# "I'm Probably The Team's AI Tool" **By Dan Cumberland** · Published May 9, 2026 · Categories: Leadership > The reason the debate over whether AI is a "tool" or a "teammate" never resolves is that both framings are partially true and individually insufficient. ... ## Why The Tool\-vs\-Teammate Debate Keeps Failing The reason the debate over whether AI is a "tool" or a "teammate" never resolves is that both framings are partially true and individually insufficient\. Calling AI a tool understates how much it reshapes role identity on a team\. Calling AI a teammate builds a kind of trust it cannot earn— it cannot push back, explain itself, or take blame\. Boston Consulting Group researchers, publishing in HBR in May 2026 \(Kropp, Bedard, Wiles, Hsu, and Krayer\)[3](/blog/blog-team-10-architecture#ref-3), found that anthropomorphizing AI as employees reduced individual accountability, increased unnecessary escalation, lowered review quality, and heightened employee uncertainty about their roles— without improving adoption\. Their summary line is the one that sticks: > "Calling AI a coworker builds trust it can't earn\. It won't push back, explain itself or take the blame\." The other pole has its own data\. Microsoft's 2026 Work Trend Index[4](/blog/blog-team-10-architecture#ref-4) \(a survey of 20,000 knowledge workers across 10 markets\) finds that organizational factors— culture, manager support, talent practices— account for 67% of AI's real workplace impact, while individual mindset accounts for 32%\. Only 26% of AI users say their leadership is clearly aligned on AI[5](/blog/blog-team-10-architecture#ref-5)\. Peer\-reviewed work in *Current Opinion in Psychology*[6](/blog/blog-team-10-architecture#ref-6) reports that teams which include an AI show lower cohesion and weaker identification than human\-only teams\. *Organizational Dynamics* names this directly[7](/blog/blog-team-10-architecture#ref-7): a real paradox where AI for improving teamwork can simultaneously restrict team dynamics\. Both poles have evidence\. Both poles have failure modes\. ```html-table
Tool framing fails this wayTeammate framing fails this way
Understates how AI reshapes role identity on a teamBuilds trust AI cannot earn (no pushback, no explanation, no blame)
Treats people who use AI as "users," not architects of how AI shows upReduces individual accountability for the work
Misses the social effects on team cohesionLowers review quality and increases unnecessary escalation
``` There's a way out of the binary\. It comes from a Dutch architect named Aldo van Eyck, who in the 1950s argued that real polarities are complementary halves of one and the same entity[8](/blog/blog-team-10-architecture#ref-8)\. Two truths\. One relationship\. The useful question shifts from *is* AI a tool or a teammate to what kind of relationship are we architecting— and that question can hold both poles at once\. Van Eyck didn't invent that move for AI\. He invented it for buildings— alongside a small group of architects who, in 1953, had the same fight in their own field\. ## Team 10 — The Architecture Metaphor Team 10 was a group of architects who in 1953 broke from CIAM— the dominant modernist movement of the early twentieth century— because they were tired of buildings that treated humans as machines for living[9](/blog/blog-team-10-architecture#ref-9)\. They proposed three replacement principles: **association, identity, and flexibility**\. Those three words are the vocabulary the AI\-on\-teams conversation has been missing\. The split happened at CIAM's 9th Congress\. By 1959, Team 10's critique had helped dissolve CIAM altogether[9](/blog/blog-team-10-architecture#ref-9)\. Core members included Alison and Peter Smithson, Jaap Bakema, Aldo van Eyck, Georges Candilis, Shadrach Woods, and Giancarlo De Carlo\. Their 1954 Doorn Manifesto articulated the three\-principle replacement[10](/blog/blog-team-10-architecture#ref-10), framed as a critique of CIAM's functional zoning— the practice of slicing human life into discrete zones for work, sleep, leisure, and circulation[11](/blog/blog-team-10-architecture#ref-11)\. Team 10 saw that as treating people as machines\. The team called themselves something the AI conversation could borrow: > "A small family group of architects who have sought each other out because each has found the help of the others necessary to the development and understanding of their own individual work\."