Managing AI-driven Team Interactions

Teams will increasingly include both humans and AI agents. We need to learn how best to configure them. A new Stanford University paper "ChatCollab: Exploring Collaboration Between Humans and AI Agents in Software Teams" reveals a range of useful insights. A few highlights: 💡 Human-AI Role Differentiation Fosters Collaboration. Assigning distinct roles to AI agents and humans in teams, such as CEO, Product Manager, and Developer, mirrors traditional team dynamics. This structure helps define responsibilities, ensures alignment with workflows, and allows humans to seamlessly integrate by adopting any role. This fosters a peer-like collaboration environment where humans can both guide and learn from AI agents. 🎯 Prompts Shape Team Interaction Styles. The configuration of AI agent prompts significantly influences collaboration dynamics. For example, emphasizing "asking for opinions" in prompts increased such interactions by 600%. This demonstrates that thoughtfully designed role-specific and behavioral prompts can fine-tune team dynamics, enabling targeted improvements in communication and decision-making efficiency. 🔄 Iterative Feedback Mechanisms Improve Team Performance. Human team members in roles such as clients or supervisors can provide real-time feedback to AI agents. This iterative process ensures agents refine their output, ask pertinent questions, and follow expected workflows. Such interaction not only improves project outcomes but also builds trust and adaptability in mixed teams. 🌟 Autonomy Balances Initiative and Dependence. ChatCollab’s AI agents exhibit autonomy by independently deciding when to act or wait based on their roles. For example, developers wait for PRDs before coding, avoiding redundant work. Ensuring that agents understand role-specific dependencies and workflows optimizes productivity while maintaining alignment with human expectations. 📊 Tailored Role Assignments Enhance Human Learning. Humans in teams can act as coaches, mentors, or peers to AI agents. This dynamic enables human participants to refine leadership and communication skills, while AI agents serve as practice partners or mentees. Configuring teams to simulate these dynamics provides dual benefits: skill development for humans and improved agent outputs through feedback. 🔍 Measurable Dynamics Enable Continuous Improvement. Collaboration analysis using frameworks like Bales’ Interaction Process reveals actionable patterns in human-AI interactions. For example, tracking increases in opinion-sharing and other key metrics allows iterative configuration and optimization of combined teams. 💬 Transparent Communication Channels Empower Humans. Using shared platforms like Slack for all human and AI interactions ensures transparency and inclusivity. Humans can easily observe agent reasoning and intervene when necessary, while agents remain responsive to human queries. Link to paper in comments.

Stop Treating AI Like a Tool, Start Onboarding It Like a Teammate! 🚀 Are you struggling to get real value from AI in your team? The problem might not be the technology, but how you're integrating it. Just like a new hire, AI needs clear roles, training, and ongoing feedback to truly thrive. : * Define clear responsibilities: What specific tasks will the AI handle? * Invest in "AI literacy": Everyone on the team needs to understand AI's capabilities and limitations. * Establish communication protocols: How will the AI share its insights and when will it need help? * Provide continuous training and feedback: Help the AI learn and improve, just like you would with any team member. * Foster collaboration and trust: Encourage teamwork between humans and AI. * Iterate and adapt: Be flexible and adjust your approach as the AI evolves. * Address ethical considerations: Be mindful of bias and ensure fairness. The key takeaway? Treat AI as a partner, not just a tool. Build a collaborative environment where AI can flourish, and you'll unlock its true potential.

