This story is based on real events in the finance/tech world. Names, companies, and details have been changed to protect everyone involved. I used AI to make it more entertaining. Just so you know. 🙂
The email from Finance landed with the soft chime of opportunity: "Project Genesis - £2.8M Approved." Alistair stared at the screen, the numbers swimming before his eyes. 2.8 million pounds. For agentic AI. He leaned back in his ergonomic chair, the one they'd bought during the "Quantum Leap" restructuring two years ago, and let out a slow breath.
His team of six, huddled in the designated "Innovation Pod" (a glass box with a dying ficus), looked to him for direction. They were the "Agentic AI Research Division," a title that sounded impressive until you remembered they'd never actually built an agent.
"So," said Priya, their lead engineer, breaking the silence. "We have the money. What's the business problem we're solving?"
Alistair felt the familiar knot tighten in his stomach. This was the ritual. The grand inversion of how things were supposed to work. The board, seduced by white papers and breathless articles about "autonomous digital workforces," had allocated the capital. Now, it was on them—the tech team—to reverse-engineer a reason for its existence. It was like being given a coffin and being told to find a worthy corpse.
"We're exploring the frontiers," Alistair said, the corporate-speak tasting like ash in his mouth. "The mandate is to research and develop a proof-of-concept for an agentic system that can... uh... dynamically optimize business workflows."
Priya didn't blink. "Okay. Which workflows? For which client?
And there it was. The question that could not be answered. They had no clients. They had no workflows. The ROI was a fictional number in a PowerPoint deck that had secured the funding. Their job wasn't to solve a problem; it was to build a problem sophisticated enough to fit the solution they were now forced to imagine.
For the next three weeks, the Innovation Pod became a factory for problems. They brainstormed. They whiteboarded. They drank endless lukewarm coffee. They came up with "Agent-based supply chain reconciliation," "Autonomous customer sentiment analysis," and "AI-driven contract lifecycle management." Each idea was a magnificent, gleaming edifice of jargon, built on a foundation of nothing.
They finally settled on "Nexus," an agentic system that would supposedly manage complex, multi-departmental project dependencies. It sounded plausible. It sounded expensive. It sounded like something that could burn through £2.8 million in research and development without ever needing to prove its worth.
Alistair presented it to the steering committee. He used words like "synergistic," "paradigm," and "cognitive orchestration." He showed them a slick UI mockup that their one designer had whipped up in a week. The UI was a lie, a beautiful, interactive fiction. The agents it depicted were lines of code in a git repository that didn't exist yet.
The board loved it. They approved the next tranche of funding for "Phase 2: Agent Simulation."
That night, Alistair stayed late. He walked through the empty office, past the desks of the "Quantum Computing Team" (who were now the "Blockchain Synergy Group" after the last company reshuffle). He thought about the tax money that flowed from his paycheck into the government's coffers, only to be routed back to them as grants for initiatives like this. It wasn't a company; it was a money laundering operation for incompetence, a perfectly circular system where public funds were transformed into private salaries for solving problems that didn't exist.
He sat back down at his desk and opened a new terminal window. He typed mkdir nexus_project and hit enter. The cursor blinked on the empty line, a tiny, digital heart waiting for a pulse. He had the money. He had the team. He had the mandate. All he needed now was a business case.
The email from Finance landed with the soft chime of opportunity: "Project Genesis - £2.8M Approved." Alistair stared at the screen, the numbers swimming before his eyes. 2.8 million pounds. For agentic AI. He leaned back in his ergonomic chair, the one they'd bought during the "Quantum Leap" restructuring two years ago, and let out a slow breath.
His team of six, huddled in the designated "Innovation Pod" (a glass box with a dying ficus), looked to him for direction. They were the "Agentic AI Research Division," a title that sounded impressive until you remembered they'd never actually built an agent.
"So," said Priya, their lead engineer, breaking the silence. "We have the money. What's the business problem we're solving?"
Alistair felt the familiar knot tighten in his stomach. This was the ritual. The grand inversion of how things were supposed to work. The board, seduced by white papers and breathless articles about "autonomous digital workforces," had allocated the capital. Now, it was on them—the tech team—to reverse-engineer a reason for its existence. It was like being given a coffin and being told to find a worthy corpse.
"We're exploring the frontiers," Alistair said, the corporate-speak tasting like ash in his mouth. "The mandate is to research and develop a proof-of-concept for an agentic system that can... uh... dynamically optimize business workflows."
Priya didn't blink. "Okay. Which workflows? For which client?
And there it was. The question that could not be answered. They had no clients. They had no workflows. The ROI was a fictional number in a PowerPoint deck that had secured the funding. Their job wasn't to solve a problem; it was to build a problem sophisticated enough to fit the solution they were now forced to imagine.
For the next three weeks, the Innovation Pod became a factory for problems. They brainstormed. They whiteboarded. They drank endless lukewarm coffee. They came up with "Agent-based supply chain reconciliation," "Autonomous customer sentiment analysis," and "AI-driven contract lifecycle management." Each idea was a magnificent, gleaming edifice of jargon, built on a foundation of nothing.
They finally settled on "Nexus," an agentic system that would supposedly manage complex, multi-departmental project dependencies. It sounded plausible. It sounded expensive. It sounded like something that could burn through £2.8 million in research and development without ever needing to prove its worth.
Alistair presented it to the steering committee. He used words like "synergistic," "paradigm," and "cognitive orchestration." He showed them a slick UI mockup that their one designer had whipped up in a week. The UI was a lie, a beautiful, interactive fiction. The agents it depicted were lines of code in a git repository that didn't exist yet.
The board loved it. They approved the next tranche of funding for "Phase 2: Agent Simulation."
That night, Alistair stayed late. He walked through the empty office, past the desks of the "Quantum Computing Team" (who were now the "Blockchain Synergy Group" after the last company reshuffle). He thought about the tax money that flowed from his paycheck into the government's coffers, only to be routed back to them as grants for initiatives like this. It wasn't a company; it was a money laundering operation for incompetence, a perfectly circular system where public funds were transformed into private salaries for solving problems that didn't exist.
He sat back down at his desk and opened a new terminal window. He typed mkdir nexus_project and hit enter. The cursor blinked on the empty line, a tiny, digital heart waiting for a pulse. He had the money. He had the team. He had the mandate. All he needed now was a business case.