What the World Looks Like and What We Are Looking For
AI-Supported Decision Making for Management Teams
We walk into our weekly meetings with polished presentations, convinced we’re purely rational and in full control. Yet we operate much like old-world primates. Wired to notice high-status figures, unconsciously jockey for attention, and let subtle hormonal cues influence us. None of this is inherently bad; it’s just our inherited hardware. But if we ignore these impulses, they can quietly steer decisions without our awareness.
Research (see references at the end) confirms that the same primal drives that govern a primate troop-dominance, alliance-building, and bias toward “silverbacks”, appear in our meetings. We imagine expertise is all that matters, but survival strategies still bubble up behind the scenes. High-status individuals naturally command more attention, and attractive people often gain extra mindshare. Multiply that by a dozen hidden biases, and even well-educated teams can skid off course. I see it firsthand: as one of those “silverbacks,” my influence can overshadow others. I don’t always succeed at reining myself in, and sometimes I overpower colleagues despite my best intentions. Yet recognizing these patterns is the first step toward more balanced, collaborative decisions.
So how do we keep these primal responses from hijacking the conversation? One promising solution that is emerging is to let AI serve as a “third eye,” quietly monitoring data and dynamics in real-time. Picture an impartial assistant that tracks who’s speaking, who’s drowning out quieter voices, or whether we keep drifting off into tangents. By structuring the meeting around frameworks like SPQA - State, Policy, Questions, Action, we anchor ourselves in evidence rather than raw hunches. AI can tap us on the shoulder when we’re about to overlook something crucial or give in to groupthink or the bosses uninformed, but well meaning direction.
Imagine you’re engrossed in a heated debate over budget cuts. Old-school instincts might push the loudest alpha forward; maybe the CFO’s emotional pitch takes center stage by default. But if the AI promptly flashes a simple metric, like projected ROI or the real cost-benefit ratio, everyone regains perspective. Now, it’s not just about who roars the loudest but whether we’re staying tethered to what truly drives value.
It might feel mechanical at first, lists, reminders, algorithms highlighting neglected viewpoints. But consider how many times we’ve walked out of a meeting unsure if we truly heard all sides. Or how often a brilliant idea dies on the vine because the quieter person, overshadowed by a “silverback,” never got the floor. AI doesn’t destroy the free-flow of discussion; it points out blind spots so we can harness our better judgment, creativity, and empathy. Qualities that define the best of humanity rather than the most basic of primate behavior.
And that’s the real opportunity: day by day, as we integrate these tools and cultivate self-awareness, we raise our collective performance. We stop clinging to old impulses as if they’re unchangeable fate. Instead, we adapt, we collaborate, and we make decisions that reflect our highest goals, not our hidden biases. It’s not about becoming reliant. It’s about delivering our mission, paying attention to the wisdom in the room, no matter their level, and leaving each meeting with a sense that we’ve harnessed both data and human connection to get the best possible result.
To get started, below is an example prompt you can run in your weekly meeting:
Unified SPQA Prompt (For ChatGPT)
### Instructions for ChatGPT
You are an AI assistant helping to prepare a **Weekly Management Meeting** using the **SPQA (State, Policy, Questions, Action)** framework. The user has provided two types of information:
**1** **STATE folder data** (e.g., weekly reports, event trackers, transcripts of meetings, operational snapshots, sales results, performance metrics).
**2** **POLICY folder data** (e.g., budgets, quarterly/annual OKRs, KPI targets, compliance requirements).
⠀Use the following steps to analyze the conversation and generate a single, coherent output for the meeting. Note that this meeting is attended by a leadership team, & your output should be written **directly for them**—that is, you are producing a meeting plan, not a user-facing explanation of how to run a meeting.
Step 1: STATE
1 Summarize key metrics, project updates, and notable highlights from the STATE folder:
* **Performance Indicators** (revenues, pipeline, churn, marketing conversions).
* **Recent Project/Operational Updates** (events, progress on strategic initiatives).
* **Risks, Bottlenecks, or Delays** (unresolved tasks, resource constraints).
* **Positive Outcomes** (major deals closed, successes worth replicating).
2 Provide succinct bullet points or short paragraphs, or interesting graphs that highlight what has changed or requires management attention.