[12](/blog/blog-team-10-architecture#ref-12) The bridge to AI is not a stretch\. The 2026 move to slot AI into "tool" or "teammate" zones is the same functional\-zoning instinct Team 10 fought against\. Different domain\. Same architectural failure\. Each of the three principles translates directly to a design question a founder can ask before AI shows up in any team workflow\. ## The Three Principles, Applied The three principles translate cleanly\. Association becomes a design question about whether AI is bringing the team closer or zoning them apart\. Identity becomes a question about whether AI protects what makes each person distinct or flattens them into prompts\. Flexibility becomes a question about whether the team's roles can evolve or get locked into tool\-shaped slots\. ### Association Association asks: does this AI move bring the team closer, or zone us apart from each other? Adding AI to a team changes the social structure of the team\. Peer\-reviewed research in *Current Opinion in Psychology*[6](/blog/blog-team-10-architecture#ref-6) reports that teams which include an AI show lower cohesion and weaker identification\. When AI gets rolled out individually— each person gets their own personal AI— the previously collaborative work of figuring something out together quietly becomes the privatized work of each person prompting alone\. Five Slack threads replace one whiteboard\. The architectural move is to design AI integrations with a team\-facing shape\. Shared prompts\. Shared review\. AI outputs the team interprets together\. Treat the model's first draft as the start of the conversation\. Our guide to [Building AI Culture](/blog/building-ai-culture) walks through what that looks like in practice\. > **Founder design question:** Does this AI integration give my team more reasons to talk to each other, or fewer? ### Identity Identity asks: does this AI move protect and amplify what makes each person on the team distinct, or does it flatten them into a tool\-shaped function? Workers' response to gen AI hinges on whether three psychological needs are met: competence, autonomy, and relatedness\. HBR \(March 2026\)[13](/blog/blog-team-10-architecture#ref-13), drawing on Self\-Determination Theory, found that when those needs go unmet, employees feel threatened, sometimes existentially\. The mechanism is specific\. When AI handles the part of someone's job that was the part they were known for— the analysis, the writing voice, the diagnostic instinct— the person becomes the human glue around an AI's outputs\. Their distinctness gets routed around\. The architectural move is to design AI integrations that take the routine off the specialist's plate so the specialist has more room for what they're known for\. In practical terms: use AI to absorb the parts of the role nobody got known for, and protect the parts they did\. Use AI as a sparring partner that amplifies what each person already does\. Tune AI on the team's distinct voice— custom GPTs, system prompts, or shared writing samples all work— so each contributor's signal gets stronger over time\. Jeremy Zug, a partner at Practice Solutions— a B2B services firm working in healthcare insurance billing— described the felt experience this way: his team integrated AI "as a sparring partner and as a tool that helps us do what we do best and magnifies what we're doing\." The team grew more comfortable as the months went on\. Each person's contribution kept showing up in the output— just amplified\. That's the identity principle as lived experience\. > **Founder design question:** After this AI integration, does each person on my team feel more distinct, or less? ### Flexibility Flexibility asks: does this AI move leave us free to evolve roles as the work evolves, or does it lock people into a slot the AI happened to make available? Microsoft's 2026 Work Trend Index[14](/blog/blog-team-10-architecture#ref-14) found that 65% of workers fear falling behind on AI, yet 45% say it feels safer to focus on current goals than to redesign work\. The result is calcification\. AI gets bolted onto current roles rather than used to question them\. A small team automates "around" each existing role and ends up with five people doing exactly what they did before, just with more outputs\. No actual redesign happened\. The architectural move is to revisit the org chart every quarter\. Ask which roles still make sense as AI absorbs more routine work\. Move people toward what only humans can do— judgment calls, relationship work, taste, accountability\. Microsoft frames the same point directly: the most effective AI users won't be the