Let’s continue our journey off the shiny AI hype machine, and onto the less exciting, but more powerful pragmatic and iterative AI implementation process…. machine. At SXSW 2025, the focus had clearly shifted from mind-blowing AI demos to real-world implementation stories. And through those implementations, we are learning how they impact company structure, team collaboration, and push us to consider openness, ethics, and innovation.
This post gathers insights from several key sessions that explored the deeper impact of AI:
- “AI-powered organizations: Redefining the future of work” by Melissa Valentine, Stanford University
- “AI and the future of consumer experiences” featuring Cherae Robinson (Flybridge), Jesse Middleton (Flybridge), and Anastasia Crew (Notion for Startups)
- “Open source AI: Legal issues at the frontier” with representatives from Meta, Databricks, EleutherAI, and Morrison Foerster
- “Recognition revolution: Blending tech and human touch” featuring Dr. Meisha-Ann Martin, VP People Research at Workhuman
For full transparency and to get meta with it, I created an outline of this blog post by loading all of my AI session notes into ChatGPT4o and working through some prompts to land on the overarching themes and sections for two recap posts. I’m using AI to enhance my content creation abilities and speed the outlining and drafting process, not replace me as a writer. The session speakers would be so proud.
AI as a teammate, not just a tool
In her session on AI-powered organizations, Stanford researcher Melissa Valentine made a compelling case: the real revolution in AI is not individual productivity, but organizational intelligence.
She described organizations as systems built on specialized yet coordinated workflows. AI, when integrated properly, can enhance these systems in three ways:
- Information gathering (input)
- Evidence-based strategy generation (analysis)
- Personalized outputs (action)
One real-world example came from Stitch Fix, which uses AI to create a “digital twin” of its warehouse and customer base. This twin helps with inventory planning, staffing decisions, and personalized recommendations at scale. The goal is not to replace humans, but to help them make smarter, faster decisions with more context.
AI can also assist with team assembly and problem-solving. Imagine an AI system that recommends the best people in your organization to tackle a new challenge, based on skills, availability, and past success.
For companies working on platforms, marketplaces, and business systems, this shift from tool to teammate is important. The next generation of product features won’t just automate tasks, they’ll coordinate effort!
AI that helps people lead better, not do less
Another standout example came from Workhuman in the session “Recognition revolution: Blending tech and human touch.” Dr. Meisha-Ann Martin described their AI chatbot as an “angel on your shoulder” designed to help managers give better feedback, not to replace them.
This assistant helps with questions like:
- What is this person doing really well?
- What skills do they need to reach their next milestone?
- How can I phrase feedback to avoid bias or microaggressions?
It draws on Workhuman’s internal data and Gallup research to provide personalized, real-time coaching for leaders. The chatbot doesn’t take over the act of recognition. It supports it with context and guidance.
This is a strong example of what responsible, human-centered AI can look like. It respects the emotional and social complexity of leadership while offering tools that improve the quality of feedback given. AI in people operations was a recurring theme of the conference, and Workhuman seems like a tool to watch in this space.
Storytelling and scalability in consumer AI
In another SXSW panel, Cherae Robinson shared her exciting journey, from founder of Tastemakers Africa to her role now as an investor at Flybridge. One of my key take-aways from Cherae was that consumers don’t just buy products, but stories, lifestyles, and aspirations.
She recommended AI in this context as a kind of creative partner. It helps brands of all sizes punch above their weight by repurposing content, scaling communication, and offering personalized experiences. She noted that she could now upload 50 Instagram captions and have Claude write the next thousand, keeping the same unique voice and tone.
Notion’s Head of Startups, Anastasia Crew, pointed out that founders often don’t realize how much structured data they already have. Platforms like Notion are sitting on rich context that could power AI-driven features like itinerary builders, onboarding guides, or content assistants. The opportunity is to connect creativity with the systems that support it.
For consumer companies and founders, the takeaway is that your data and your voice are your greatest assets. AI is nowhere near ready to replace that, but it can amplify it.
Open source, closed doors, and the future of AI governance
One of the most thought-provoking sessions at SXSW was a panel on the legal and ethical considerations of open source AI. Representatives from Meta, Databricks, EleutherAI, and Morrison Foerster debated what openness really means, how to balance safety with innovation, and what guardrails should be in place.
Here are a few takeaways:
- There is no one-size-fits-all approach to open source licensing for AI models. Meta’s models, for example, have tighter usage terms for tools like Ollama, intended to prevent misuse.
- EleutherAI, on the other hand, advocates for full openness and transparency, arguing that open models empower safer and more meaningful experimentation.
- Databricks emphasized that the real asset is not the model itself, but the data. They want their customers to build on top of shared models with confidence, supported by observability tools and safety protocols.
The panel also discussed the importance of red teaming, acceptable use policies, and safety frameworks like Meta’s internal risk assessments. These tools help determine whether a model poses a step change in risk and whether it should be released at all.
The spicy question of the day was: where is the line between safety and censorship? As AI becomes more powerful, this tension will grow. For now, developers and product teams must decide which values to prioritize and how much control to retain.
Building toward an agent-powered future
Several sessions echoed the same prediction: we are heading toward an agent-to-agent economy. Imagine a world where your personal assistant AI agent talks to your travel AI agent, your business planning AI agent, your sales forecasting AI agent - and they all coordinate to act on your behalf.
While this is more future-thinking and requires significant human oversight today, foundational examples are emerging:
- ERP.ai, which helps businesses plan for growth by anticipating needs across inventory, staffing, and logistics
- Creative workflows that use Runway, Midjourney, and Replit to build full marketing campaigns in a day
- Open source agent platforms that automate end-to-end tasks with no-code inputs
For product leaders, this means thinking not just about what your app does for the user, but what it can coordinate on the user’s behalf.
Open, human, integrated
At SXSW 2025, the biggest ideas weren’t about the latest AI models. The hot topics revolved around how to design better systems. How can AI enhance human systems, organizational systems, and open source systems?
The best practices that surfaced were:
- Treat AI as a partner, not just a feature
- Lean into your data and your brand’s unique voice
- Make deliberate decisions about openness, safety, and governance
- Focus on coordination and collaboration, not just output
- Support humans in doing their best work, not replace them
This kind of thinking is what will differentiate great products in the next wave of AI. We’d love to work with you on your AI initiative - let’s chat!
