One Agency, 15 Screens: How We Shot a Client Team We Never Met in Person

The short version: Anchovies, a branding agency, hired us to photograph 15 people for a client of theirs, a team scattered across different cities and countries. No studio, no travel, no generated faces. Every session was shot by one photographer, Elina, live: the person joins a video call for the direction, but the capture itself runs through our own app, which shoots RAW straight off the phone's camera. She set up each camera with the person, directed them the way she would in a studio, and shot remotely. Our retouch team then finished every file mostly by hand. The result went onto a client-facing website an agency had put its own name behind, and that is the highest quality bar we know.

Here is the part that surprises people about my job on a project like this: it was over in the first week. I set the project up, and then Elina ran it. On an office shoot I stand in the room all day with a call sheet. Here, once the spec was agreed and the booking links went out, the machine did not need me, and it did not need the client either.

Key takeaways

  • A remote headshot here is a photograph, not a generation. Every delivered pixel started as a RAW frame captured live from the person's own phone, directed by a real photographer.
  • The direction happens over video; the capture happens in our own app. The app shoots RAW from the phone's camera, so the files are camera originals, not compressed video frames.
  • A session takes about ten minutes per person. The person needs a phone, a window, our app and nothing else.
  • AI is pointed at the camera, never at the person. Perspective, lens distortion, light and color are corrected algorithmically; everything that touches a face is done by hand.
  • Fifteen rooms, one photographer, one standard. The same photographer shot every session, to one spec, through one retouch pipeline. That is why the set looks like it was shot in one place.
  • The client did nothing after the first email. No coordination, no chasing, no scheduling on their side.

In this guide

  • The project at a glance
  • Why did a branding agency hire us for someone else's team?
  • Can a phone camera really produce a professional headshot?
  • How did the sessions actually run?
  • What does "retouched by hand" actually mean?
  • How do fifteen photos from fifteen rooms end up matching?
  • Why not just use an AI generator?
  • What did the agency actually have to do?
  • When is remote not the right tool?
  • Frequently asked questions
  • The bottom line

The project at a glance

  • 15 people photographed, each in their own city, several time zones between them

  • One photographer for the whole set: Elina, our remote photographer, shot every session herself

  • Ordered by an agency for their client: Anchovies ran the brand, we ran the photography

  • About 10 minutes per session, directed live over video

  • Captured in RAW through our own app on each person's phone, with the light the room already had

  • Retouched mostly by hand, with AI correction used only to undo what the phone and the room got wrong

  • Nothing generated. No pixel in the delivery was synthesized, and no face was altered

  • Zero coordination on the client side after the initial list of names

The figures on this page come from our production records for this project. The end client is not named at their agency's request. Employee faces appear with permission.

Why did a branding agency hire us for someone else's team?

Because their name was on the deliverable.

Anchovies was building their client's brand and website. The team page needed fifteen faces that looked like one company, and the fifteen people making that company sat in different cities. The usual answers all failed at once: flying a photographer to every city costs more than the website, hiring a local photographer in each city produces fifteen different-looking sets, and AI generation was ruled out immediately, because an agency selling creative work cannot put synthetic faces on a client's public site and hope nobody ever finds out.

That last sentence is not us editorializing. It is the reason we get calls from agencies at all. An agency vouches for everything it ships, so the photography vendor behind it has to be someone whose output survives scrutiny: from the client, from the client's employees, and from anyone who zooms in.

So the brief to us was simple and strict: fifteen matching, client-facing headshots, no travel budget, no studio visits, and quality that would sit comfortably in a brand book. That is our remote headshots service, and this is what it delivered.

Can a phone camera really produce a professional headshot?

Yes, and the phone was never the hard part.

Modern phone cameras closed most of the gap with dedicated cameras years ago; DXOMARK, which tests both on the same protocols, puts flagship phones at a level where the difference is hard to tell in good conditions. The sensor stopped being the bottleneck. What actually separates a selfie from a professional headshot is everything around the sensor: where the camera stands, where the light comes from, what the person does with their face and shoulders, and who decides which frame is the one.

Which is to say: the bottleneck is direction, and direction is exactly what a live session supplies. Left alone with the same phone, the same person produces the photo you have seen on a hundred team pages, shot from below, lit from the ceiling, smiling like a hostage. Ten minutes of a photographer saying "higher, toward the window, chin down a touch, now look just past me" produces the photo in the gallery above. Same phone. Same room.

How did the sessions actually run?