ones who do more things faster— they'll be the ones who redefine their value around what only humans can do[14](/blog/blog-team-10-architecture#ref-14)\. Our [AI Decision Framework for Founders](/blog/ai-decision-framework-founders) sketches when to make that call\. > **Founder design question:** What roles on my team should look different in six months because of how AI is reshaping the work? All three questions point at the same uncomfortable truth: someone in the room has to be the architect— and on a small team, that someone is the founder\. ## Architect, Not User The choice in front of a founder is what kind of team to architect now that AI is in the room\. The "whether to use AI" question is already answered by the market, by the calendar, and by the team\. Outsourcing the architectural call to IT, to HR, or to a vendor is itself a dehumanization move— those parties optimize for their own KPIs, and the team's identity is rarely on the list\. Reworked\.co[15](/blog/blog-team-10-architecture#ref-15) puts the diagnosis hard: > "The team is the subject, not the beneficiary\." That's the article's hardest sentence\. When AI shows up in a small workplace, the team is usually the subject of the deployment\. Only the founder is positioned to flip that equation\. The Microsoft data backs the structural claim: just 26% of AI users say leadership is clearly aligned on AI[5](/blog/blog-team-10-architecture#ref-5)\. The architectural step is being skipped almost everywhere\. Team 10 called themselves "a small family group of architects who have sought each other out because each has found the help of the others necessary\."[12](/blog/blog-team-10-architecture#ref-12) That's the frame to borrow for the founder's team\. A small family group of distinct people whose work each makes the others' work possible— with AI in the room amplifying each one of them\. If architecting AI's role on your team feels like one more thing on the founder's plate, that's exactly the kind of work a fractional AI leader can take on alongside you\. [Founders making the architectural calls AI demands](/for-founders/) deserve a partner for the drafting— more on what that role looks like in our breakdown of [What Is a Fractional AI Officer](/blog/what-is-a-fractional-ai-officer)\. The architecture is yours\. The drafting can be shared\. The small\-team specialist who realized they sounded like the AI deserves a team where they don't\. ## FAQ ### What is Team 10 in architecture? Team 10 was a group of architects active 1953–1981 who broke from the CIAM modernist movement to advocate for human\-scale, community\-centered design[9](/blog/blog-team-10-architecture#ref-9)\. They proposed three principles— association, identity, and flexibility— articulated in the Doorn Manifesto of 1954[10](/blog/blog-team-10-architecture#ref-10)\. Core members included Alison and Peter Smithson, Aldo van Eyck, Jaap Bakema, and Giancarlo De Carlo\. ### What are Team 10's three principles? Association, identity, and flexibility\. Association prioritized human relationships and social clusters over isolated functional zones\. Identity emphasized cultural and historical context in design\. Flexibility promoted adaptable structures that accommodate evolving needs\. The principles were articulated in Team 10's 1954 Doorn Manifesto[10](/blog/blog-team-10-architecture#ref-10)\. ### Should AI on my team be treated as a tool or a teammate? Both framings are partially true and individually insufficient\. "Tool" understates how much AI reshapes role identity on a team; "teammate" builds a kind of trust AI cannot earn— it cannot push back, explain itself, or take blame[3](/blog/blog-team-10-architecture#ref-3)\. The more useful question is what kind of working relationship you are architecting and whether each role on your team comes out of it more distinct\. ### Why does AI make some workplaces feel less human? Because absent deliberate architecture, AI integration usually reshapes work around the AI's capabilities rather than around human roles\. 