8 min Devops

Pega deploys GenAI across platform: integrations, development, adoption

Pega deploys GenAI across platform: integrations, development, adoption

Pegasystems says it is working to enable the autonomous enterprise. GenAI plays an important role in this. This involves not only adding GenAI functionality to its own products and services, but also offering more integrations with the “outside world” and enabling organizations to get everything out of Pega’s software itself.

At PegaWorld iNspire, Pegasystems’ annual event, it’s pretty clear that GenAI plays a central role. Virtually all the news it brings out has to do with it. That’s only natural, too. 2024 is an excellent year for companies that want to get started with GenAI in a thoughtful and pragmatic way, like Pega. This is in contrast to last year, when everyone seemed to be mostly busy building castles in the sky.

A company like Pega has no business at all in such a world. Pega CTO Don Schuerman aptly points this out during a session prior to PegaWorld: “We are not an AI generalist who builds big models, we focus on how to transform our customers’ business.” All with the goal of enabling an autonomous enterprise. On the GenAI front, Pega announced some things this week that should be able to help with this. We list these below.

Models from AWS, Google Cloud and Anthropic available through Pega GenAI

Not every company wants or is allowed to use the same LLMs. Therefore, when developing applications for use within those organizations, it is important to be able to offer choice. Pega wants to make that better possible with today’s updates to Pega GenAI. It has built connectors that allow it to link Pega GenAI to a whole range of external GenAI models. It will be possible to add models from AWS, Google Cloud and Anthropic to applications. This is in addition to OpenAI on Azure, which was already available.

The integration with Amazon Bedrock, a fully managed service that offers about thirty LLMs (including Amazon Titan) and Vertex AI, a similar but even more extensive offering from Google Cloud, are important to mention here. With these, Pega unlocks a huge range of additional GenAI capabilities in one fell swoop. There are also connectors for Google Gemini and Claude (from Anthropic).

Alan_Trefler_Pega
Alan Trefler, the CEO of Pegasystems, called today’s announcements the most important ones ever at the annual event of the company.

The new models will be available in Pega Connect GenAI, the name the company coined for the architecture in order to deploy them quickly. This central architecture gives organizations not only greater efficiency, but also an environment to responsibly deploy public and proprietary LLMs. After all, Pega GenAI as a whole falls under the management of Pega, with all the governance and management involved. Pega GenAI is available to all customers using Pega Cloud.

Expansion of Pega GenAI Blueprint

Pega announced Pega GenAI Blueprint earlier this year . With it, the company aims to greatly accelerate application development. If Pega-CPO Kerim Akgonul is to be believed, the company is aiming for a 50 percent efficiency gain for organizations when developing those applications. That is certainly an ambitious goal. If actually achieved, it will be a huge step forward. Also, it’s a SaaS application hosted by Pega, so they can update it very quickly with new features.

Pega GenAI Blueprint is there primarily for ideation around application development. Akgonul states quite explicitly that Pega GenAI Blueprint does not write code. He thinks that’s a “horrible idea” anyway. According to him, it is complicated enough to check the code written by humans, adding millions of lines of code written by an AI does not make life any more fun or easy, according to him.

So Pega GenAI Blueprint is mainly there to get an impression of what an application might look like. With a few mouse clicks, you can indicate what you want in it and ask the tool to make suggestions about what else it could include. It will then search for any interesting additions. It is also possible to change the language with a single mouse click. Starting in late June, users will also be able to get live previews of what the application will look like on any channels (web, mobile, etc.). Furthermore, it will then also be possible to add inputs from legacy environments such as Business Process Modeling Notion (BPMN) models. More legacy inputs will follow over the course of the year.

As mentioned, Pega GenAI Blueprint is primarily an environment to get ideas and try them out. The blueprint still has to go to Pega’s App Studio after that to turn it actually do the coding and turn it into a working application.

The market responds very well to Pega GenAI Blueprint, according to the company. Since its introduction earlier this year in April, more than 30,000 blueprints have already been created at nearly 500 customers. That makes it the most popular product the company has created to date, according to Pega. That is, no product or service had faster adoption.

We certainly see the appeal of Pega GenAI Blueprint as well. We have yet to see to what extent it will actually lead to the efficiency gains promised. It could also result in too much uniformity and people will stop thinking critically about what they want an application to look like. There is always the risk with this type of development that the lowest common denominator will become the standard. That’s not what the industry wants and needs. You may get the efficiency gains during the development of an application, but not the application you want or need at the end. It is therefore important that the people who get to work with Pega GenAI Blueprint really know what they’re doing. It should not become a place where everybody thinks up run-of-the-mill applications.

Taking corporate learning to the next level with Pega GenAI Socrates

In saying that corporate employees need to know what they are doing with Pega GenAI Blueprint, we have a nice segue to the latest GenAI announcement that we want to highlight in this article. Pega GenAI Socrates wants to make this possible for the entire Pegasystems portfolio.

Pega GenAI Socrates can be thought of as a personal tutor for the people within a company working with Pega software. The idea is that this teacher ensures that they can get as much as possible out of the software. This is good for their personal development, but obviously also good for the development of the organization they work for.

This is not just another online course that employees can take, by the way. It is an interactive dialogue between the employee and Pega GenAI Socrates. Hence the somewhat lofty name for this new tool. The Socratic Method is one of question and answer. The idea is that knowledge and expertise sticks better this way. This is not only because the interactive method of question and answer works better than simply sending a huge amount of knowledge in a student’s direction. It also ensures, if all goes well, that each student can start at his or her level. And of course grow from there to a higher level.

By naming the new corporate learning product after the namesake of the Socratic Method, one of the most famous concepts in the world of philosophy, Pega is certainly aiming high. Let’s hope that the answers and insights students receive honor the name of this sage. After all, such a name creates rather high expectations. GenAI still tends to spread some untruths. With Pega GenAI Socrates, this will undoubtedly be contained as much as possible. But it is virtually impossible to trust GenAI blindly, which is something you should be able to do with a teacher. If not blindly trusting a teacher, you should at least be able to question the veracity of what he says to you. That’s also not always easy with GenAI. On the other hand, Socrates was also known for his assertion of basic ignorance. So perhaps we expect a little too much from it.

Pega GenAI Socrates is available in some Pega Academy courses starting today. The first course is the most popular one, System Architect. More will follow in the coming months. There is support for ten languages: English, French, Spanish, Italian, Portuguese, German, Dutch, Swedish, Polish, and Japanese.

GenAI is everywhere at Pega

The announcements Pega made today at PegaWorld iNspire show that GenAI is making its mark at multiple levels within the company. From integrations with external LLMs, through increasingly deep functionality in Pega products to supporting employees in learning what is possible with Pega’s software. We do have some questions about the actual impact and value things the company has announced, which we plan to explore in more detail over the next few days here in Las Vegas. Basically, however, the direction Pega is moving in on GenAI is a good one, we think. It is adding GenAI in a pragmatic way, in places where it adds value quickly and is easily manageable. As far as we’re concerned, that’s how to approach GenAI in 2024.