PhotoShelter's New AI System 'Celine' Shrinks Weeks Win/Loss Research Down to Hours for its GTM Teams

For go-to-market teams, AI is quickly evolving from a basic administrative helper into a heavy-lifting strategic engine, from helping with simple time-saving tasks such as email management toward unlocking brand new possibilities that didn't exist before, including analyzing win/loss data at lightning speed and shaping market expansion plans. At PhotoShelter, that evolution came in the form of turning something that once would take weeks of grueling desk research into a single overnight sprint.

Christina Kyriazi, Chief Marketing Officer at PhotoShelter, has seen firsthand how her team connected AI directly into their sales data to compress their strategic timelines. Rather than treating the software as a simple productivity hack, she uses it to run a deep win/loss program and spin up detailed go-to-market playbooks in hours instead of weeks, something that would once have been impossible.

"We went from using AI to save time typing up emails to focusing on going to market faster. We can be much more efficient about acquiring and growing our customers, and that velocity can have a real impact on revenue," says Kyriazi. To support that approach, PhotoShelter deploys a custom AI go-to-market system named Celine, which they connect to their internal tech stack.

• Enter Celine: Celine processes customer and prospect calls to run deep win/loss analyses. The setup helps solve a fundamental issue of human bandwidth. Sales teams simply move too fast to manually categorize messy, qualitative context at scale accurately. By handing this over to Celine, the team extracts deeper, more accurate insights without sacrificing data quality. "Celine will listen and read all the communications and extract insights that our reps could never compose effectively in a simple dropdown menu. A sales rep is moving a million miles an hour in a day. They might select whatever on the dropdown menu, but Celine is not going to lie to you," says Kyriazi.

• Signal over static: The win/loss program combines quantitative scoring with qualitative analysis. On the quantitative side, the AI evaluates 50 different factors on a scale of one to ten, measuring how strongly each variable actually ties to winning or losing a deal. "We can rank based on not only the volume of mentions, but also their impact on deals. The primary thing it helps us do is hyper-focus on the top two or three variables that we can address to have an impact on the metrics that matter." A budget objection that clearly influences the outcome might score a ten, while a factor that barely appears in the interaction might score a one. That allows the team to see not just what gets mentioned, but what actually moves the needle on revenue.

• Playing the brain game: Kyriazi notes that asking the AI to explain the difference between won and lost deals is a massive analytical lift. "From there, I can adjust my go-to-market from the very beginning to target the exact types of leads that are converting best." Those qualitative insights feed directly into how Kyriazi targets the top of the funnel. Instead of guessing which acquisition sources might perform best, the team can see exactly which types of deals tend to convert. The AI surfaces behavioral patterns that would be nearly impossible to piece together manually across hundreds of conversations, a tactic other teams use to understand why customers stay, expand, or churn.

But people still have to actually use the tools. As some industry reports highlight, the AI scaling gap in customer experience frequently stalls on human habit rather than technical capability. At PhotoShelter, adoption started organically with a single product marketer experimenting with Celine before spreading to the rest of the group. Kyriazi focused on getting the interface into people’s hands and teaching basic prompting so they could target specific deal types. Over time, she worked to build a new reflex into the team’s workflow.

• The mental shift: "As a leader, whenever we are going through projects or getting stuck, I always have to remember to say, 'Wait, did you ask Celine?' That has been the biggest mind shift. It is about remembering to stop and actually ask that question to either myself or the team." For Kyriazi’s team, building that muscle memory led to massive timeline compression.

• Overnight overhaul: She points to a recent example planning an expansion into new market segments. Traditionally, building a comprehensive strategy required weeks of grunt work to understand market gaps, evaluate competitive positioning, and define target accounts. By using AI as a strategic thought partner, she turned that research into an overnight exercise without disclosing any proprietary details about the segment itself. "I took a couple of hours chatting with Claude and gave it the context to create a template for a go-to-market playbook. At the end of the night, I got a 37-page, detailed, personalized playbook that I literally rolled out to my team overnight."

Kyriazi estimates that this approach has likely saved her team upwards of six months between analysis and execution. Instead of spending that time building the plan, they can now use it to test and refine the strategy in-market. Looking ahead, she expects the next maturity phase to be less about planning and more about production. Kyriazi observes that while early AI adoption heavily favored content generation, some marketing teams are now graduating to automating workflows and asset creation, a trend that aligns with signal-driven go-to-market experiments.

For companies leaning into AI-assisted planning, such velocity creates a new competitive baseline. The organizational bottleneck has moved away from researching the plan and now sits squarely on executing the plan in-market. Kyriazi notes that among forward-thinking leadership teams, there is often less patience for traditional execution timelines. It is a factual change in how fast modern business moves, and she is candid about the executive pressure that creates. "The pressure is crazy because leadership knows this technology is out there, and it can help us. You need to keep up with it," Kyriazi concludes. "It's almost like you're racing against the AI."