Introduction

We introduce a new colleague to our clients who is capable of handling a significant workload of administrative tasks, and in some cases, can autonomously perform entire administrative workflows.

The simplest way to deliver knowledge provided by artificial intelligence to our clients is in the form of a physical device deployed on the client's premises that requires no operation or supervision. This device can run multiple collAIgues with varying levels of expertise, and can work with multiple data sources. One of them reads documents from SharePoint, processes them, and returns them. Another one reads an IMAP mailbox and processes incoming emails, while a third one interacts with an SQL database. These collAIgues can also be integrated into workflows, allowing their specialized knowledge to complement each other for complex tasks.

With security in mind, we have encapsulated artificial intelligence in a closed box, where it performs tasks under complete control. It cannot leak confidential information, and direct communication with it is not possible. Its operations are regulated by an AutoMeta framework.

collAIgue refers to colleagues trained for administrative tasks, job types, primarily document and email processing. CollAIgue operates on a specialized, high-performance device connected to the client's network, capable of complementing or replacing the administrative work of one or more colleagues.

Sample

Alfresco CMIS integration, extracted DublinCore, Summarizable and Effectivity aspects:

Contact

If we have piqued your interest in the product, please contact us at info@collAIgue.com.

Tasks and Jobs

Administrative tasks can include:

• Creating summaries from documents or emails in the desired language.
• Determining filing metadata, such as type, sender, recipient, subject, date, identifier, etc.
• Creating short subjects and descriptions for documents.
• Further collAIgue tasks based on document type, such as:
  • defining periods of validity,
  • extracting invoice data,
  • analyzing utility bills,
  • extracting consumer locations and consumption data,
  • analyzing health and lab data,
  • sorting based on recognized metadata with automatic folder structure creation,
  • extracting personal document data (ID cards, passports, residence permits, etc.),
  • extracting vehicle identification document data
  • searching in document repositories,
  evaluating search results, and
  • generating reports.

In the application, multiple collAIgue profiles can be created for different job types. Profiles can specify the collAIgue's name, working hours, the server or application type to connect to, connection details, and required source data for processing (e.g., folder name, search expressions, etc.). The administration is no more complicated than configuring an email mailbox, and we provide online support if needed to set up the configurations.

Continuous reports (timesheets) are generated for the activities of individual collAIgues, including the timestamp, the name of the processed document, the job type, processing status, and the time spent.

Summary

• collAIgue is not a server and is not a service accessible by others.
• collAIgue does not connect to the internet, and it does not require it.
• collAIgue does not store data during its operation, ensuring the complete security of the client's data.
• Configurations created for the client and timesheets for individual collAIgue profiles are stored on removable storage.
• In the event of hardware replacement, the new device can be used immediately.
• When using collAIgue, there are no IT integration costs, since it can access the systems it needs to work with, much like a human. Using the same method, its activities can also be monitored.
• collAIgue continuously logs its activities and, if needed, provides timesheets that include completed tasks and the time spent on them.

Demo

If we have piqued the prospective client's interest, we provide them with an account on the demo server. Here, they can upload their documents to pre-configured folders and observe how collAIgue processes them. They can also test multiple collAIgues with varying knowledge and job types, allowing them to see how artificial intelligence handles their documents. Following a successful demo, a contract can be established, where the client commits to leasing collAIgue device(s) for the long term (minimum 12 months).

Device Selection

Based on the client's organization size and the amount of data to be processed, they can choose a device from various performance categories. Each device has limited performance, so if a need arises for higher speed, such as processing incoming documents within a specific time frame, the client can opt for a more powerful device or deploy multiple ones.
In addition to processing power, the context size is also crucial. In cases where the standard context is insufficient, we provide a tailored offer for a device with expanded capacity.
We do not store client-specific data on the devices; all data related to the client is stored on removable media, such as a USB drive.
In the event of a malfunction or unsatisfactory performance, the device will be replaced, and the new device can immediately take over with the data from the removable media.

Device Setup

We deliver the device to the client in a state where they only need to provide power and a network connection. The device does not require a display or keyboard connection, and there is no need for administration. The device can be turned off at any time without a shutdown procedure.
During device setup, it obtains an IP address from the client's network using DHCP.
The application running on the device can be configured by accessing it on the HTTPS port, and pre-configured collAIgue profiles can be modified or new ones added. These unique data are stored on removable media.
If the client needs assistance with profile setup, we can provide support through screen sharing.

Device Operation

Operating the device does not require specialized knowledge; it only needs a power source and an Ethernet network connection.
The client's own configuration is stored on removable media, and the client is responsible for backing it up; no other data needs to be saved.
The device does not require an internet connection, but it needs to reach the network endpoints configured in the profiles.
To access the device, only the HTTPS port needs to be enabled on the network. We provide a monitoring URL through which the device's status can be queried.