Skip to main content
Version: 10.3.2

View automation technology classification

The following levels of robotic labor classification are measured by the complexity and stability of the underlying technology.

  • Stability of Technology: Level 1 is the most stable, and Level 5 is the least stable.
  • Execution Speed: Level 1 is the fastest way to execute automation, and Level 5 is the slowest.
  • Resource Utilization: Level 1 is the most efficient in utilizing computing power, while Level 5 causes the highest hardware overhead.

The above complexity, stability, and resource utilization comparison is only applicable to RPA automation levels (Level 1 through 5).

Level 1: API Automation (APIs, File Exchange, Database, ETL)​

Examples​

  • Collect information from Social Network API(s), web services, SOA architectures, or legacy APIs:
    • REST, SOAP, SNMP
  • Automate file exchange:
    • FTP, SFTP, S3, SMB
  • Connect to a database and build a report:
    • JDBC, ODBC
  • Extract, Transform, Load operations and document conversion:
    • Read, write, convert, process PDF, XML, DOC, DOCX, XLS, XLSX, HTML, TXT, images, and so on
  • Send and retrieve emails programmatically:
    • Microsoft Exchange, SMTP, IMAP, POP3
  • Place and retrieve a message to or from a queue:
    • AMQP, STOMP, JMS, ActiveMQ, RabbitMQ, and so on

Traditional solution​

  • ESB tools, ETL tools, protocol-specific libraries, custom scripts, and code

WorkFusion​

  • Pre-built connectors to most or all of the common user interfaces and APIs
  • Out-of-the-box plugins to handle file operations and transition, for example, file, S3, zip, HTTP, SOAP, REST, Excel, JSON, and so on

Pros​

The pros of the approach are as follows:

  • It is a native robot language, meaning the most stable, robust, and effective integration and function.
  • There is no need for separate hardware resources to run Agents with the third-party software installed.
  • You can use any operating system like Linux, Windows, and so on.
  • APIs have strictly defined and often standardized protocols. It also means APIs do not tend to change. When APIs change, they preserve backward compatibility for an extended amount of time.
  • Most of the protocols have connectors to a higher level for humans. For example, you can open an XLS file with Excel.

Cons​

The main disadvantage is that IT is involved in integrating systems on this level or developing interfaces for human interactions.

Level 2: Web scraping and crawling​

Examples​

  • Fetch public data from static websites:
    • HTTP, HTML
  • Use web search crawlers and indexers, full-text search:
    • Solr, HTTP

Traditional solution​

  • Transport via HTTP, web scraping, and crawling tools

WorkFusion​

  • Pre-built set of plugins to handle HTTP data exchange and results, for example, HTTP, HTTP-extended, XPath, CSS selectors, and so on
  • WorkFusion crawling infrastructure

Pros​

  • Refer to pros from Level 1.

Cons​

  • Sites tend to change. Automation scripts require re-writing, which is not applicable for crawling.

Level 3: Web Automation (dynamic web applications and sites)​

Examples​

  • Fetch data from dynamic websites:
    • HTTP, HTML, AJAX, Javascript
  • Take and compare screenshots to monitor websites.

Traditional solution​

  • Browser (headless or actual)

WorkFusion​

  • Headless browsing:
    • PhantomJS
    • Headless Chrome
    • Headless Firefox
  • Using a real browser run off the desktop:
    • Chrome
    • Firefox

Pros​

  • Hardware utilization can run multiple instances of automation in parallel on a single machine or in a single session.

Cons​

  • Prone to breaks due to the increased number of variables.
  • Requires a virtual desktop. *

* No hardware is required for headless browsing.

Level 4: Desktop automation (legacy applications and Internet Explorer applications)​

Examples​

  • Windows applications, .Net applications
  • Web applications that only work with Internet Explorer
  • Legacy applications
  • Citrix applications *
  • Java applets
  • Terminal and command-line applications
  • Mainframe
  • Microsoft Office applications

* When Citrix allows access to clipboard and Windows properties.

Traditional solution​

  • Give up
  • Instrumentation APIs:
    • Java Access Bridge, Win32, Microsoft Active Accessibility, Microsoft UI Automation, and so on
  • RPA software solution:
    • AutoIt, AA, BP, UIPath, and so on

WorkFusion​

Pros​

  • It mimics human interaction.
  • No extra IT support is required.

Cons​

  • Requires Windows environment for RPA Units.
  • Single robot per user or session at a time.

Level 5: Surface Automation (screen-only tools, image recognition)​

Examples​

  • Websites and systems that try to block or prevent bots from access
  • Legacy applications that cannot be inspected
  • Citrix applications *
  • Exotic Operating Systems
  • Flash applications

* Requires remote access to an uncontrolled environment.

Traditional solution​

  • Give up
  • Image-based robotics
  • Screen scraping

WorkFusion​

Pros​

  • Mimics human interaction.
  • No extra IT support is required.
  • End-to-end automation for uncontrolled legacy environments.
  • Flexibility, meaning the same approach is applicable to any applications.

Cons​

  • More complex for headless unattended execution:
    • Dependency on screen size, resolution, fonts

Level 6: Cognitive Automation (unstructured data)​

Examples​

  • Unstructured documents:
    • PDF, DOC, Email, TXT, HTML, and so on
  • Computer vision (OCR, image recognition, and so on):
    • Scans, photos
  • Captcha, re-captcha

Traditional solution​

  • Give up
  • IT project(s)

WorkFusion​

  • AutoML:
    • Data Extraction
    • Classification
    • Ranking and Scoring
    • Computer Vision
    • OCR

Pros​

  • Automated
  • Continuous model re-training
  • Built-in quality controls

Level 7: Conversational Automation​

Examples​

  • Conversation understanding:
    • Chat and voice
  • Fulfilling back-office requests
  • Question-Answer systems

Traditional solution​

  • Process manually

WorkFusion​

  • Chatbots

Pros​

  • Direct integration with RPA and Cognitive Automation Levels (1-6):
    • Chatbots collect data, enter information, process transactions, and deliver results directly in a chat window, via a text or email.
  • Chat UI integrations, including:
    • Website
    • Mobile
    • Third-party (Slack, on-premise messaging systems, and so on)

Appendix​

What else?

Level H: Physical​

Examples​

  • Manufacture lines
  • Press physical buttons on a hardware token
  • Paper document as an input (Canon or Epson)

WorkFusion​

  • Human-in-the-loop (HITL)

Level Z: Alien intelligence​

Some other examples​

  • Golden Robot spans along levels 2-6 (with HITL).
  • Humans operate on levels 6, 7, and H.
  • Business Process Management (BPM) is used to connect levels.

Dynamic workload balancing (server-based orchestration)​

With dynamic workload balancing, the RPA platform can distribute the workload among RPA Units in real time. It improves productivity as work is optimally distributed among available RPA Units, thus enabling more autonomous functioning of automations requiring less manual support to manage them.

Enterprise-level architecture​

  • Disaster Recovery, High Availability, Scaling, Monitoring, Docker, Security (+SSO, +PAM)
  • Packaging (artifactory, continous delivery and integration, configuration management, release management, unit testing of Business Processes)
  • Cloud-based or on-premise or hybrid deployment

Language-agnostic ML​

  • Non-technical
  • Per process license (no volume limit or on amount of hardware utilized)
  • Included cloud-based hosting
  • 24/7 support
  • Industry-adapted tools: Eclipse, Selenium, AutoIt
  • Dev community
  • Knowledge base
  • Online and offline training