Low code grew. Although software application development has been able to draw upon libraries, reference architectures and various forms of predefined ‘methods’ [a programming term used to describe a software function of one kind or another]

Not the same as no-code - platforms designed to enable completely non-technical business users to ‘create’ typically smaller and restricted application functionalities, often via simplified drag-and-drop visualization tools - low-code is still for software developers who have generally spent serious years gaining degree-level qualifications and competency in hard coding. Although low-code platforms do also encompass the options for so-called citizen developer businesspeople to perform some application tuning and creation, this technology space remains a place we developers.

As low-code now becomes codified (pun appreciated, but not intended) into the fabric of enterprise software workshops around the world, a new responsibility arises. We must now look for the niche elements of software that can be more accurately optimized. We must look for dovetail joints that sit between business productivity measures and the ability for low-code to deliver functions that will Improve the commercial bottom line. We must filter through our low-code systems to make the Developer eXperience (DX) and User eXperience (UX) more intuitive… and, in general, we must continue to build better software.

Styling itself around a branded tagline that reads Build for Change, Pegasystems Inc. is a low-code platform provider that aims to fulfil upon as many (if not all) of those low-platform developments as possible. In full, Pega describes itself as a low-code platform for AI-powered ‘decisioning’ and workflow automation.

The company has this month enhanced its Pega Infinity software suite with new low-code muscle that it hopes will accelerate low-code development in line with the need to continuously optimize existing processes - with those processes being a) initially software-based processes but b) ones that lead directly and adroitly towards improved business processes.

Birth of the autonomous enterprise

What Pega (and to be honest, the rest of the enterprise software space) is working towards is the dream of the so-called autonomous enterprise i.e. a vision of a company of any reasonable size that can rely upon software systems and processes that work autonomously, update themselves autonomously and enable business operations to run autonomously, either with businesspeople being able to self-service adapt technology to their working environment needs or not have to engage with smart systems at all that just ‘figure out’ our needs in any given workplace scenario.

Pega suggests that the concept of the autonomous enterprise is a state of being and mind which unifies Artificial Intelligence (AI) and automation to help organizations become more self-optimizing to drive maximum efficiency and effectiveness. The company bids to make this possible with extensions to its generative AI technology - the previously announced Pega GenAI - which now comes with a set of 20 new generative AI-powered ‘boosters’ across Pega Infinity ‘23 to make application development easier, faster and more powerful.

Referred to in this way, we can take AI boosters to refer to connection points between Large Language Model (LLM) datasets and the use of Natural Language Processing (NLP) technologies that enable users - and in this case developers too - to humanly speak (which often also means type) intentions and requests of software development environments in order to get them to work in specified ways.

For clarification here, the company has created boosters including the AI-assisted development functionality now available in its Pega App Studio and Prediction Studio software tools - as well as elsewhere throughout its platform. This offers Chat-GPT style text queries to provide developers with help as they create and build. The company says it gives developers and data scientists the tools to build and maintain sophisticated predictive models and algorithms. According to this technology’s product specification sheets, the AI-assisted development panel recommends what the developer needs as they build workflows. “It aggregates information from a number of different sources, provides relevant information where and when it is needed the most, and nudges the developer to adopt best practices,” notes the company.

The idea that low-code caught on in the first place is somewhat interesting suggests Kerim Akgonul, chief product officer at Pega. After all, there is no low-smoking or most other low-form activities as such - although there is perhaps some low-driving with all the smart automobile technology around now. But when it comes to the Pega platform and its usage across the hard-core (low-code) developers using its tools vs. the citizen developers also now touching its platform, how should we really think about this juxtaposition point?

“If you think about the way people build software these days, it’s mostly meetings - the coding time is usually far less than the planning time, so a generative AI-driven application that is 95% right before a human gets into the loop to tweak it makes a lot of sense,” said Akgonul. “The Pega platform’s set of accelerators that get functionality into an application quickly are many and varied. Whether citizen developers are involved at the keyboard level or not, it’s crucial that they are at least involved in order to be able to address the lifecyle of work and articulate the vision of the company’s business unit and their workflows.”

Akgonul further states that his team’s mission is to optimize the underwriting process behind the creation of software development such that it is insured for success. This is primarily because an organization’s business experts have been able to capture their vision in a structured way as part of the total coding process. If companies do this right, they get to a place where business outcomes are quantified and encoded from the start… and at that point the right code can be built.

