No-code Intelligent Document Processing (IDP) platform enables innovative digitalisation

Platform overview

Metamaze is capable of ingesting multiple document types, from all common sources and in all common file formats. Whether it is an Office 365 document, email, or scanned PDF, Metamaze will interpret the contents of the documents with an above-human precision and speed. The extracted values are sent back to your back-end in a structured format using API, RPA or any of our standard connector integrations.

Optional human validation can be configured depending on the accuracy level you need.

Configurable and extensible pipeline system

The brains in the background

Where Metamaze makes the biggest difference is in their AI expertise. Users only need half the time and effort to reach the same automation results as competitors in the Intelligent Document Processing (IDP) market that lack the “brains in the background” to achieve the same results.

As it stands, competitors fall short in the most important aspect: how to build custom machine learning models improved by the data of the clients. In this regard, Metamaze software is distinguished by its proprietary deep learning algorithm that requires only a few exceptional documents to be annotated to gain the most learnings possible. In essence, one only needs 50% of the documents to reach the same accuracy in Straight Through Processing (STP) as other IDP platforms. As a result, AI will free up office workers to do more value-adding activities across all departments.

Differentiators

Best in class Artificial Intelligence

Metamaze introduces unprecedented levels of intelligent document processing

Next-generation deep learning

Metamaze is built by a stellar team of AI experts that bring more than 100 years of combined experience into the field of Natural Language Processing.

This team, considered to be at the crest of European talent, is behind some of the latest breakthroughs in the field of IDP, including state-of-the-art multi-modal transformer-based deep learning algorithms enabling the Metamaze technology to process both simple as well as complex and unstructured documents.

Blending multiple domains in AI, such as computer vision, the Metamaze team keeps benchmarking and performing bleeding-edge R&D to continuously stay ahead of the competition and give our users the unique capabilities to extract, understand and process entire documents, including text, dates, currencies, tables, signatures, stamps, page numbers and much more.

Reducing the cost of annotation through active learning

Introducing Metamaze’s Optimal Document Selection Strategy (MODSS), only those documents that will improve the quality of the solution the fastest as suggested for annotation.

This significantly reduces the cost by over 50% in terms of annotation-effort, compared to other solutions, while obtaining the same level of quality.

This breakthrough in active learning has been responsible for drastically cutting short the cost and time it takes to go to production.

Intelligent data quality improvement

We use advanced machine learning techniques by continuously monitoring which documents contain annotation anomalies that risk degrading the overall quality and accuracies of your solution.

This intelligent detection mechanism feeds into Metamaze’s advanced Quality Assurance module, enabling to define custom review tasks to validate problematic entities or documents with a human in the loop.

This approach protects your solution against model drift, degraded quality assurance, and more.

Multilingual by design

Using deep learning and transfer learning, the Natural Language Processing capabilities of the Metamaze solution have been revolutionized to support multiple languages.

This cross-lingual transfer enables the capacity to learn multiple languages and generalize across languages.

In this way, the Metamaze solution reduces the cost of maintaining the quality and accuracy of supported languages by over 60% compared to others.

REMARK: At this moment Metamaze only supports Western-European languages. Other languages are planned. Don’t hesitate to reach out for more information on this.

Full end-to-end control

A unique no-code interface to control manage your intelligent document processing solution

Easy and granular configuration

Allows administrators to configure the pipeline, enable input and output security settings, add new use cases, define entities that need to be extracted, customize the accuracy thresholds for classification and extraction of entities, and modify business validation rules, create label patterns and manage user access controls.

Allows managers to upload the documents by batches, observe dashboards and human worker performance, label, and train AI models, perform quality assurance tasks.

Allows labelers and human validation operators to annotate and validate the data.

All this from within one single, user-friendly user interface.

Business-enablement for any type of use-case

Metamaze does not confine you to work with a pre-defined set of documents or use-cases. Instead, we let the business-user enjoy the unrestricted freedom of processing virtually any type of document over the Metamaze IDP platform.

We are able to do this by offering our users a unique and user-friendly user interface that doesn’t require AI or development skills to configure, release or maintain your solution.

If you don’t want to start building your solution from scratch, you can rely on a continuously expanding library of document model templates that will get you started. To even further granulate your unprecedented levels of control, you can configure all threshold settings, enabling even better control over the required quality of document classification.