63% of U\.S\. workers expect AI to make work feel less human, and 42% specifically cite dehumanization of work as a top concern \(Resume Now / Pollfish, n=1,003 employed U\.S\. adults, October 2025\)[1](/blog/blog-team-10-architecture#ref-1)[2](/blog/blog-team-10-architecture#ref-2)\. ### Who should design how AI shows up on a small team? On a small team, the founder is best positioned to architect AI's role\. They hold both the authority to make the call and the relationship to the team to know what's at stake\. Defaulting to IT, HR, or a vendor optimizes for their KPIs, and the team's identity is rarely on the list\. Only 26% of AI users say their leadership is clearly aligned on AI \(Microsoft Work Trend Index 2026\)[5](/blog/blog-team-10-architecture#ref-5)\. ## References 1. Resume Now \(survey via Pollfish\), "63% of Workers Say AI Will Make the Workplace Feel Less Human in 2026 — AI and Workplace Humanity Report" \(March 16, 2026; survey fielded October 2025, n=1,003 employed U\.S\. adults\) — [https://www\.prnewswire\.com/news\-releases/63\-of\-workers\-say\-ai\-will\-make\-the\-workplace\-feel\-less\-human\-in\-2026\-302713735\.html](https://www.prnewswire.com/news-releases/63-of-workers-say-ai-will-make-the-workplace-feel-less-human-in-2026-302713735.html) 2. Resume Now / Pollfish, "AI and Workplace Humanity Report" \(2026\) — [https://www\.prnewswire\.com/news\-releases/63\-of\-workers\-say\-ai\-will\-make\-the\-workplace\-feel\-less\-human\-in\-2026\-302713735\.html](https://www.prnewswire.com/news-releases/63-of-workers-say-ai-will-make-the-workplace-feel-less-human-in-2026-302713735.html) 3. Kropp, Bedard, Wiles, Hsu, and Krayer \(Boston Consulting Group\), "Research: Why You Shouldn't Treat AI Agents Like Employees," Harvard Business Review \(May 6, 2026\) — [https://hbr\.org/2026/05/research\-why\-you\-shouldnt\-treat\-ai\-agents\-like\-employees](https://hbr.org/2026/05/research-why-you-shouldnt-treat-ai-agents-like-employees) 4. Microsoft / Edelman Data x Intelligence, "Agents, Human Agency, and the Opportunity for Every Organization — 2026 Work Trend Index Annual Report" \(May 5, 2026\) — [https://www\.microsoft\.com/en\-us/worklab/work\-trend\-index/agents\-human\-agency\-and\-the\-opportunity\-for\-every\-organization](https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization) 5. Microsoft / Edelman Data x Intelligence, "Agents, Human Agency, and the Opportunity for Every Organization — 2026 Work Trend Index Annual Report" \(2026\) — [https://www\.microsoft\.com/en\-us/worklab/work\-trend\-index/agents\-human\-agency\-and\-the\-opportunity\-for\-every\-organization](https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization) 6. "AI\-teaming: Redefining collaboration in the digital era," Current Opinion in Psychology \(Elsevier, 2024\) — [https://www\.sciencedirect\.com/science/article/pii/S2352250X24000502](https://www.sciencedirect.com/science/article/pii/S2352250X24000502) 7. "Paradox of artificial intelligence as teammate," Organizational Dynamics \(Elsevier, 2024\) — [https://www\.sciencedirect\.com/science/article/abs/pii/S0090261624000913](https://www.sciencedirect.com/science/article/abs/pii/S0090261624000913) 8. Aldo van Eyck \(Team 10 Online\), "Twin Phenomena" — [http://www\.team10online\.org/team10/eyck/index\.html](http://www.team10online.org/team10/eyck/index.html) 9. "Team 10," Wikipedia \(cross\-verified with Architectuul\) — [https://en\.wikipedia\.org/wiki/Team\_10](https://en.wikipedia.org/wiki/Team_10) 10. "Team 10 and the Critique of CIAM," Fiveable Modern Architecture Study Guide \(2024\) — [https://fiveable\.me/modern\-architecture/unit\-10/team\-10\-critique\-ciam/study\-guide/dZ0DQP3llC2UmaAp](https://fiveable.me/modern-architecture/unit-10/team-10-critique-ciam/study-guide/dZ0DQP3llC2UmaAp) 11. "Team 10 and the Critique of CIAM," Fiveable \(2024\) — [https://fiveable\.me/modern\-architecture/unit\-10/team\-10\-critique\-ciam/study\-guide/dZ0DQP3llC2UmaAp](https://fiveable.me/modern-architecture/unit-10/team-10-critique-ciam/study-guide/dZ0DQP3llC2UmaAp) 12. "Team 10," Wikipedia / Team 10 Online — [https://en\.wikipedia\.org/wiki/Team\_10](https://en.wikipedia.org/wiki/Team_10) 13. "Why Gen AI Feels So Threatening to Workers," Harvard Business Review \(March–April 2026 issue\) — [https://hbr\.org/2026/03/why\-gen\-ai\-feels\-so\-threatening\-to\-workers](https://hbr.org/2026/03/why-gen-ai-feels-so-threatening-to-workers) 14. Microsoft / Edelman Data x Intelligence, "2026 Work Trend Index Annual Report" — [https://www\.microsoft\.com/en\-us/worklab/work\-trend\-index/agents\-human\-agency\-and\-the\-opportunity\-for\-every\-organization](https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization) 15. Reworked / Simpler Media Group, "That Helpful AI on Your Team? It Doesn't Work for You" — [https://www\.reworked\.co/digital\-workplace/ai\-coworkers\-is\-a\-misnomer/](https://www.reworked.co/digital-workplace/ai-coworkers-is-a-misnomer/) --- Source: https://dancumberlandlabs.com/blog/team-10-architecture/