Most of us have experimented with ChatGPT for basic tasks, but moving beyond obvious use cases to truly strategic applications requires more sophisticated thinking. In our latest Women @ Tinder event, we brought together experts who've moved past the "AI 101" phase to share advanced insights and honest perspectives on implementation challenges.
Alongside my fellow Women at Tinder ERG lead Kenny Imery Bonilla, I was thrilled to host a panel of expert speakers from Open AI and TinderⓇ, including:
- Hillary Paine, VP of Product at Tinder, who drew from her extensive experience as a product leader at Tinder, Zillow, and other platforms to share how AI helps leaders scale themselves by offloading some of the things that they currently have on their plate.
- Jeannie Andrews, Digital Native Sales Leader at OpenAI, who offered a unique perspective from the forefront of AI innovation, providing practical examples such as using AI for complex international translation and transcription, along with nuanced technical details about how various ChatGPT models work.
As a domain expert in data, machine learning, and artificial intelligence, I was curious to hear from these industry experts about all the ways they’ve evolved their AI usage to truly strategic applications that impact their businesses.
Here are key takeaways from our conversation about AI's role in scaling product leadership.
Shifting from tool to thought partner
Many teams started using AI as a more sophisticated search engine or writing assistant. But Hillary and Jeannie both emphasized something organizations are just beginning to discover: AI's real value emerges when you treat it as a strategic collaborator rather than just another productivity tool.
"I view ChatGPT as a really great thought partner and a teammate," Hillary says. Instead of one-off requests, she treats interactions conversationally, as if sitting in a room whiteboarding with a colleague. These models are trained on global data, which makes them incredibly diverse and gives you access to perspectives that even human teams might miss.
This mindset shift—from efficient tool to strategic partner—separates teams seeing incremental improvements from those achieving transformational results. It's a distinction many of us are still learning to navigate.
Advanced applications beyond the obvious
While we've all heard about AI helping with content creation, our panelists shared several sophisticated use cases that many teams haven't explored yet.
The conversation began with a few wins in their personal lives, like using ChatGPT to create a 3-day itinerary for an 8-month-old during a stretch of solo parenting or wedding vows that became so much more thoughtful and personalized with added context. But the professional applications showed where experienced users are finding unexpected value.
Jeannie uses AI for complex translation work and transcription work, particularly when generating next steps and action items for international customers. This approach effectively opened doors across borders—the kind of strategic application that goes far beyond basic translation tools most of us might reach for.
For leaders who've mastered basic AI tasks, Hillary says AI helps you scale yourself by offloading some of the things that you currently have on your plate. The goal isn't just efficiency—it's creating mental space to be more human, strategic leaders.
Refining your prompting approach
If you've been using AI for a while, you've probably noticed that getting consistently good outputs requires more finesse than the basic tutorials suggest. You’ve also probably noticed that even the best AI models can’t spit out great outputs after just a prompt or two.
Jeannie says that adding more context to her prompts or attaching an example of the output you’re looking for helps her get to the end goal faster. She adds that she treats AI not as a robot, but like a highly skilled intern who just needs a little extra direction. This approach helps her calibrate expectations, and the interactive back-and-forth with AI is far more effective in her day-to-day life.
Hillary takes a different but equally sophisticated approach—she prefers starting with a question and then continuing to refine it rather than trying to perfect a single prompt. This conversational refinement process, like asking for less realistic results or less professional tones, reflects the kind of iterative thinking that got most professionals to where they are today.
The reality is that even with experience, AI responds better to clarity and iteration than perfection on the first try—something many of us are still learning to embrace.
Navigating sophisticated implementation challenges
Even at a tech-forward company like Tinder, AI implementation presents nuanced challenges that go beyond basic adoption hurdles.
When developing Tinder's “The Game Game”TM feature, the team encountered something many of us working on AI-powered features will recognize: AI personas that initially surfaced as too direct or confrontational. Their solution required layering in EQ models and different psychology frameworks to achieve a campy vibe that was neither too real nor too fake, and clearly designed as a training run for IRL conversations.
During the Q&A session, a more universal implementation concern emerged around the risk of over-reliance on ChatGPT. This begged an almost existential question: are we at risk of losing ownership over our work when two AIs are just talking amongst themselves?
According to Jeannie, this is exactly why it’s critical to use AI as a thought partner and not a replacement mechanism for doing good work. This requires managers to hold their teams accountable for responsible AI use—treating AI as an amplifier of human judgment, not a substitute for it. It's a balance many of us are still figuring out.
Moving beyond individual tasks to workflow integration
For teams ready to advance their AI implementation, the panel offered a path forward that many of us haven't fully explored yet. Rather than just using AI for isolated tasks, they're seeing success with more integrated approaches.
Our panelists shared several examples, which made it difficult to pick just a few to explore in this recap. Hillary unpacked how she created a function where all project briefs must be started through prompting and ChatGPT, while Jeannie discussed her approach to systematic experimentation with anything that you hate doing.
As exciting as it might sound to automate away the work you dislike, both of our panelists emphasized that building AI use responsibly requires diverse perspectives. Jeannie says that OpenAI focuses on ensuring teams are diverse and have diverse perspectives because this means that we're going to build and deploy better technology.
Hillary urged anyone using ChatGPT to actively consider cultural perspectives and to ask what they might be missing when they develop their ideas. When done correctly, Hillary says that this gives us the ability to use the models’ diverse training data to spot blind spots we would have otherwise missed.
The emerging frontier: AI agents
Looking ahead, the discussion revealed where AI implementation is heading next: agents that can complete a task for you end-to-end. For example, Jeannie’s recent hackathon project was an agent that ingests details about all product launches from a Slack channel and then builds responses to each of her customers.
For those of us who've moved past basic AI adoption, this event offered refreshingly honest perspectives on the sophisticated challenges and opportunities many teams are navigating. Even as AI becomes more accessible, thoughtful implementation grounded in diverse perspectives and human judgment remains the key to transformational results.