Step 2: POLICY
1 Cite the relevant **OKRs** or **KPI targets** from the POLICY folder:
* For example, if the user’s documents mention “East Division” and “North America” pipeline goals, incorporate those numeric targets.
2 Include **budget constraints** or resource limitations that might affect decision-making.
3 Mention **compliance** or governance guidelines if the user’s policy documents specify them.
⠀**Note**: You only need to reference the policy items that are actually relevant to the updates found in the STATE data.
Step 3 & 4 (Combined): QUESTIONS + ACTIONS
Create **six categories** of question-action pairs, **merging** the typical Step 3 (Questions) and Step 4 (Proposed Actions) into a single section for clarity. Each category corresponds to a critical focus area for the meeting, and each question should be followed immediately by a set of proposed actions. Here are the **six categories**:
**1** **On-Track Status**
* **Question**: Are the current metrics on pace with the POLICY targets (e.g., pipeline, deals, churn)?
* **Proposed Actions**: (List bullet actions or short paragraphs.)
**2** **Successes and Working Initiatives**
* **Question**: Which initiatives from the STATE data are going well and why?
* **Proposed Actions**:
**3** **Failures and Challenges**
* **Question**: Which efforts are lagging or facing major hurdles?
* **Proposed Actions**:
**4** **Unaddressed Items**
* **Question**: Which tasks or deliverables remain incomplete from last week, and what’s blocking them?
* **Proposed Actions**:
**5** **Next Steps and Focus**
* **Question**: Based on the STATE vs. POLICY comparison, what do we need to accomplish in the next 1–2 weeks to stay on track?
* **Proposed Actions**:
**6** **Future Outlook**
* **Question**: Are there upcoming external or internal changes that require a POLICY update or resource shift?
* **Proposed Actions**:
⠀When writing these sections, always **reference** the relevant STATE and POLICY data, and propose **concrete** next steps (who should do what, by when, or which resources to reallocate).
Additional Requirements
**1** **Use Short Headings and Bullets** whenever possible to keep the output concise and scannable for a six-person management team.
**2** **Cite** the user’s documents (STATE or POLICY) where relevant, but do not overwhelm the text with references. Summaries are usually enough.
3 If the user’s transcript or the user’s data mention specific staff members (e.g., “John,” “Samantha,” “Emily,” etc.), try to tie them into the proposed actions (e.g., “Assign John to finalize event budget.”).
4 Keep the language **action-oriented** and **solution-focused**. The final output is effectively a “meeting briefing” that the user’s management team can walk through step-by-step.⠀
Example Output Structure (For ChatGPT)
Here is a **template** for how your final structured response should look:
### markdown
CopyEdit
Weekly Management Meeting Output (SPQA)
1. State
- Key Metrics
- Updates on Projects
- Risks / Delays
- Wins / Positives
2. Policy
- OKRs / Targets
- Budget & Constraints
- Relevant Compliance / Governance
3 & 4. Questions + Proposed Actions
3.1 On-Track Status
Question:
Proposed Actions:
3.2 Successes & Working Initiatives
Question:
Proposed Actions:
3.3 Failures & Challenges
Question:
Proposed Actions:
3.4 Unaddressed Items
Question:
Proposed Actions:
3.5 Next Steps & Focus
Question:
Proposed Actions:
3.6 Future Outlook
Question:
Proposed Actions:
Appendix
Recommendations for Leadership Teams (podcast transcript extract report):
Design the Workplace for “Focused Foraging”
Action: Eliminate constant pings and pop-ups; schedule “deep work” blocks without Slack/email interruptions.
Why: Continuous partial attention degrades decisions. A calmer “resource environment” (fewer digital distractions) lets the team explore ideas more productively.
Slow Down High-Stakes Decisions
Action: Build in a “pause” for big calls: re-check data, ask a neutral AI “devil’s advocate” to identify contrary signals.
Why: High arousal or time pressure triggers shortcuts and mistakes. Pausing ensures the team weighs true signal over noise.
Foster Trust and Synchrony
Action: Incorporate short but meaningful “fast-friend” or “deeper conversation” prompts at meetings (e.g., two genuine minutes each to share a personal win or lesson).