The fastest-growing profession of this decade won't be creating AI, instead it will be: Managing the agents it spawns. Management has always evolved with technology: → Foremen directed the construction of buildings → Industrial supervisors organized factory production → Corporate managers optimized business operations → Agent Managers now orchestrate artificial intelligence This evolution marks a fundamental shift in how we organize work and create value. People who orchestrate workers are managers. People who orchestrate software are engineers. But what do we call those who orchestrate AI agents? While we figure out the terminology, this represents a new job category emerging from the advancement of AI. The distinction matters because: → Engineering builds systems with predictable outcomes → Management guides humans with emotions and incentives Agent management bridges these worlds, directing intelligence that scales like software but reasons unpredictably. What do agent managers actually do: → Provide strategic direction that AI still struggles with → Design frameworks for AI teams to operate within → Make high-level decisions about resource allocation → Create evaluation systems for quality and safety → Optimize collaboration across specialized agents This role will explode in demand because: → Enterprises are deploying specialized agent teams → Powerful AI will require more sophisticated oversight → AI is becoming a mission-critical business function → Orchestration becomes a competitive edge → Returns from effective AI management exceeds costs The most effective agent managers will: → Communicate with exceptional precision → Design robust feedback systems → Think systemically about agent interactions → Learn to anticipate how AI "thinks" differently → Balance innovation with appropriate guardrails This isn't just another tech job. We are entering an era where algorithms and data are table stakes. The true competitive edge lies in developing capabilities others can't easily replicate. Agent management is exactly this, the bridge between human strategy and AI execution that will define tomorrow's market leaders.

The most powerful use of AI at work won’t be solo. It will be shared. Ben Thompson recently wrote about a compelling use case: how he and his assistant collaborated with a single LLM chat. An example of a shared assistant for team coordination and synthesis. I’ve been thinking about this a lot too. At Dropbox, we’re building toward this future with Dash, our new AI workspace, and specifically with Stacks, a way for teams to organize, track, and reason across all the work happening in a project. Stacks are designed for collaborative intelligence. Teams can pull in docs, links, and tools from anywhere, ask questions about the work, and get AI-generated summaries that evolve as the project does. It’s a persistent shared memory that helps teams move faster, stay aligned, and reduce the drag of context loss. But coordination is just the first step. There are four basic configurations for how humans and LLMs might collaborate: 1. One person working with many agents. The classic orchestration model. Think of a PM using agents for research, writing, and planning. Most solo AI workflows live here today. 2. One agent working with many agents. A tool-using agent. This is the core of agentic infrastructure work. AutoGPT, Devin, and others. A lot of current technical energy is focused here. 3. Many people working with one LLM. A shared assistant for a team. Ben’s focus. This supports team-level memory, project synthesis, and aligned decisions. It’s emerging now. 4. Many people working with many agents, all coordinated through a shared LLM. This is the frontier. Imagine a team approves a campaign plan. Their shared LLM doesn’t just spin up agents. It engages the creative director, strategist, and producer, plus their teams (human and AI). The LLM knows the full context. It routes tasks, surfaces blockers, loops people in, and maintains alignment across the entire system. This isn’t a person using a tool. It’s people and AI, working together, across roles and workflows, with shared direction and shared memory. The shift is from individual productivity to shared intelligence. And the opportunity doesn’t stop at coordination. Negotiation. Conflict resolution. Team morale. Goal tracking. These are the complex, often messy parts of work where tools today tend to disappear. But this is exactly where AI can help. Not by replacing humans, but by holding context, clarifying intent, and accelerating momentum. That’s the future we’re building toward with Dash. AI that doesn’t just respond to prompts. It shows up in the group chat. It remembers the project goals. It knows what’s next. And it helps the whole team move. The future of work is multiplayer. And the most powerful teams will be human and AI, together, all the way down.

If AI won’t take your job, it may take your team’s psychological safety. A recent HBR article by Jayshree Seth and Amy C. Edmondson highlights that adding AI to teams can undermine trust, coordination, and communication if leaders are not deliberate. The research shows that when AI becomes a “teammate,” human teams often struggle more - not less as we assume. - Coordination breaks down. - Mutual understanding weakens. - People put in less effort, assume someone or something else will take responsibility, and speak up less. This matters because effective teamwork has never been only about task execution. It depends on fluid, adaptive collaboration and on people anticipating each other’s needs, adjusting their behavior in real time, and building shared understanding through everyday interaction. AI disrupts this process: 1. AI does not read context. 2. AI does not sense hesitation, tension, or uncertainty. 3. AI does not adapt its communication style to team dynamics. 4. AI does not participate in the informal, human moments where trust and psychological safety are built. What I want you to realize is this: 👉Psychological safety does not automatically survive the introduction of AI. Leaders cannot outsource trust and teamwork to technology. So, the current leadership question is no longer whether to use AI, but: How do we redesign teamwork so that humans still feel safe to think, question, and learn together when AI is in the room? That is where the future of effective teams will be decided. P.S.: what are your thought on that? Do your think organizations even realize this potential challenge?

