRM

  • Processing. Daily searches are launched on social networks of events that happen in the world with certain characteristics and the result is analyzed with NLP techniques, completed with information from the internet and compared and filtered with information from the CRM to finally be qualified (ranking).

  • Final result: Selection of events of interest to the client with a rating.

  • Initial Configurator Panel: There is a default configuration but the business user can launch searches adapted to the needs of each moment.

 

Insurance company

We've applied machine learning models for churn, segmentation and cross-selling, including NLP analysis and advanced analytics to customer service emails

 

Online gaming company

We have applied reinforcement learning in an online recommendation game system (PoC)

 

Facilities Company

We have developed a system to inventory objects through 360 video images by training a neural network with surprising results, superior to those achieved with commercial tools and much cheaper (PoC)