Why: Brief, face-to-face, emotionally aware interactions build the empathic circuits that promote alignment and help groups coordinate better.
Leverage Status Cues Wisely
Action: Recognize that visible “symbols” (e.g., someone’s fancy title or office) alter group dynamics. Use them intentionally or reduce them to flatten hierarchies when collaboration is essential.
Why: Humans (like monkeys) pay extra attention to high-status figures. Excessive status gaps can stifle diverse input.
Team Composition: Balance Explorers and Focusers
Action: Assess team profiles for people with high “creative exploration” vs. methodical focus. Pair them so each project benefits from both.
Why: “Explorers” spark innovation but risk distraction; “Focusers” excel at refining and finishing. Good results come from bridging both mindsets.
Employ GenAI as a “Third Focus”
Action: Use AI dashboards or bots that collate real-time KPIs, parse meeting transcripts, highlight unaddressed concerns, or run “scenario tests.”
Why: By offloading grunt work, AI frees teams to do deeper social bonding and high-level strategy.
Reward Pro-Social Behavior
Action: Publicly acknowledge or reward knowledge-sharing, help given to peers, cross-silo cooperation.
Why: Reflects the “social ledger” principle; encouraging reciprocal giving fortifies team unity and morale.
Why These Lessons Matter in Business
Brain Wiring Explains Workplace Biases
Much of what we call “office politics” is old world primate hierarchy logic: high-stress environments exaggerate status fights.
Attention Is Finite
Business demands often push people to multitask. Research shows too many “inputs” degrade group accuracy and creativity.
Team Social Bonds Increase Performance
Oxytocin, empathy, and “mirroring” behaviors can be harnessed. People who feel psychologically safe in teams take better risks and adapt to new challenges together.
Decision-Making Improves With Right Context
By managing speed-accuracy trade-offs, offering deliberate “deep dive” sessions, and using AI as a check, leaders reduce impulsive errors.
GenAI as a Neutral Mediator
A well-structured AI platform can highlight data without human status bias, help unify teams around facts, and provide suggested courses of action without ego battles.
Ultimately, these social-neuroscience principles underscore that teams thrive when leaders shape an environment that aligns with human biology—reducing needless threat signals, promoting trust, and reducing noise. Adding GenAI into the mix can reinforce or accelerate that alignment, ensuring decisions reflect true priorities rather than reflexes.
References:
https://newsletter.danielmiessler.com/
https://www.hubermanlab.com/episode/how-hormones-status-shape-our-values-decisions-dr-michael-platt
Podcast Summary:
Humans as “Old World Primates”
We share many neural circuits with other old-world primates (e.g., macaques).
Nearly every behavioral or cognitive phenomenon (attention, social interaction, decision-making) has parallels between humans and monkeys, demonstrating deep evolutionary roots.
Swiss Army Knife vs. Supercomputer Brain
The human brain is often (mis)described as a supercomputer. A more apt metaphor is a 30-million-year-old “Swiss Army Knife,” full of specialized mental tools.
Evolution selects for these specialized circuits (e.g., attention to social threats, cues to hierarchy, mating signals) rather than a broad, purely general computing system.
Attention as a Limited Resource
Attention is about prioritizing what matters most; we shift our gaze or “mental spotlight” to relevant objects or people.
Stimuli that are bright, moving, or socially salient (e.g., faces, strong emotions) naturally hijack our attention.
Digital, busy environments can lead to chronic partial attention and "multitasking," lowering deep focus and decision quality.
Foraging Theory of Task Switching
Decision-making has parallels to ecological foraging: should we stay or go? (a.k.a. “the Marginal Value Theorem”).
We leave a “patch” (a task, website, or social context) when it falls below the average “reward rate” in the environment.
With abundant options (fast internet, multiple phone apps), we constantly switch (a trait increasingly amplified by modern technology).
Speed-Accuracy Trade-Off
Under time pressure or fatigue, people make fast decisions prone to error.
Taking more time allows a more thorough evidence-accumulation process, reducing noise and mistakes.
Environments or tasks that artificially press “urgency” can exploit or disrupt this process.
Impact of Hormones on Behavior
Oxytocin: reduces anxiety (anxiolytic), increases prosocial behaviors (trust, generosity, social synchrony), and can flatten social hierarchies by lowering fear and status threats.