What kills collaboration faster than conflict? Silence. How AI can fix it.     We've all been there: a meeting ends, everyone nods, no one asks questions... and yet, the project still goes sideways. The truth? Silence doesn’t mean clarity. Silence in teams can feel like alignment, but it's often confusion in disguise. It usually means someone didn’t feel safe or empowered to ask for it.   Even the best teams hit roadblocks:   Misunderstandings from assumptions Hesitation to ask questions Miscommunication that leads to rework   These challenges aren't new, but the way we tackle them can be.   This is where AI can quietly transform how your team collaborates. By acting as a neutral, judgment-free assistant, AI makes it easier for people to understand questions, clarify tasks, and stay aligned without fear of “looking dumb.”    Here's how:   ✅ Clarify complexity – AI can quickly summarize dense threads, documents, or meeting notes. ✅ Encourage curiosity – With the right prompts, AI makes it safe and easy to ask “obvious” questions. ✅ Keep teams in sync – AI can reinforce shared goals and priorities without sounding repetitive. It’s like adding a smart, impartial facilitator to every meeting, every teams thread, every project doc.   💡 Try this prompt to get started: "You are a helpful team assistant. Whenever I ask a question, respond with a reasonable amount of detail to help the team work together effectively." Simple but powerful to make missing information to all team members visible.     Ready to bring this into your team culture? Start with these steps:   1. Pick one team ritual (e.g., weekly meeting, retros, or docs) and layer in AI support. Let AI summarize, generate follow-up questions, or identify unclear points. 2. Encourage “clarifying questions” as a norm, not a nuisance. Use AI to increase curiosity and good inquiry. 3. Train with prompts. Craft a few go-to prompts your team can use in AI tools like Co-Pilot or whatever tool you use.   Collaboration doesn’t break down because people don’t care. It breaks down when people don’t feel clear and get frustrated.

In a world where AI announcements seem to drop every 15 minutes (seriously, it’s so hard to keep up), I've been reflecting on what actually matters beyond the hype. As a people leader navigating this landscape, I've learned that the challenge isn't just adopting AI tools quickly—it's adopting them thoughtfully. This is especially important at HubSpot, where helping our employees move faster helps our customers win faster. I'm seeing AI reshape not just what we do, but how we make decisions and prioritize our people. Here are some approaches that have worked well for us as we continue to test and learn: 1. Expedite access to AI tools and encourage experimentation. We're experimenting with the latest versions of Claude, Gemini, ChatGPT, and more—providing teams access within hours of new releases, not weeks. This creates a culture of experimentation and keeps us ahead of the curve. 2. Foster knowledge-sharing. We've created dedicated channels where employees share their AI wins and habits. Our People team sends a weekly "MondAI" digest featuring different employee use cases that inspire others across the organization. 3. Prioritize leader enablement. We've built AI-first resources, starting with People Leaders who then cascade knowledge to their teams. This isn't just about tools—it's about developing judgment for when AI enhances human work and when human expertise should lead. 4. Seek external expertise. We regularly bring in experts from companies like Anthropic and Google to share insights with our teams. We've cultivated a culture of learn-it-alls, not know-it-alls. 5. Integrate AI into existing workflows. We're incorporating AI tools directly into team processes, focusing on high-impact, repetitive tasks first. Our AI support bot now handles over 35% of tickets while maintaining high customer satisfaction. The most exciting part? Watching our teams develop the discernment to make AI work harder for them, not the other way around. When people and technology make each other stronger—that's the sweet spot. Fellow people leaders: How are you balancing rapid AI adoption with thoughtful implementation that truly empowers your people? Other insights we can learn from?