Transparently manage your processing pipeline

From training your model, all the way to the actual processing of your documents in a production environment, you are able to configure and manage every step of the way.

Using Metamaze’s unique granularity and radical transparency of its model training and processing pipeline, it gives its users a closer look at what’s happening under the hood, while remaining user friendly.

While this enables business users to better engage with other departments in their organization, it ensures the user to remain fully and continuously informed about what parts of the IDP solution need attention.

Control and support any type of information

Metamaze is able to identify, interpret and extract any kind of text, image, and handwritten information. As the user is capable to define custom entities and entity classes (text or image), with any kind of parsing and validation rules, for any type of document, a full and complete set of robust entities can be tailored to any and every use-case.

Furthermore, Metamaze introduces an advanced set of entity-features to support the parsing of more complex documents.

As such, we support composite entities to interpret tabular fields, such as order lines or other tables, while enabling to aggregate entity values gives way to a novel approach for collecting one or more related entities in an intelligent manner.

Data validation and enrichment using business rules

At Metamaze we understand your information isn’t isolated, and how proper processing of information is subject to validation of a business context.

Therefore, we enable our users to define and set business rules to validate entities through an easy-to-use interface, as to enforce sets of validation rules before sending the data to external services.

But even more complex validations can be extended by adding metadata from your own IT systems and databases. This level of enrichment and validation of both data and metadata leverages the quality of the Metamaze IDP solution with other business and backend systems within your organization.

Data and model quality

Continuously improve performance and quality by managing a unique blend of artificial intelligence and human validation

Advanced data and model management

Metamaze supports your organizational context and offers each and every team member a powerful toolset to build, train and manage the data and models that support your document processing solution.

Metamaze enables its users to track, compare, monitor, and optimize model development, from training all the way into production. Hereby, Metamaze never stops providing its users with granular insights down to the level of who labeled what entity for what model, all the way up to high-level performance scores, successfully enabling to compare improvements against previous models.

This level of advanced data and model management ensures your organization is always and transparently in control over the quality and performance of your Metamaze overall solution.

Supporting fast and easy data labeling

As the quality of the model is directly linked to the quality of data labeling, Metamaze continuously aims to improve the way data labeling is done.

By using artificial intelligence, we continuously push the boundaries for optimal labeling, through means of predictive labeling and automatically detecting wrong annotations.

Being able to manage data labeling guidelines to facilitate the data entry users with step-by-step instructions, has greatly reduced the error-rate for our customers, while increasing the efficiency and effectiveness of labeling. Support for multiple monitors, as well as a set of annotation keyboard-shortcuts further increases the efficiency of labeling and data-input.

Increasing automation through human validation

Metamaze enables various levels of automated validation, such as business rules, but also enables you to set a fine-grained granularity in quality thresholds in order to support a human-in-the-loop process for manual validation.

You can configure your project so that a human must always intervene at a certain step in your process or choose to have a straight-through processing pipeline without any human validation whatsoever.

Metamaze is capable of supporting any degree of human intervention you find desirable for your project, while using the feedback from the human-in-the-loop process as input data for the deep learning models, so you don’t have to solve the same problem twice.

Using a blend of AI and humans, our customers reach unprecedented levels of quality in record times. The interface displays the confidence levels for classification and extracted entities that failed to meet the defined thresholds, highlights the entities that violate business rules or fields with incorrect/missing data, and allows business users to manage the work queue of processed documents.

Advanced quality assurance

Metamaze’s advanced Quality Assurance module defines intelligent tasks to validate problematic entities or documents where a human in the loop was used.

Using Artificial Intelligence, Metamaze assists the QA operator by automatically prioritizing what problematic documents have the biggest impact in augmenting the overall quality of the model.

QA operators can focus on increasing the quality as effectively and efficiently as possible while controlling the cost of quality assurance.

Enterprise grade

The security, stability and maturity any organization expects

Security by design, SSO & MFA

Metamaze understands how data protection and security is a vital cornerstone of your organization and has a security-by-design approach. This means that security testing is an integral part of our software development lifecycle, security requirements are constantly considered as part of our roadmap.