Each person got a calendar slot, opened our capture app on their phone, and joined a video call for the talking part. From there it worked like a studio session that happens to be very small and very far away:

  • Minutes one to three, the setup. Elina looks at the room through the camera and picks the spot, almost always a window. The phone gets propped, leaned or held by a books-and-tape rig she talks the person through. Angle and height are set to the frame our spec calls for.

  • Minutes three to eight, the shoot. She directs exactly as she would in the studio: posture, shoulders, chin, where the eyes go, when to breathe out. The shooting itself runs through our app, which captures RAW files from the phone's camera while the conversation continues over video.

  • Minutes eight to ten, the check. She reviews frames on the spot, reshoots what needs reshooting, and confirms the session is done before the person leaves the call.

The app is the part of this that does not get enough credit. Services that run a session through a browser link capture, at best, what a video stream can carry, and a video stream is built to survive bad Wi-Fi, not to hold detail. Our app records camera originals: RAW files with everything the sensor saw. The person talks to Elina over video; the photographs never travel through the video call at all. That one architectural decision is most of the reason the finals survive retouching, printing and a designer zooming to 100%.

Nobody needed equipment. Nobody was asked to buy a ring light. The brief to each of the fifteen was one line long: a phone, a window, our app, ten minutes.

And here is the honest division of labor. I produced the front of the project: the roster from the agency, the spec, the booking links. After that, Elina ran it end to end, sessions, reschedules and retakes included, working through the list across time zones at whatever hours suited the fifteen people in it. On an office shoot a producer stands in the room all day because seventy people share one physical line. Fifteen people on fifteen separate calls do not need that person, and we do not bill for one standing around. If a session needs a retake, that is a ten-minute calendar slot, not a reshoot day. For the step-by-step version of the process itself, we wrote up how a live-directed session works.

What does "retouched by hand" actually mean?

It means the AI is aimed at the camera, and the hands are reserved for the person.

Let me be precise, because "no AI" would be a lie and "AI headshots" would be a worse one. A phone does predictable things to a face: wide-lens perspective stretches whatever is closest, distortion bends the edges of the frame, and computational color makes its own decisions about skin. We built our own AI correction pipeline for exactly that job and nothing else: perspective, lens distortion, light and color, undoing what the phone and the room did. It runs first, it works because RAW files give it real latitude, and it never touches features. What we will not do is feed a mediocre photo to a generator and let it invent a better-looking human.

Everything after that is a person with a tablet and a deadline. Skin is retouched the way we retouch studio files: temporary blemishes out, texture kept, pores visible at full zoom. Stray hairs, shine, a collar that sat wrong. No reshaped faces, no resculpted bodies, no synthesis. The person in the photo is exactly the person who joins the meeting, minus a bad night's sleep.

This split, AI for the camera's mistakes and hands for the human, is the part of our process that took the longest to build, and it is the reason the finals do not read as phone photos. It is also the opposite of how most of the category works, where AI does the finishing and the face drifts a little with every pass.

How do fifteen photos from fifteen rooms end up matching?

This was the agency's real fear, and it is the right fear. Fifteen decent headshots that do not match are useless to a brand book; buyers describe this exact failure as photos that look fine individually and wrong side by side.

The answer is boring and structural, which is why it works. One photographer: Elina shot all fifteen sessions herself, so the eye behind every frame was the same eye. One spec, written before the first session: framing, crop, the direction light should come from, the expression register. One art director, Alex Karnyukhin, who owns our lighting and retouching standard and reviewed this set the same way he reviews studio output. One retouch pipeline, so every file passed through the same hands and the same color decisions. The person shot on a Tuesday in one country matches the person shot on a Friday in another, because nothing about the standard was left to the room they happened to sit in.

Compare that with the obvious alternative, a local photographer in each city: fifteen rooms and five photographers is five different opinions about light. Fifteen rooms and one Elina is one opinion, fifteen times.

We keep that spec on file afterward. When the client hires three people next quarter, those three sessions produce files that drop into the same grid without anyone flying anywhere. Consistency across time is the same problem as consistency across a set, and it is solved the same way: by the spec, not by luck.

Why not just use an AI generator?

Honest answer: for some people, that is a fine choice, and the tools are getting better. Anchovies ruled it out for a different reason than quality, and it is worth understanding.

The research on AI faces is uncomfortable in both directions. People are bad at spotting them: a study in PNAS found participants performed below chance at telling synthetic faces from real ones. So the risk is not that visitors will squint at a website and catch it. The risk is the reveal. In a survey of 1,087 recruiters, 76.5% actually preferred AI headshots when they did not know what they were looking at, and 66% were put off the moment they learned a headshot was AI; 88% wanted disclosure. And on a team page the reveal is not a risk, it is a certainty, because the fifteen employees know what their own faces look like, and so do their colleagues, clients and families.

