Low Code is Dead… Long live the Agents!
Highlights from Microsoft’s Power Platform Community Conference 2025
Las Vegas isn’t known for understatement, and neither was this year’s Power Platform Community Conference (PPCC 2025). Over four days, more than 7,000 attendees packed the halls to witness a movement in full transformation. What began a decade ago as a grassroots low-code experiment has matured into an AI-powered ecosystem redefining how software is imagined, built and governed.
The headline? Low code has evolved beyond code at all. AI agents are now the makers’ new collaborators, and Power Platform has become the proving ground for a future in which software builds itself.
The End of Low Code as We Know It
Charles Lamanna, Microsoft’s President of Business and Industry Copilot, opened the week with a line that set social feeds alight: “Low code as we know it is dead.” Far from a funeral, it was a call to arms.
Lamanna’s point was that intelligence is now “an API call away.” With GPT-5-class models and reasoning engines underpinning every tool, the creative act of software development has become a conversation. “Ideas are becoming limitless,” he said. “Whether you write code or drag-and-drop, those boundaries are blurring.”
He reminded the audience that Power Platform now counts 56 million monthly active users, a 40 percent jump year-on-year — evidence that the maker movement has entered the mainstream. Demos showed agents completing data-entry tasks in seconds, apps auto-generating screens and data models, and workflows emerging from a single sentence of instruction.
From Copilot to Colleague
If Lamanna declared the revolution, Ryan Cunningham and Dan Lewis showed what it looks like in practice. Cunningham, Microsoft’s Corporate VP for Power Platform, described this as the “middle innings of AI” — no longer the hype cycle’s dawn, not yet the era of maturity, but a messy, creative phase where breakthroughs multiply daily.
Copilot Studio, he explained, has moved from a design tool to a genuine agentic environment: a place where autonomous but governable agents reason, act and learn inside the enterprise boundary. “It’s not a feature,” Cunningham said. “It’s the thing.”
Behind the scenes, Microsoft engineers now generate and test hundreds of new apps every night, allowing AI graders to evaluate and score the results — a continuous-improvement loop that refines both the agents and the platform itself. The focus has shifted from flashy demos to sustained impact and measurable ROI.
Lewis, who leads Copilot Studio, outlined how Microsoft is packaging pre-built “first-party agents” — for sales, service, HR and finance — so customers can start quickly and customise safely. His advice to executives starting late: give employees self-service AI in their daily flow of work, dedicate a small team to experimentation, and build from there. “Start somewhere, but start now,” he urged.
Intelligence on Tap
Bryan Goode coined the phrase of the week: “Intelligence on tap.” He likened AI to electricity in the early industrial age — once scarce and costly, now abundant and ubiquitous. With 64 percent of business leaders already using AI daily, and IDC projecting one billion enterprise agents by 2028, Goode argued that the question is no longer if firms adopt AI, but how fast.
Microsoft’s new App Builder feature for Microsoft 365 Copilot epitomises this shift. Type a natural-language description of the tool you need, and the system assembles it instantly — data model, workflow, user interface and all. Edit by conversation, share with colleagues, and watch it improve through usage feedback.
“It’s never been a better time to be a builder,” Goode said, underscoring that AI is now a universal productivity layer. The next wave of “frontier firms,” he added, are those putting intelligence at the heart of every process.
Becoming a Frontier Firm
That concept of the frontier firm ran through the week. Phil Topness unpacked Microsoft’s framework for becoming one, built on four pillars: empowering employees, reinventing customer engagement, reshaping processes, and bending the curve on innovation.
His workshops showed that the journey is measurable. In pilots with enterprise customers, AI agents have delivered 18-hour faster resolution times in HR queries, cut onboarding effort by double digits, and enabled executives to shift from reactive to predictive decision-making. “Ten percent cost savings are table stakes,” Topness said. “Frontier firms look for 10x transformation.”
He also warned that governance and trust must mature alongside velocity: “Administrating this platform is a little bit magic and a little bit policy. You need Security, HR and Legal in the room early.”
Real-World Proof
Customer panels translated vision into reality.
At NFP AON, Head of Engineering Jayaprakash Subramaniam described a rapid journey from 1,200 to 8,000 Copilot licences across a workforce of 8,000 people. Weekly usage now tops 90,000 Copilot interactions, and adoption is spreading organically. “It’s not just numbers,” he said. “It’s a cultural shift in how our people perform their work.”
That cultural change rested on foundations laid years earlier: consolidated CRM systems, unified data management in Azure, and migration to modern device management. Subramaniam credited this groundwork for enabling fast AI rollout when opportunity struck.
Other customers echoed the same playbook: data consolidation first, AI second. Financial-services firms spoke of automating onboarding and compliance; manufacturers shared savings in service-desk operations of 20–30 percent after deploying Copilot agents; and government agencies emphasised the productivity and well-being dividends of reclaiming hundreds of hours for staff development.
Makers, Engineers and the New Craft
By Day 4 the focus turned to the builders. Simon Matthews from Microsoft’s developer-agents team demonstrated the latest Plan Designer, where specialised agents — requirements, data and code — collaborated live to generate a functioning conference-scheduler app from a single paragraph of text.
In minutes, the system analysed user roles, produced a six-table data model, generated sample data, and created working screens — complete with colour palette and navigation design. The message was unmistakable: agents need apps, and apps need agents.
At a parallel panel, veteran MVPs predicted that natural language will replace low code as the default interaction model. “We’re moving from forms to conversations,” said one. The pro-dev community’s new role, they agreed, is becoming AI orchestrators — shaping prompts, validating models, and ensuring the “human in the loop” remains strong.
The Human Element and the AI Operating System
Despite the buzz of automation, every panel returned to the same truth: success still depends on people. Governance, security and culture dominated the “Ready, Set, Scale” adoption panel, where leaders argued that centres of excellence for AI are as vital as those for low code.
One executive noted that without common architecture and telemetry, small pilot projects risk stalling. The companies achieving sustainable scale — some with 7,000 apps and 18,000 automated processes — are those building on a single, governed platform with shared data policy and clear accountability.
The closing Ask Me Anything turned to practicalities: data-labelling, SharePoint permissions and how to scale Azure AI Foundry alongside Copilot Studio. The consensus was clear — pro-code and low-code paths are converging fast. React-based code apps can now coexist with Power Apps under unified governance and security.
“It’s not either-or,” one Microsoft leader said. “It’s the same story told through different skill sets.”
Our Time to Build
By the final morning, Ryan Cunningham’s closing keynote captured the mood. He recalled Power Platform’s first hackathon in 2017, when a student’s novelty fidget-spinner app proved the platform’s creative potential. Eight years later, a thousand developers were still up at midnight in the same ballroom, furiously building agents instead of spinners.
“This is our time,” he told them. “We’re standing at the centre of how all software will be built.”
The numbers back him up: tens of millions of makers, billions of AI agents on the horizon, and a platform evolving faster than any in Microsoft’s history. PPCC 2025 wasn’t just another conference. It marked the moment low code grew up — when makers and machines truly met, and together began writing the next chapter of enterprise innovation.