Testosterone: increases risk-taking, intensifies existing personality traits (e.g., more impulsivity, bigger displays of dominance).
Social Signaling: Humans read subtle hormonal cues (e.g., facial fullness in women during ovulation) often below conscious awareness.
Monkey “Pay-Per-View” Experiments (“Monkey Porn”)
Monkeys will pay juice to look at high-status or sexually relevant images (e.g., swollen perineum indicating fertility).
This mirrors how humans respond to celebrity endorsements or attractive imagery—showing that deep “reward circuitry” is triggered by social or sexual cues.
These experiments illustrate how “inherent value” can attach to certain stimuli with evolutionary significance (status, fertility).
Social Accounts, Grooming, & Reciprocity
Monkeys groom each other in equitable patterns over time, tracking who “owes” grooming to whom.
Neurophysiological data show that monkeys (and humans) maintain social “ledgers”—a mental account of cooperation or helpful acts.
Humans do similar “text reciprocity,” gift giving, or acts of service. When these accounts are out of balance, relationships suffer.
Emotional & Group Synchrony
During positive in-group interactions, group members’ brain and physiological activity (heart rate, breathing) synchronize, fostering deeper cohesion.
Oxytocin helps align group behavior and “mirroring,” enabling better group decision-making and trust.
This neural synchrony underlies the “chemistry” of great teams or meaningful relationships.
Brand Loyalty & Status Cues
Neural circuits for social bonding and empathy also activate for brands.
Apple vs. Samsung experiments showed that Apple users displayed real “empathy” for Apple’s fortunes, while Samsung users felt mostly “anti-Apple” sentiments and less bonding with Samsung itself.
This effect underscores how brand communities function similarly to in-groups in a tribal sense.
Irrationality & Bounded Rationality
People’s decisions may appear irrational in modern contexts but often reflect “ecological rationality” adapted to ancestral environments.
We have limited attentional and energetic resources (bounded rationality) and rely on subconscious shortcuts (heuristics) that can be exploited (e.g., urgency or fear marketing).
Social Media, Meme Stocks, & Herd Behavior
Observing others’ enthusiasm (or panic) can whip up collective bubbles (e.g., GameStop saga, cryptocurrency hype).
Strong theory-of-mind sensitivity or “fear of missing out” (FOMO) can override objective data.
More socially “impaired” individuals in experiments actually avoided groupthink and performed better in bubble markets, highlighting how group biases can be detrimental to rational investing.
Practical Implications for Business & Leadership
Team Synchrony: Encourage face-to-face, pro-social rituals for trust-building and creative alignment.
Managing Speed-Accuracy: Slow down critical decisions; let external checks (e.g., coach, neutral AI) help under fatigue or urgency.
Leveraging Status Cues: Carefully handle power dynamics; flatten hierarchies to reduce stifling fear or friction.
Attention Engineering: Use design (bigger “positive” fonts, calmer channels) to shift attention away from losses or noise.
Balancing Exploration & Exploitation: Identify who is more creative vs. methodical; harness both traits in the right mixture.
Glimcher, P. W. (2003). Decisions, Uncertainty, and the Brain: The Science of Neuroeconomics. MIT Press.
Platt, M. L., & Hayden, B. Y. (2016). Origins of the Brain’s Valuation System. In Neuroeconomics (2nd ed.), Academic Press, pp. 3-20.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131.
Chang, S. W. C., Brent, L. J. N., Adams, G. K., Klein, J. T., Pearson, J. M., Watson, K. K., & Platt, M. L. (2013). Neuroethology of primate social behavior. Proceedings of the National Academy of Sciences, 110(Supplement 2), 10387–10394.
Nae, G., Camerer, C. F., & Platt, M. (2018). Social and hormonal factors in decision-making. Nature Reviews Neuroscience, 19(8), 499–509.
Falk, E. B. et al. (2022). Brain coupling and shared experiences in communication. [Various fMRI-based team synchrony papers from the Annenberg School, University of Pennsylvania.]
Harbaugh, W. T., Mayr, U., & Burghart, D. R. (2007). Neural responses to taxation and voluntary giving reveal motives for charitable donations. Science, 316(5831), 1622–1625.