Humans-in-the-loop

The question that inevitably arises at this point, with all this automation and AI enrichment flying around, is just how safe are these technologies when we come to apply them in the workplace, sometimes in environments that may be bordering on the mission or life-critical? Pega is adamant on this point and explains that the company’s approach will help organizations leverage the power of generative AI while building controls, governance and auditing to keep humans-in-the-loop.

Pega GenAI will offer a secure, plug-and-play architecture that will allow for low-code development of AI prompts, the ability to infuse generative AI functions into their workflows and the freedom to swap in different Large Language Models (LLMs). This approach will help users drive value from generative AI now and into the future. So has the company locked down the security and robustness factor here enough?

“Pega has been working hard to put the right architecture in place with regard to every aspect of generative AI safety, control, management and oversight that is possible in both human and machine terms,” clarified Don Schuerman, Pega CTO.

Schuerman made his comments this month while speaking directly on the subject of whether ‘safer’ generative AI needs to feature not just Large Language Models and open public AI engine power, but also incorporate (and in some cases insist upon) a proprietary data approach where Small Language Models (SLMs) are also used to work on company-specific sensitive data sources - a process some are referring to as private AI as opposed to the public alternative.

“We are well-planned and on the right trajectory for this particular spiral arm of the AI galaxy,” enthused Schuerman. “While some industry vendors are entering the AI space for the first time, ‘private AI’ isn’t a new idea for Pega. We’ve had clients running our Customer Decision Hub technology and our Process AI tools to build and run their own private, self-learning AI models for over a decade. During that time, we’ve built up an enviable bank of experience, best practices and product offerings for aspects of software architecture such as MLOps (our software code revision manager tool manages deployments, testing and retiring of models), AI explainability (our transparent AI ‘T-switch’) and bias test controls to govern AI models. Our clients have deployed tens of thousands of these private models over the years.”

Going further, Schuerman said that when it comes to generative AI for Pega, the company’s architecture is built around a technology known as a Connect-Generative AI layer. This allows software developers and data engineers to build their own prompts using low-code tools and will abstract the prompt itself from the different AI models. Essentially, this is the sort of air-gapping guardrail that those of us in the data science space have been calling out for as a sort of software safety belt all year.

“We have also included mechanisms to automatically replace sensitive Personally Identifiable Information (PII) data with anonymous data in calls to public models. Requests flow through a proxy service that provides an additional layer of protection. In the future, we plan to offer our own private Gen-AI services [in the realm of the Small Language Model approaches already noted here] as well as the ability to for users to run their own local models,” added Schuerman, in what is an arguably more comprehensive response to this entire point of sensitivity than many other commentators.

Inside workings in low-code

In addition to its aforementioned new generative AI-powered boosters, Pega Infinity ‘23 will also include updates to Pega Platform, Pega Customer Decision Hub and Pega Customer Service - all of which form central building blocks to the company’s core IT stack offering.

The types of functionalities on offer here provide us with some insight into how low-code platforms actually work. A reuse library in the Pega App Studio tool enables developers to find reusable software reusable components and pre-engineered business logic while also sharing those resources with others. Pega says that this same function can also be used by citizen developers, although perhaps not quite extending to the no-code end of low-code. Allied to this technology is the Pega Constellation offering and these are software tools focused on improving the User eXperience (UX) factor in enterprise applications through the use of built-in templates and patterns.

There’s an inclusivity factor here in low-code with Pega Infinity ’23, as there of course should be. As such, Pega now offers its approach to enhanced accessibility compliance and does so by following in line with Web Content Accessibility Guides (WCAG) 2.1, a set of guidelines laid down by the World Wide Web Consortium (W3C). There are also connector technologies here designed to integrate Pega Customer Service (technology that provides functions such as self-service workflows to automate customer query resolutions) with Pega Process Fabric (technology designed to unify the discoverability and prioritization of human - and increasingly now also, machine - workplace tasks) all of which are built to work under the Pega low-code platform umbrella.

“Pega Infinity unifies customer engagement, customer service and intelligent automation capabilities within a single platform to help accelerate digital transformation,” notes Pega, in a technical product statement. “By engaging customers with the right message, at the right time, across any channel, clients can Improve customer satisfaction, increase customer lifetime value, and boost productivity.”

Application iteration liberation

At the risk of coining a marketing phrase that one or other low-code vendor (Pega or its competitors) might be desparate to adopt, what all these developments push towards is something we could call an application iteration liberation.

In simple terms (and we learned this so well during the pandemic, obviously) organizations in every vertical need to be able to ‘iterate’ their applications to progress them to support new functionalities, connect to new data sources, draw power from new cloud services, work on new form factor devices when they come around and reduce the complexity for accelerated deployments everywhere.