An agency cannot carry that exposure on a client deliverable, so the decision made itself. The same survey has a finding we quote more often, though: 67.5% of recruiters were also put off by bad real headshots. Real is not enough either. The bar is a real photograph, professionally made, which is the entire argument for putting a photographer on the call. Our longer comparison lives in AI headshots vs real photography.

What did the agency actually have to do?

Send a list of fifteen names and email addresses, and answer one question about the brand's visual direction. That was the entire engagement on their side. After that first email, neither Anchovies nor their client touched the coordination at all.

We turned the list into booking links, the booking links into sessions, and the sessions into a proofing gallery. The agency's designer got files cut to the site's specs. Nobody at Anchovies chased an employee across time zones, rescheduled a no-show or explained to anyone how to prop a phone against a stack of books, because Elina handled all of it directly with the fifteen people themselves. Agencies bill their hours to clients, so the fewer of those hours our process eats, the better we look on their books.

The saved spec is part of the deal: next quarter's new hire books a ten-minute slot and lands in the same grid.

When is remote not the right tool?

Credibility requires saying this part out loud.

If your whole team sits in one building, an on-site day is better: more frames per person, more setups, and the group photos that remote cannot do. That is a different production, and we documented one at a 70-person office shoot. If you need sculpted executive portraiture for an annual report, with controlled studio light and multiple setups per person, that is executive portrait territory, and a phone by a window will not get you there. And some teams simply want the shared experience of a studio day, which no video call replaces.

Remote wins when the team cannot be in one room: distributed companies, international hires, an agency's client three time zones away. That was exactly this project.

Frequently asked questions

Is this good enough for a client-facing website? That was the acceptance test of this exact project: an agency put the output on a site carrying its own reputation. The before/after pairs above are the evidence. Zoom in; the texture is the tell that it is a photograph.

What does each person need? A phone made in the last few years, a window, our capture app, and about ten minutes. No ring lights, no tripods, no other equipment.

Why shoot through an app instead of just a video call? Because a video stream is compressed to survive bad Wi-Fi, and photographs deserve better. Our app captures RAW files directly from the phone's camera while the direction happens over video, so the retouch team works on camera originals with full detail and color depth. Browser-link sessions work with whatever the stream delivers; we never do.

What if someone's room has terrible light? That is the first three minutes of every session. The photographer looks at the room through the camera and finds the light, which almost always means repositioning the person relative to a window. In fifteen sessions there was no room Elina could not solve.

Can you match headshots you shoot later to this set? Yes, that is what the saved spec is for. New hires book a slot, and their files pass through the same pipeline to the same standard. The grid grows without drifting.

Who actually holds the camera? Nobody, usually. The phone is propped or leaned at the angle the photographer sets, because a held phone moves. A colleague or partner can hold it if the room demands it, but the photographer is watching the frame live either way.

How is this different from AI headshot tools? AI tools synthesize a face from your uploads; we photograph the face you have. Every pixel we deliver started as a RAW frame captured during a live session, and the only AI in our pipeline corrects the phone's optics and color. It never generates, reshapes or "improves" a person.

The bottom line

My definition of a good remote project is the same as for a good office shoot: the client does almost nothing, the people in front of the camera are done in minutes, and the finals go up without a single "can we fix this one" email. Fifteen people, several countries, ten minutes each, one photographer, no travel, nothing generated, and a set an agency was comfortable signing its name under.

If you are an agency with a client team to photograph, or a distributed company staring at fifteen mismatched LinkedIn crops on your own site, talk to us. I will set the project up, a photographer like Elina will find the window in every room, and your part of the whole thing will be one email.


"We sold our client a brand, and the team page had to live up to it. Match ran the whole thing off one email from us, and the set came back looking like it was shot in one studio on one afternoon. Nobody believes it was fifteen phones." — Anchovies


About Match Production

Match Production is a New York City studio specializing in corporate, team and executive headshots, shot in our Midtown studio, on location at your office, and remotely with live direction. Our remote photographers cover the US, Europe and Asia-Pacific, so a distributed team books sessions in its own working hours. Elina, our remote photographer for Europe and the US, shot the set described above; Irina Chistikina, our Lead Producer, set the project up. Published with permission from Anchovies.

Irina Chistikina, Lead Producer at Match

Irina coordinates every project we run: the room, the schedule, the sign-up sheet, the order your people move through in the time you actually have, and the gallery afterwards. If the day runs smoothly, that was her. She is also who your quote comes from, so she is the one to email.

Next
Next

Tips for Staff Headshots: How to Prepare Your Team and Get Consistent Results