Using Single-Sign-On you can rest-assured Metamaze access is integrated within your federation, identification, and authentication landscape. With support for SAML SSO, Metamaze integrates seamlessly within common enterprise environments, such as Active Directory, OpenSSO, Red Hat Directory Servers, and more.

In order to further secure the access at an organizational level, Multi-Factor Authentication can be enabled, forcing users towards a more secure authentication for accessing Metamaze. Introducing these, and more, security features, Metamaze fits perfectly into your organization’s ISO27000 compliance.

Tenancy, User and Role Management

Having advanced role and user management capabilities, Metamaze provides granular access controls to its individual users and ensures that access to some of Metamaze’s features is properly allocated to the corresponding users and shields your IDP solution from accidental misuse or intentional abuse.

Whether small or large teams, Metamaze’s user and role management system supports your organization.

Queueing

Work queues provide a real-time and continuous overview of what is happening inside Metamaze while providing an organized manner to control and manage human intervention and QA tasks for large and small-sized teams.

In addition, Team Queues ensure a smooth collaboration with other team members by providing features to better organize working together. Using Metamaze’s signature easy-to-use interface, operators are able to quickly filter and handle the relevant documents and/or languages in the team queues.

Furthermore, if an operator requires assistance during labeling or human validation, documents can be parked and queued for further assistance or review from managers.

Detailed audit trail

A detailed audit trail provides end-to-end insights into the security and operational processes of Metamaze. These audit trails not only include all its automated activities of the platform, but also all human actions, such as manual validations, logons, labeling activities, and more.

This doesn’t only support security organizations, but equally provides valuable insights when interfacing with technical teams in terms of debugging.

Any activity on the platform, human or AI? Metamaze logs it.

Monitoring dashboards

Metamaze brings real-time dashboarding to inform members of your organization with all relevant metrics in terms of costs, savings, performance, and quality of your IDP solution, so your team can continuously measure and improve quality and performance.

Analytics dashboards provide your managers a view that allows tracking of various parameters such as STP rate, document-level and information-level processing, human validation operator performance, label annotating performance, data quality, numbers of errors fixed, time taken to fix the errors, etc.

Built to integrate

Interface with your existing IT landscape through easy integrations

REST APIs

Metamaze is built to be the integrated brain that augments your document processing flow with AI capabilities. As it ingests documents from a variety of sources and sends the extracted information to other IT systems, support for REST API’s and Webhooks was integral to the development of Metamaze.

Metamaze offers robust, scalable and secure APIs that drive your business.

Different API security solutions are configurable and thanks to our extensive documentation and professional support it’s easy to integrate your tools and data.

RPA Integrations

AI and RPA make a fantastic combination that continues to convince more and more customers. Therefore, Metamaze welcomes a variety of RPA integrations, such as BluePrism, MS Power Platform, and more.

Using our trusted RPA partners, we are able to integrate Metamaze’s AI into your existing RPA solution or other legacy systems.

Extending Metamaze’s capabilities using webhooks

Webhooks are user-defined callbacks that inform other applications and services about events happening on your Metamaze project in near real-time. They can be used to extend the functionality of your automated document processing workflow.

With Metamaze you can build custom plugins through its easy-to-use user interface to make customized integrations.

Using Webhooks, you can interface with your own application solutions, data validation microservices, or enrich your data from external sources in real-time.

No-code Intelligent Document Processing (IDP) platform enables innovative digitalisation

Platform overview

Configurable and extensible pipeline system

The brains in the background

Differentiators

Best in class Artificial Intelligence

Metamaze introduces unprecedented levels of intelligent document processing

Next-generation deep learning

Full end-to-end control

A unique no-code interface to control manage your intelligent document processing solution

Easy and granular configuration

Data and model quality

Continuously improve performance and quality by managing a unique blend of artificial intelligence and human validation

Advanced data and model management

Enterprise grade

The security, stability and maturity any organization expects

Security by design, SSO & MFA

Built to integrate

Interface with your existing IT landscape through easy integrations

REST APIs

Keep up to date with the latest releases

We're constantly updating our platform. Subscribe for notifications about our latest releases.

Subscribe to our newsletter

Get in touch

Ready for Intelligent Document Processing?

Ready for Intelligent Document Processing?

Ready for Intelligent Document Processing?

Let's build the next unicorn!

Let's join forces