How a 160-Person Tech Company Switched to AI Headshots (Case Study)

This article is part of our Team & Company Headshots collection.

A mid-size cloud software company (we'll call them Client A) had 160+ employees across three offices and a growing problem: their team photos were a mess.

Some people had studio headshots from two years ago. Others had phone selfies. A bunch of new hires had no photo at all. Every time they put together a client deck or updated the website, someone had to chase people down for photos — and the results were never consistent.

They decided to try AI-generated headshots for the entire company. Here's what happened over eight months — the good, the messy, and what they'd do differently next time.

What was the actual problem?

Client A needed employee headshots for:

• Their internal directory and HR systems
• Client presentations and proposals
• The company website and marketing materials
• LinkedIn and professional networking

The old approach was scheduling a photographer to come in every 6-12 months. But with three office locations and constant hiring, there were always people who missed the shoot. Remote employees basically never got professional photos. And every round produced slightly different results — different lighting, different backgrounds, different quality.

By the time they'd finish one cycle of photos, the earliest ones already looked dated compared to the latest batch. It was a never-ending chase.

Why did they choose AI headshots?

The leadership team looked at three options:

1. Keep doing traditional photoshoots — More of the same headaches, just on repeat
2. Hire a photographer to visit all three offices — Expensive and still doesn't solve the consistency problem or cover remote employees
3. Use an AI headshot tool — Everyone uploads a photo on their own time, AI generates consistent results

They went with option 3 after evaluating several vendors and choosing BetterPic for its image quality, API capabilities, data security, and user-friendly interface.

How did the rollout work?

The project ran from January to August 2024:

January — Kickoff. HR, IT, and Marketing aligned on goals: cut costs, save time, and get every employee a matching headshot.

February — Vendor selection and setup. They picked BetterPic and started integrating it with their HR system via API. IT handled the technical side — setting up secure connections, mapping employee data fields, and configuring access controls.

March-April — Testing. They ran a pilot with two volunteer departments. This is where they caught early issues: some employees used low-quality source photos, and the initial API integration had a few compatibility hiccups that needed fixing.

May — Company-wide rollout. After fixing the pilot issues, they opened it up to everyone. HR sent instructions, ran a couple of live demos, and offered one-on-one help for anyone who needed it.

August — Review. They looked at the numbers and gathered employee feedback.

What pushback did they get?

This is the part nobody talks about in these case studies, but it's the most useful part.

Privacy concerns were the biggest issue. People were uncomfortable uploading photos of their face to an AI service. Fair enough — it's a reasonable concern. Client A addressed this by:

• Being transparent about exactly where photos were stored and who had access
• Confirming BetterPic's GDPR compliance and data encryption (end-to-end)
• Setting up a consent management system so employees could control their own data
• Running security audits to verify everything was locked down

Some people just didn't trust AI photos. About 15% of employees had reservations about the "realness" of AI-generated images. They worried the photos wouldn't look like them, or that something would feel off.

Long-time employees preferred the traditional way. A few people who'd been at the company for years liked having a photographer come in. It felt more personal to them.

The solution? Client A gave people a choice — participate in the AI headshot program or provide a traditional photo that met the brand guidelines. Most people who were skeptical actually came around after seeing how good the results looked on their colleagues.

What technical problems came up?

A few things went sideways during implementation:

Bad source photos. When employees uploaded dark, blurry, or heavily filtered selfies, the AI output wasn't great. The fix was simple — HR created a one-page guide: "Face a window, shoulders up, solid background, no sunglasses." Problem solved for 90% of cases.

API integration quirks. The initial connection between BetterPic and Client A's HR software had some compatibility issues. It took a few rounds of back-and-forth between the IT team and BetterPic's support to get everything running smoothly.

System slowdowns during rollout. When 100+ people tried to upload photos in the same week, things got sluggish. The IT team upgraded their server capacity and staggered the rollout by department, which fixed it.

Legacy data headaches. Migrating old employee photos into the new system and mapping them correctly took more time than expected. Custom scripts had to be written to handle edge cases.

What were the actual results?

Here's what the numbers looked like after three months:

Time savings:

• New employee onboarding (photo portion) went from a multi-week process to same-day — a 75% reduction
• Real-time updates when someone changed roles or appearance — no more waiting for the next scheduled shoot

Cost savings:

• 60% reduction in annual photography costs
• No more expenses for company-wide photoshoots
• Less admin time spent chasing people for photos

Consistency:

• Every employee now has a headshot with the same style, background, and quality level
• Client decks, the website, and marketing materials all have a unified look
• New hires get their headshot during their first week, matching everyone else from day one

Employee satisfaction:

• 70% reported being happy with the convenience
• Most of the initial skeptics came around once they saw the results
• The 15% who still preferred traditional photos were accommodated without drama

What did they learn?

Five lessons from Client A that are worth knowing if you're considering this for your own company:

1. Start with volunteers, not mandates. Rolling out to enthusiastic departments first created internal champions. When skeptics saw their colleagues' results, most resistance melted away on its own.

2. Invest in the source photo guide. The single most impactful thing they did was create clear, simple instructions for how to take a good upload photo. Bad inputs = bad outputs, and no amount of AI can fix a dark, blurry selfie.

3. Don't force it on everyone. Giving employees the option to submit a traditional photo (as long as it met brand guidelines) reduced pushback dramatically. In practice, very few people actually took that option — but having it available made people feel respected.

4. Front-load the privacy conversation. Don't wait for people to raise concerns. Address data security proactively — where photos are stored, who has access, how long they're kept, and how to opt out. Transparency kills anxiety.

5. Budget time for the technical integration. Even with a clean API, connecting a new tool to existing HR systems takes longer than you think. Build in buffer time and plan for a few rounds of troubleshooting.

Would they do it again?

Without hesitation. Client A's take is that the eight-month project paid for itself within the first quarter through photography cost savings alone — and the ongoing benefits (instant onboarding photos, consistent branding, zero scheduling headaches) compound over time.

They're now looking at expanding the integration to include project management tools and internal social platforms, and exploring more style options for different use cases (casual for internal comms, formal for client-facing materials).

For any mid-size company dealing with the same "team photos are a mess" problem — which is basically every mid-size company — this case study is proof that AI headshots work in practice, not just in theory. You just have to plan the rollout right.