Cognizant Showcases AI-Powered Dynamic Workflow Solution

By leveraging Unqork and Google Cloud’s Vertex AI platform, the Cognizant team quickly built a solution that dynamically transforms natural-language prompts into end-to-end workflows, including the design and user experience–all with zero code.

At At AI Unlocked: A Google Cloud x Unqork Innovation Day, Cognizant showcased  Orchestrate.AI: An AI-powered Dynamic Workflow Solution in order to demonstrate what you can achieve by leveraging Unqork’s enterprise application platform and Google Cloud’s AI solutions. Their team was able to build their use case quickly and securely, while unlocking the enormous power of AI-enabled experiences—all with zero code. 

As one of world’s leading professional services companies, Unqork’s partner Cognizant leverages its expertise in AI and other technologies to transform experiences, reimagine processes, and modernize technologies. Cognizant’s Jerome Mathias, Global Competency Leader for AI Solutions, Low-Code Platforms, and Process Orchestration, showcased the AI-powered dynamic workflow solution, which:

  • Leverages AI  to augment the workflow and influence it beyond native workflow decisioning
  • Demonstrates both smart routing and adjustments to workflow based on AI decisioning

Watch Mathias demo the solution: 

What Cognizant built & why

It typically takes organizations 10 to 12 weeks (or more) to build even a relatively simple workflow. Long development timelines often result in delayed product releases and/or a final product that fails to harness planned or potential business value. 

In addition to significant back and forth between IT and business teams, a traditional approach requires significant time and effort for formatting and design as well as development. Meanwhile, changing requirements–an inevitable part of the process—delay time-to-market due to complex rework. 

To accelerate the long and inflexible process, Cognizant built a fully AI-powered framework that gathers business needs expressed in natural language and dynamically turns them into end-to-end workflows, including user experiences (UX). With the click of a button, Google Cloud’s Vertex AI platform then translates the prompt into steps in the workflow. Best of all, the solution eliminates the bottlenecks that changing requirements inevitably entail. You simply check the prompt, make any corrections or changes, and hit send. The solution instantly delivers a dynamically created application, all without any static workflows to  slow down the process.

As a result, the Cognizant solution supports: 

  • Multi-language capabilities by leveraging Google’s Gemini models
  • Complex, multi-step processes 
  • Multiple personas
  • Any industry or use case

How Cognizant made it happen: The Vertex AI platform x Unqork 

To deliver such a powerful and flexible functionality, Cognizant’s team built Orchestrate.AI as a layer of abstraction that sits on top of the Unqork platform. The Vertex AI platform, which provides a deep bench of flexible models, instantly maps them to a library of standardized components in Unqork, all with zero coding. As a result, the Vertex AI platform doesn’t simply automate development, it becomes the architect of the process, designing everything from screen flow to data element fields and validation. 

Unqork is the ideal application development layer to leverage the power of the Vertex AI platform. Unqork makes it fast and easy to build secure, standardized components as required. For example, Unqork’s GenAI Connector enables you to build easily scalable integrations with providers such as Google Cloud’s AI solutions—all within a single view. And because Unqork’s data-driven architecture provides a layer of abstraction, that allows us to take on the burden of maintenance across the SDLC, from updates to security and compliance. As a result, Unqork’s customers can focus on creating innovative solutions instead of managing technical debt.

Three use cases (and endless possibilities) 

As mentioned, Cognizant’s solution can be used for most workflows in most industries. However, Mathias demos three use cases to show what is possible. In each case, the solution generated workflows dynamically in real time to meet the needs of the user, the processes and requirements of the organization—all based purely on real-time prompts. 

1. Billing dispute workflow

Mathias showed how the solution could be used to automatically create a billing dispute workflow. The resulting workflow leverages AI to gather relevant information from the transaction being disputed and then dynamically prompts the user with fields to provide any missing information required to make a determination. Once the customer hits submit, the solution creates a case that is automatically sent to the back office. 

With a click, a reviewer can quickly review relevant case details, see all the steps required for resolution, and take the next necessary steps, including requesting more information from the customer. Any required navigation automatically connects the step in the process. And when the reviewer takes action, the workflow even generates an automated status update that is routed to the customer and any internal stakeholders. 

2. Special payment plan

For this scenario, the Cognizant solution leverages AI to gather all relevant customer and account information from the account in question. Then it uses AI to dynamically create a monthly payment plan based on logic articulated in natural language. 

This is used to generate a case, which is forwarded to a back-office agent to review the case. The agent can accept or amend the AI-generated plan. Once approved, the offer is automatically extended to the customer and a status update is forwarded to all relevant stakeholders.

3. Complaint processing

After gathering a natural-language prompt, the Cognizant solution creates a workflow and UX, including customer input, back-office validation, and review and next actions. 

The solution enables users to file a complaint via a chatbox using natural language. Then, it dynamically creates a complaint submission form based on that prompt, including fields necessary to gather more required information.

Further, the solution creates a case and organizes the information in a way that facilitates a rapid back-office review. Finally, the reviewer can immediately take action to resolve the case or take further steps, including requesting more information and/or forwarding it for further review. Once the reviewer takes action, the solution automatically updates the customer and other stakeholders about any change in status and further steps, if required. 

To learn more, watch Cognizant’s Jerome Mathias demo Orchestrate.AI: An AI-powered Dynamic Workflow Solution. 

Experience the Unqork platform up close

Take the guided tour