We may be approaching AI the wrong way. Most people still treat AI like a tool. But the people getting the best results treat AI like a teammate. TOOL mindset → Ask a question → Get an answer → Accept or discard the result TEAMMATE mindset → Give context → Coach the response → Iterate together Three weeks ago, I made a small change in how I work with AI. I stopped thinking about prompts. I started thinking about briefing a co-worker. Instead of asking one-off questions, I began interacting with AI the way I would with a new team member. Giving context. Providing feedback. Correcting it when it goes wrong. Helping it understand how I think. Over time, something interesting happened. The amount of AI slop dropped significantly. The number of iterations required to get high-quality output has reduced dramatically. This idea was reinforced when I listened to Jeremy Utley from Stanford University (Beautiful 13 mins video - Shared the link in comments) His research found something surprising. In many cases, AI actually made people less creative. So they compared the underperformers vs the outperformers. The difference wasn't the model. It was their orientation toward AI. Underperformers treated AI like a tool. Outperformers treated AI like a teammate. And when you treat AI like a teammate, your behavior changes. You coach it when the output is weak. You give feedback. You ask it to challenge your thinking. You even ask it: “What questions should I be asking about this problem?” At Capillary Technologies, we've been sharing several internal AI adoption stories. When I compare AI experts vs AI experimenters, one pattern keeps appearing. The experts don't just prompt AI. They work with it. For leaders, managers, and non-technical roles, this shift might be especially important. The skill of the next decade may not just be using AI. It may be managing AI like a teammate. And this coaching inspiration is packaged into our AI products (AiRa) by default before being handed over to our clients. If you're experimenting with AI, try this simple shift. Don't just prompt it. Coach it. Curious to hear from others here — Do you currently treat AI more like a tool or a teammate?

Leadership’s new reality? You’re managing people and algorithms. And no, there’s no playbook for it (yet). Leading a hybrid team is one thing. Scaling one? That’s where it gets interesting. Because bolting AI onto workflows isn’t enough anymore. The leaders I admire most right now are the ones redesigning collaboration, rethinking trust, and rewiring success metrics from the inside out. Here’s a Hybrid Leadership Playbook I’m seeing take shape — from the field, the boardroom, and every conversation in between: 📖 Rule 1: Build for Trust First, Outputs Second If your teams don’t trust their AI teammates, no amount of automation will save you. 📖 Rule 2: Orchestrate Tasks, Not Just Assign Roles Hybrid teams thrive when humans and AI each play to their strengths. 📖 Rule 3: Prioritize Human Learning Curves Your AI might be ready before your team is. (Train for the future — now.) 📖 Rule 4: Measure Momentum, Not Just Milestones Collaboration rates. Trust signals. Innovation loops. These are your new metrics. 📖 Rule 5: Align AI Success to GTM and Customer Impact If your AI isn’t helping drive personalization, pipeline, or loyalty — it’s just a gadget. 💬 Which of these 5 rules resonates most with how you're leading today? Or — what would you add to this list? 👇 Let’s evolve the playbook together. #Leadership #HybridTeams # #HumanAndAI #FutureOfWork #GTMStrategy #CMO #ScalingInnovation #AgenticAI #HCLTech

CIOs: What if your next leadership assessment involved managing AI agents, not just people? A recent Harvard study (NBER #33662) found that leadership skills can be accurately assessed by observing how someone manages a team of GPT-4o agents. The results were striking: performance managing AI strongly correlated (r = 0.81) with human team leadership. Why should leaders care? Because leadership in the enterprise is shifting. It’s no longer just about managing headcount, it’s about guiding distributed, intelligent systems. AI agents, automation, and orchestration layers are now part of every operating model. In the study, effective leaders didn’t micromanage. They asked questions, coordinated interactions, and adapted on the fly. The same behavioral traits that drive performance in human teams also improved outcomes with AI agents. This has real implications for CIOs: • How do you assess leadership readiness for a hybrid human-AI operating model? • Are your future leaders fluent in managing both people and intelligent systems? • Is your org structure evolving fast enough to reflect this shift? Leadership in this context isn’t about command-and-control. It’s about influence, framing, and system-level thinking. The line between managing humans and machines is blurring. The best CIOs will build leadership pipelines that don’t just scale with headcount but with intelligence, both artificial and human.