Pegasystems (Pega) clearly draws its name from the winged divine stallion depicted in Greek folklore and Pegasus itself was ultimately brought to Mount Olympus by Zeus where it was stabled next to the thunder god's thunderbolts. Can we see the parallels here? At least low-code is no myth.

Low-code automation software firm Pegasystems Inc. posted disappointing second-quarter financial results that saw it miss analyst’s targets on earnings, revenue and adjusted operating income.

The company reported a net loss for the period of $46.8 million, improving on the loss of $268.2 million it posted one year earlier. However, its earnings before certain costs such as stock compensation came to just a penny per share, below the forecast of five cents per share. Revenue rose 9%, to $298.3 million, but came in below the analyst’s target of $309.6 million.

Pegasystems is a provider of low-code automation software that helps people who aren’t programmers automate manual tasks, unify business processes and customer journeys. The software helps companies to get a better handle on the multiple business applications and systems they use. By creating a configurable platform that sits above those other systems, the Pega Platform provides businesses with a single view of their customers, cases and workflows, together with all of the associated data.

Although Pegasystems software is often compared with that of robotic process automation firms such as Automation Anywhere Inc. and UiPath Inc., it’s generally seen as a more bespoke offering.

In a statement, founder and Chief Executive Alan Trefler (pictured) spoke of an “uncertain and changing environment,” and said it’s more important than ever to focus on client success. “Our low-code platform for AI-powered decisioning and workflow automation uniquely empowers clients to embrace emerging trends like generative AI and, at the same time, reduce costs and Improve customer engagement,” he said.

Trying to find a few bright spots from the quarter, Pegasystems pointed to annual contract value growth of 13% from the same period last year. It also said Pega Cloud’s gross margin expanded to 73%.

Though the results missed expectations, analyst Holger Mueller of Constellation Research Inc. said he sees evidence that Pegasystems might have turned the corner, so to speak, with revenue growth that almost reached double-digits. “It made progress on cost reductions and it showed good product execution, too,” the analyst said. “Now the onus is on Trefler and team to make better progress towards profitability by the end of the full year.”

Investors were simply not convinced, however, and Pegasystems’ stock fell more than 10% in the morning session after the report.

Prior to today’s drop, Pegasystems’ shares had gained more than 59% in the year to date, compared with the S&P 500’s average gain of just 19%.

Photo: SiliconANGLE

Some of AIA's 2023 Young Architect award winners from left: Shannon Gathings, AIA, Jason Takeuchi, AIA, Kate Thuesen, AIA, Chris Haedt, AIA, Stephanie Vito, AIA, and Caroline Shannon, AIA.

Earlier this month, the Financial Times released a damning exposé revealing a long history of abuse committed by one of architecture’s biggest stars, Ghana-born British architect David Adjaye. The news came as a shock, as Adjaye, who is Black, has long been seen as a beacon of progress in the overwhelmingly white profession of architecture. Best known in the United States for designing the National Museum of African American History and Culture in Washington, D.C., he has been accused by three women of sexual exploitation, harassment, and creating an unbearably hostile working environment at his firm, Adjaye Associates. Through a lawyer, he has denied all allegations.

The details of the report are horrifying, painting a picture of a man who, after accruing power and success, abused that power for his own gratification for many years and bought his victims’ silence. His pattern of abuse continued unabated for so long because, as their employer, he had disproportionate power over these women—all three of them Black, a not-incidental element of how he viewed and treated them (that is, as disposable). One woman said he called Black women “low-hanging fruit,” and said, “If I was white he would have had respect for my body.” He also had a hold over the immigration status of his workers in Accra, Ghana; an FT follow-up revealed that he gave the names of his victims to the Ghanaian government.

The report contained some especially lurid, violent details, which immediately prompted outrage. The financial fallout was immense. Adjaye’s role in many of his projects under development was terminated, including some blockbuster works such as Westminster’s Holocaust Memorial and Harlem’s Studio City. Of course, no amount of scuttled commissions will ever be enough to supply those three brave women years of their lives back.

Many were relieved to see any allegations against a powerful male architect see the light of day. In some circles, admittedly unsubstantiated rumors about Adjaye were circulating as early as 2017, at the height of the #MeToo movement. Many women in architecture will remember the now-deleted “Shitty Men in Architecture” list, which featured many big names in the field. Such whisper networks among women in practice and students in the academy are often the only form of information-gathering on gender-based violence in the field.

