Pegasystems Inc., the low-code platform provider empowering the world’s leading enterprises to Build for Change, announced Pega GenAI – a set of 20 new generative AI-powered boosters to be integrated across Pega Infinity ‘23, the latest version of Pega’s product suite built on its low-code platform for AI-powered decisioning and workflow automation. It will be showcased next month at PegaWorld iNspire and available in Q3 2023. Pega GenAI will provide organisations with the architecture and integrated use cases to drive value from generative AI now and into the future.
For example, a bank looking to automate its loan processing operations would
traditionally need to start by identifying, designing, and developing dozens of workflows
from scratch. With Pega GenAI, they simply tell Pega they are building a “loan processing
application,” and Pega will automatically create the related workflows, data models, user
interfaces, sample data, and more based on responses from generative AI models like
those from OpenAI. Because the responses from generative AI are mapped directly into
Pega’s powerful model-driven architecture, low-code developers can easily configure and
change these suggested starting points to rapidly deliver a completed application. Pega
GenAI boosters like these will be infused throughout Pega Infinity, allowing users to
accelerate their low-code application development, enhance customer service, and
improve customer engagement.
A new API abstraction layer, called Connect Generative AI, will allow organisations to get
immediate value from generative AI with a plug-and-play architecture that allows for low-
code development of AI prompts. Rather than directly calling OpenAI, or other APIs directly from UIs or workflow steps, Pega uniquely provides an API abstraction layer so developers can easily swap out large language models running on both public and
private clouds and build reusable generative AI components that can be leveraged across
applications. Connect Generative AI will be able to automatically replace personally identifiable information (PII) data with placeholders in generative AI prompts, helping
organisations enforce their data protection policies and advancing the secure use of public and private models.
Generative AI-powered boosters in Pega Infinity ’23 facilitate the rapid development of
innovative new capabilities and give low-code developers the power to infuse generative AI
functionality into decision-making and workflow automation. As large language models,
cloud services, and data privacy needs continue to evolve, and this “AI choice” architecture
allows Pega and its clients to continuously innovate new security solutions. Pega will initially offer connectors to OpenAI’s API and Microsoft Azure’s OpenAI APIs and will be
supplemented by additional downloadable connectors to other providers on Pega
Marketplace.
The 20 new Pega GenAI boosters will help make Pega Infinity ‘23 even easier, faster, and
more powerful to use, including:
Faster low-code application development in Pega Platform:
AI-prompted workflows: When users type in the name and short description of the
new application they want to build, Pega GenAI will provide relevant suggestions for
different workflows and build out their associated stages and steps.
AI-generated personas: Pega GenAI will provide definitions for relevant users and
personas associated with a workflow, allowing low-code developers to route work
appropriately and generate relevant user interfaces.
Automatic data modeling: Pega GenAI will provide suggestions for different types of
data likely needed for an application and model it into the workflow and user
interfaces.
Back-end integration assistance: Pega GenAI will speed application integration by
automatically understanding the target system’s APIs and mapping the
corresponding workflow data fields.
Sample data generation: Users will be able to test and demo their application using
sample data automatically generated by Pega GenAI, reducing hours of low value
work by developers and improving the quality of testing.
Enhanced customer engagement relevancy in Pega Customer Decision Hub™
Treatment creation assistant: Pega GenAI will help users produce better treatments
by providing text and image suggestions. Users can select from Cialdini’s principles
of persuasion to create variations in tone for the copy to appeal to different
customers.
Autonomously recommend higher impact actions: Pega GenAI will present a list of opportunities to improve the Next-Best-Action library along with automatically
generated suggestions for how to make those improvements. Suggestions will come
from analysing underserved customers and identifying actions that don’t have
enough treatment variation for the AI to optimally work.
Explainable AI analytics: Pega GenAI will provide visibility into how AI decisions are
made by analysing decisioning data and providing easy-to-understand explanations
of why certain actions would (or would not) be presented to a customer.
Population targeting and validation: By automatically generating sample personas for different populations, Pega GenAI will help users more quickly test the logic in their
engagement policies. Improved customer service in Pega Customer Service.
Automatic interaction summaries: Pega GenAI will automatically summarise message transcripts between agents and customers. This will reduce the need for customer
service agents to manually write notes and significantly improve handle times in
contact centers.
Accelerated chatbot training: Pega GenAI will generate labeled training data for
chatbots, anticipating likely utterances and phrases from customers. This will reduce
the time and effort needed to deploy new customer service bots.
Customer interaction simulator: Organisations will be able to use Pega GenAI to role-play with their customer service agents to train their staff, identify agents in need of
additional training, find process bottlenecks, and evaluate staffing levels.
Advanced chatbot response: Pega GenAI will assist a chatbot that can’t identify a
suitable answer on its own. For example, it could help the chatbot better
understand the customer’s intent so it can improve its chances of a relevant
response. Generative AI output can further train and enhance the natural language
models to teach the chatbots how to respond to new inquiries.
More productivity in Pega Sales Automation
Email reply generator: Pega GenAI will scan incoming emails and offer suggested
replies.
Meeting summary generator: Salespeople can use Pega GenAI to generate sales
meetings summaries and action items for meeting attendees by analysing meeting
transcripts from Pega Voice AI, Zoom, or WebEx.
Faster bot creation with Robot Studio
Script generation: Pega GenAI will automatically generate custom scripts that can be
called directly from the automation surface, which helps simplify development and
debugging.
Test data generation: Robotics developers will save time by using Pega GenAI to
automatically create input parameters for testing bot automations.
Enhanced knowledge management libraries in Pega Knowledge
Semantic search: Users will be able to ask Pega GenAI natural language questions
and receive concise, summarised answers based on knowledge management
articles.
Generation of new answers: Knowledge management content authors will be able to
use Pega GenAI to help create new, succinct articles based on summaries of existing
content.
Easier access to robust operational insights
Automatic insight generation: Users will be able to generate reports via Pega GenAI by making requests such as, “Show me the sales figures for the Midwest region in Q2.”
Pega’s approach to generative AI allows organisations to confidently deploy their AI models
of choice in a responsible and governed way while minimising risk. It incorporates auditing,
rules-based governance, and workflow-managed human approval to advance safety,
security, and reliability. Pega will allow for all AI-generated text to be reviewed, edited, and
approved by authorised staff to mitigate the risk of inaccurate or biased text from being
exposed to customers.
Pega GenAI builds on Pega’s decades of experience in applying computer intelligence,
including rules and data-driven AI, in responsible and effective ways. It complements the
already powerful AI-powered decision engine in Pega’s low-code platform, which brings
together decision management, predictive and adaptive analytics, natural language
processing, voice recognition, business rules, and a robust set of MLOps and testing
capabilities for monitoring and governing AI models.
Full details on additional Pega Infinity ‘23 features will be showcased at PegaWorld iNspire,
the annual Pega user conference, from June 11-13 at the MGM Grand in Las Vegas.
Quotes & Commentary
“We’ve added a staggering number of powerful generative AI-powered boosters across
Pega Infinity helps organisations quickly leverage the power of generative AI to work
faster and more efficiently,” said Kerim Akgonul, chief product officer, Pega. “Our clients will have the flexibility and security to power our generative AI features and build their own
using their large language models of choice. The Pega GenAI boosters released in Pega
Infinity ‘23 is just the start — we plan to add new ones on a regular basis as we continue to
evaluate new ways to responsibly and securely leverage generative AI and as new models
come to market.”
