(Some) software developers may now evolve to become orchestrators
Coding is dead. No, it’s not.
Good, that’s out of the way and crystal clear. Of course, software application development is not going away, despite the rise of AI and the multifarious layers of predictive, generative and agentic intelligence, automation and acceleration services that now pervade.
Software developers will still drive the core constructs of any given codebase, they will determine the algorithmic logic that underpins the execution of applications and data services, they will regulate the interconnection points that software has with third-party components and they will dictate the aesthetics of the user-level functions that applications have, once we get our hands on them.
All that said, some software development is changing and the role of AI in areas such as code completion, alongside syntax error and misconfiguration analysis, will be supremely useful. With system maintenance and updates, anomaly detection and everything through to user acceptance testing now being ripe for a degree of automation, some aspects of coding will benefit from AI, but deciding not to learn Java, Python, SQL or C++ is still arguably foolhardy.
If these core constructs are even 20% true, what will actually happen when AI “starts work” inside the development team?
A raft of new methodologies
Some argue that platform engineering will prevail and we’ll see a massive focus on higher (in fact lower) infrastructure as the primary facilitator for code generation management. With standardised tools delivered from an internal developer platform, an IDP embodies platform engineering’s approach to developer self-service capabilities, where automation drives the developer’s curated workflow environment.
Others advocate so-called vibe-coding i.e. where developers “express” the functions they want from a resultant piece of software and sit back as AI does the rest, at least until humans need to come in and debug it. Others still say that developers will now step forward into a kind of orchestrator role where AI does a lot of the work and the coding team gets more closely aligned with business goals and user requirements.
For some developers, progressing to these new constructs will not sound like progress; these are the command line purists who will use AI-augmented development to a degree, but won’t be comfortable talking to or working with users that directly. There’s a reason developers talk about “operator error” when an app doesn’t work properly… but let’s move on.
As the dust on this dustup starts to settle, a number of technology vendors will be keen to ride into the new dawn of developer orchestrator functions.
A system of orchestrated intelligence
Ali Arsanjani
“In this [new developer orchestrator] world, code authorship becomes a smaller – but still meaningful – part of a broader system of orchestrated intelligence. Tools like Google’s open source Agent Development Kit and the Agent Engine (Runtime) are making this orchestration accessible. These frameworks allow engineers to deploy, configure and compose agents that autonomously reason, plan and interact via real APIs – reducing the complexity of managing agent lifecycles, giving optionality in either designing or deploying on OSS frameworks,” writes Ali Arsanjani, director of applied AI engineering at Google, on his own blog post.
As the reality starts to bite here, developers may start to look at their “typical day” and agree that they spend a lot less time coding than they might like i.e. there are a glut of primary tasks and roles that need to be addressed first such as system maintenance, cloud provisioning, scalability reasoning, roadmap planning and eating pizza. With so much integration and configuration to shoulder in the increasingly cloud-native world, there are a huge number of pre-flight checks that every coding team needs to go through, perhaps the shift to orchestrators of intent won’t be so bad?
AI-native enterprise software company DevRev claims to be “collapsing the gap” between product engineering and customer experience, giving developers a direct line of sight into how their work shapes user value. The company’s logically-named Computer service aims to unify structured and unstructured enterprise data to enable AI agents that answer natural language queries. With the drive to automation quite prevalent across its platform, what does DevRev co-founder and CEO Dheeraj Pandey think about the state of developer orchestration today?
“I’m watching developers make a subtle but profound shift. They’re moving from writing code to architecting intelligence. Developers are defining what an AI agent should learn from every interaction, every ticket, every commit. The challenge isn’t writing code faster. It’s building systems that remember, that connect context across every tool a business uses, that get smarter with each decision,” said Pandey.
Pandey: New company, new sweater, new smartwatch… same engaging warmth, charm and insights.
“The craft is different now. Less about lines of code. More about judgment calls: what should the system learn on its own? When should it act? When should it ask for help? Developers aren’t disappearing behind the curtain of AI-generated code. They’re becoming the architects of how humans and machines reason together. They’re building systems that think with us, not just run for us,” added Pandey.
Developers govern guardrails
Other work in this space is being surfaced by AI-automated business process management company Kognitos. The organization says it wants to remove the language barrier in automation, so that business logic can be expressed in English while developers design the guardrails and governance.
According to the company’s AWS marketplace descriptor, using Kognitos means teams no longer need to spin up infrastructure or hire RPA developers. Because users develop in natural language English, it opens up the door to who can automate; all exceptions are now handled at the business level, reducing the need for a large center of excellence.
Binny Gill, CEO of Kognitos is confident that we’ve moved on from what he calls the “era of brittle automation and unpredictable AI” where automations fail to deliver the requisite services needed inside mission-critical environments. He states that his company’s platform delivers “powerful, non-hallucinating AI agents” that understand business intent with a capability to document their operations and meticulously follow processes. Just the sort of thing that users will like, it’s less clear how software application developers will feel about these services if the spectre of agent sprawl is indeed real.
Binny Gill
“Just as assembly language faded from view, today’s programming languages will become invisible layers of abstraction. The new source code is English, or whatever natural language people think in. Developers won’t vanish; they’ll define the structures, policies, and reasoning that AI systems follow. What changes is the medium, not the mindset. The future of software isn’t about typing syntax; it’s about expressing logic in the language of intent. When machines understand that language, everyone becomes part of the development process and developers become the architects of how intelligence works, not just how it’s written,” said Gill.
Enterprise back-office automation
Agentic AI-based finance automation specialist Auditoria wants to create new ways to operate enterprise back-office automation, applying AI agents to finance and operations so that developers can embed intelligence directly into the workflows that underpin business execution.
“AI is redefining how developers contribute to enterprise value, especially in finance. At Auditoria, we’re moving from building isolated tools to orchestrating systems of intelligent agents that manage and optimise core financial workflows. For us, that means developers focus less on manual process automation and more on designing governance frameworks that ensure accuracy, compliance, and adaptability. The outcome isn’t just faster close cycles or fewer manual tasks, it’s a finance office that learns and evolves continuously. And developers are now building the architectures of operational intelligence, systems that shape how finance teams think, decide, and act in real time,” said Rohit Gupta, CEO and Founder, Auditoria.
Taken together, these platforms suggest a future where developers aren’t displaced by AI, they’re elevated into conductors of complex, intelligent systems.
Free image.