To say that architecture is a profoundly misogynistic discipline is an understatement. As a woman in the field, I have experienced it firsthand. In many instances I have been treated with less respect than my male counterparts and belittled by men, even at public events. And misogyny in the workplace isn’t just limited to sexual harassment. In the United States alone, despite their receiving 48 percent of all architecture degrees, women make up only 17 percent of the architects registered with the American Institute of Architects, and of those, fewer than .05 percent are Black women. Women in architecture only make 86 cents to every dollar earned by men. In Adjaye’s case, institutional misogyny is compounded with racial contempt. His alleged predations may best be described as what Black feminist scholar Moyra Bailey calls “misogynoir,” the form misogyny takes when directed at Black women in particular.

It’s no mystery how Adjaye was able to fly under the radar for so long. Architects like him achieve nigh-mythical status as solo geniuses whose groundbreaking work changes the very shape of art, if not the world. It does not matter to the public that he is merely a synecdoche for his entire firm, Adjaye Associates, a company made up of hundreds of subordinate employees. Adjaye himself is the only one who receives public recognition for the work done by the army of laborers responsible for making architecture actually happen. When public focus rests on one man and erases architecture’s collective project, that man gets elevated to a level of power and influence that makes it harder and harder for victims to speak out.

To make things more complicated, the field’s exclusion of minorities—coupled with its predilection for patronizing tokenism toward them—has also given Adjaye, a Black man, an extra layer of protection and insulation from criticism. He is not just a famous architect; he is a trailblazer. No one wants to be the employee who ruins the career of one of the few famous architects of color. It’s reasonable to anticipate that their allegations will be weaponized by bigots to reinforce their stereotypes and prejudices. The fact that all three women were Black should remind us that the role race plays in architecture is more nuanced and intersectional than Adjaye’s photo ops with Barack Obama would have us believe. His fall should put to bed the idea that one man’s success in a world that views him as an other is in any way tantamount to a real reckoning with race in architecture.

It bears repeating: Adjaye is an employer. His abuse is workplace abuse—it cannot take place without the infrastructure of the workplace to provide him power and access to victims. While his case is particularly spectacular in its violence, it is of a kind with abuse that happens in architecture firms around the world. All workplace abuse is irrevocably linked with worker precarity. The environment of architectural workers is particularly conducive to exploitation, since it encourages self-abuse: long hours, unpaid internships, unnatural devotion to “the project,” and identification of the self with the workplace.

Adjaye’s cult of personality and status as an employer had far-reaching consequences for his workers’ financial and personal autonomy. The FT report details how many of his workers were tied up in the vulnerabilities of immigration-based contracts. Many were not getting paid or were getting paid late. Then he heaped loads of hush money on workers he had made financially fragile—one woman described having to choose between diapers and food.

These allegations should not be viewed as the ignoble and unfortunate end of what was once a fairy-tale story. They should not be viewed as an isolated instance of brutality. They should be viewed as a wake-up call. All of the elements that allowed Adjaye’s harm to go unpunished for so long are present in one way or another in all firms. They are inherent in the very culture of the discipline, which has become increasingly stratified, with entry-level workers seen as especially disposable and exploitable. Young architects, after being told all through school that they will be embarking on a journey to change the world and shape the built environment, instead find themselves working 10-hour days using mind-numbing software to catalog how much insulation is needed in a given wall. Receiving any scrap of acknowledgement from the great masters who run their firms more like despots than artists feels especially rewarding in such an uninspiring environment. Combine this dynamic with a culture of virulent racism and misogyny, lack of financial security or upward mobility, and precarious employment visas, and you have an environment that is primed for exploitation. The fact that this exploitation takes on a sexual dimension is no surprise when domination—over the workplace, perhaps over the built environment itself—is the order of the day.

A solution to these problems requires a world in which architectural workers see themselves as workers and where starchitects like Adjaye are no longer seen as gods. It also requires labor organization in the workplace; as in all corporate settings, institutions like HR (if firms even have it) are designed to protect the company, not its workers. Whether through activist organizations such as the Architecture Lobby or through unionization, architectural workers need accountability and support from outside their firms. Unionism in architecture is in its infancy, but solidarity among the field’s workers is rising year by year.

The only way to keep such awful acts from happening again is to change the very structure of architecture. We owe Adjaye’s survivors a reckoning beyond punishing their abuser for his misdeeds. We owe it to them to create an architecture that is supportive and safe. We owe it to them to reform the workplace into one where opportunities for exploitation and race- and gender-based violence are eliminated. We owe it to them to change the field for the better, to make sure that what has happened here can never happen again.