When Someone Explains the Impossible
My first week at Citizen, someone explained what the operators actually did. I thought they were messing with me.
"So they listen to police, fire, AND EMS radio channels. Simultaneously. At 3x speed. They memorize the different department ten-codes on the fly — NYPD, Fire Department, EMS all have different lingo they use. They listen for locations, descriptions, severity levels. Then they type it all into the incident app on a second monitor. For 8-12 hours straight. Processing 100+ clips per hour during peak periods."
I stared at them. "That's... not humanly possible."
"Yeah, but they do it. They're essentially EMS, police, and fire dispatchers combined, handling every emergency type that crosses NYC airwaves."
I had to see this for myself.
Watching Meghan Work
The first time I shadowed Meghan, I couldn't believe what I was seeing.
Click. Police radio: "Central, we got a 10-53 at Queens and Northern, possible 10-54 on scene..."
3x speed. Her fingers flying across the keyboard while her brain parsed: "vehicle accident with injuries." Possible 10-54 — ambulance might be needed soon. Northern Boulevard intersection.
"Wait — what? When did they say... how did you...?"
Second monitor. She's already typing incident details, estimating location, setting severity level.
Next clip. Fire dispatch: "Battalion 14, 10-36 Queens Boulevard and Northern Boulevard, multiple calls coming in..."
Then immediately: EMS radio: "Dispatch, 23-Adam requesting 10-32 at Queens Boulevard Northern, standby for patient count..."
Three different departments. Three different codes. Same accident. Her brain instantly connects them while her fingers keep moving.
I watched her handle 15 clips in just a couple minutes. Police shootings, medical emergencies, building fires. All at 3x speed, all decoded instantly, all input accurately.
"This is incredible," I told her during a break. "How do you even—"
Her screen froze mid-sentence.
She didn't curse. Didn't complain. Just let out the quietest sigh, hit Ctrl+Alt+Delete, and started her browser over. For the third time that shift... like it was part of the job.
The Color That Broke My Heart
That's when I noticed the live dashboard on a monitor on the wall. A grid of operator names, each with a color indicator next to it.
Meghan's name had turned red.
"What does red mean?" I asked.
"Falling behind on clip queue," someone explained. "Yellow means struggling to keep up. Green means on track."
I watched Meghan's screen boot back up. She pulled up the audio interface, found her place in the queue, and dove back into 3x speed police radio without missing a beat.
The red indicator stayed red.
Not because she wasn't fast enough. Because the tools kept crashing.
That night, I couldn't sleep. These people had adapted to doing the impossible. They'd gotten SO good at this insane cognitive process that when the technology failed them—which happened constantly—they just absorbed it. Incorporated crashes and memory leaks into their workflow like it was normal.
And management had built a system that labeled them as "behind" when their tools failed.
The Normalized Breakdown
Over the next few days, I watched the pattern repeat:
Sarah's audio interface crashes during a major incident. Quiet sigh. Restart browser. Continue processing emergency clips.
David has to mouse-click to change playback speed from 2x to 3x, interrupting his flow every few seconds. He's adapted by anticipating which clips need faster processing.
Lisa refreshes her page every hour to prevent memory leaks, losing her place in the queue each time.
They'd all adapted to dysfunction. These brilliant, dedicated professionals had incorporated technical failures into their mental model of the job.
The most heartbreaking part? They thought this was their fault.
"I'm just not as fast as I used to be," Meghan mentioned one day after her fourth browser restart.
I wanted to scream. She was processing emergency audio at 3x speed while instantly recognizing that NYPD 10-53, FDNY 10-36, and EMS 10-32 were all the same incident, and she thought SHE was the problem?
What If We Just... Fixed the Small Things?
I started with something tiny. Keyboard shortcuts for playback speed.
Instead of mouse-clicking to change from 1x to 2x to 3x speed—interrupting flow every few seconds—what if operators could just press 1, 2, or 3?
"Want to try something?" I asked Meghan.
She was skeptical. "Management's tried to 'improve' things before."
But she agreed to test it. I sat next to her during her shift and watched her process clips without touching her mouse. 1 for normal speed when audio was unclear. 2 for standard processing. 3 for rapid-fire during peak periods.
After an hour: "This is... smooth."
No mouse hunting. No clicking and typing. Just seamless flow between speeds as her brain processed emergency content.
Her color indicator stayed green the entire shift.
Building Trust, One Fix at a Time
Word spread quietly. Other operators wanted to try the keyboard shortcuts. Then they started mentioning other friction points:
"The memory leaks are still killing us after a few hours..."
"Audio fails to load sometimes and we don't know if we should wait or skip..."
"The waveform visualization would help us scrub to important parts faster..."
Each fix was small. Memory management so browsers didn't crash. Predictive buffering so audio loaded reliably. Clear error states so operators knew exactly what was happening. Keyboard left and right for skipping a second backward or forward to replay.
But together, they transformed something fundamental: these professionals could finally start to actually trust their tools.
The Day Meghan Didn't Sigh
Three weeks later, I shadowed Meghan again.
8 hours. Police shootings in Brooklyn. Medical emergencies in Manhattan. Structure fires in Queens. The usual superhuman cognitive load at 3x speed.
But something was different.
No browser crashes. No memory leaks. No mouse-hunting for speed controls. No frustrated restarts.
At the end of her shift, Meghan looked surprised. "I didn't restart my browser once today."
She said it like she was noticing for the first time that her computer hadn't failed her. Like reliability was a novelty.
That's when I understood what we'd really accomplished. We hadn't just fixed technical problems. We'd given these incredible professionals permission to do their jobs again.
What They Deserved All Along
The transformation wasn't dramatic. No applause, no celebration. Just professionals doing impossible work, finally supported by technology that worked as hard as they did.
Meghan processing hundreds of clips in an 8-hour shift without interruption.
Sarah switching playback speeds seamlessly during a major incident, staying in flow state throughout.
The color indicators reflecting actual processing pace, not tool failures.
Operators who'd been labeled "behind" suddenly performing at their natural, superhuman level.
The saddest part? They'd been capable of this excellence all along. They'd just been fighting broken tools while doing work that should have been impossible in the first place.
And it was the start of a really amazing collaboration.
The Bigger Picture
This audio processor became part of the complete emergency response suite I rebuilt at Citizen:
- Emergency incident command center - Real-time crisis coordination
- Video monitoring system - Live incident verification
Together, these tools supported 50+ operators processing emergency content 24/7, enabling accurate, timely alerts for nearly 500,000 users across NYC.
But the real victory was smaller and more human: professionals doing extraordinary work, finally trusted and supported by the technology they deserved all along.
Some jobs are impossible. But when people figure out how to do the impossible anyway, the least we can do is build tools that don't make it harder.
