Your weekly meeting should not be a status report. It should be a rapid-fire session where every role, energized by a human or AI, surfaces what is blocking its work and walks away with a clear next step. Here is the complete format, step by step, with everything you need to facilitate it yourself.
A Tactical Meeting is a structured operational meeting designed to do one thing well: remove immediate barriers to getting work done.
It is not a brainstorming session. It is not a strategy discussion. It is not a space where the most confident speaker dominates for forty minutes while everyone else checks their email. It is a tightly facilitated process where every participant, whether a person or an AI agent filling a role, can surface a tension and get what they need to move forward.
The meeting has two distinct halves. The first half is the preamble: a snapshot of current reality through checklists, metrics, and project updates. No discussion, just information. The second half triages tensions: the felt gaps between how things are and how they could be. Every tension gets resolved or turned into a next action or project. Every next step gets an owner. That is it.
There is a critical distinction embedded in this design that most teams miss: the Tactical Meeting is for triaging, not per se for solving. The purpose is not to analyse issues in depth. It is to quickly identify the minimum next step that allows someone to move forward. This single principle is what keeps a thirty-item agenda manageable in forty-five minutes.
There are two large forces that have emerged the past few years that make structured meetings more important then ever. First, people are working from home more and more, and second, organisations are starting to deploy AI agents alongside human team members. As these trends play out, the operational rhythm of the team has to accommodate role-fillers who do not absorb context through hallway conversations or body language. An AI agent filling a role needs the same structured space to report on its projects, surface obstacles, and request what it needs from other roles.
The Tactical Meeting provides exactly that. It treats every role equally, regardless of who (or what) fills it, and makes sure that all information relevant to all roles surface. Project updates from an AI agent monitoring customer onboarding metrics are processed the same way as a tension from a human team member about a delayed supplier. The structure does not care who speaks. It cares that every tension gets a next step.
Every Tactical Meeting requires two specific roles to function well.
The Facilitator holds the process. Their job is to guide the meeting through each step, keep conversations focused, enforce the rules, and ensure every participant gets space to process their tensions. The Facilitator does not contribute content. They do not share opinions on what should be done. They do not resolve tensions for people. They create space for each agenda-item owner to process their own tension, and they protect that space from being hijacked by others. The two most important phrases in the Facilitator's toolkit are "What do you need?" and "Did you get what you needed?" Everything else is in service of those two questions.
The Secretary captures outputs. Every action, project, and governance tension that emerges from the meeting is recorded by the Secretary. This creates the accountability trail: after the meeting, everyone knows what was decided, who owns what, and what structural issues need to be addressed in the next Governance Meeting. The Secretary also schedules meetings and distributes outcomes to all circle members afterwards. Do not save time by skipping the capture step. The discipline of recording every output is what makes the meeting useful beyond the room.
These are defined roles within the team, elected through the governance process. They are not the "team lead" by another name. Anyone in the circle can fill them.
In a team of multiple people, the Facilitator opens the meeting by inviting each participant, one at a time, to briefly share whatever has their attention. The purpose is simple: call out distractions and get present for the work ahead. One person speaks at a time. No discussion, no responses, no reactions. It is a moment for each person to arrive.
Facilitator phrase: "One at a time. Call out distractions, share how you're showing up today. No discussion."
If you are a solo founder running a team of AI agents, or there are no other human team members, the check-in round can be skipped.
Duration: 30 to 60 seconds per person in a multi-person team. Two to five minutes total.
The Facilitator reads through a list of recurring actions and asks the relevant role-filler for each one: "Check or no check?"
"Check" means the recurring action was completed in the preceding period. "No check" means it was not. No explanations, no justifications, no discussion. If someone wants to discuss a missed checklist item, they add an agenda item for the triage phase.
In practice, AI agents should report their checklist status before the meeting, not during it. In Nestr, agents tick off recurring actions as they complete them throughout the work period. By the time you sit down for the meeting, the checklist is already populated. You are scanning it for "no check" items, not waiting for each agent to report live. This turns a two-minute round into a ten-second scan. Any "no check" that concerns you becomes an agenda item for triage.
This step creates visibility into the operational heartbeat of the team. Are the recurring activities the organisation relies on actually happening? A thirty-second scan of "check, check, no check, check" reveals more about operational health than a thirty-minute status update.
There is a critical rule here that many teams overlook: checklist items cannot set new expectations on a role. If you want someone to do something on a recurring basis that is not already part of their role, that expectation must be proposed as an accountability through governance. The checklist only tracks what has already been agreed to.
Examples of checklist items might be:
Each role assigned a metric reports briefly on the latest data. A short highlight of what the number shows. Clarifying questions are allowed. Discussion is not.
Roles should have their metrics ready before the meeting. In Nestr, roles that track operational metrics, such as customer response times, conversion rates, or system health indicators, post their latest data as part of their regular work. By the time you sit down, the numbers are there. You scan, note anything that looks off, and add it to the agenda if needed.
Facilitator phrase: "Clarifying questions only. If you want a discussion, add an agenda item."
The purpose is to get a shared view of reality. If a metric raises a concern, that concern belongs on the agenda as a tension, not in a side conversation during the metrics review. Separating the surfacing of data from the discussion about what to do with it is what keeps this meeting fast.
The Facilitator goes through the list of active projects and asks each project owner: "Any updates?"
Facilitator phrase: "Share just what's changed since the last time we met. Not the whole history, just what's changed."
The role-filler either says "no updates" or shares briefly what has changed since the last meeting. Clarifying questions are allowed. Discussion is not. If anyone needs to discuss a project further, they add an agenda item.
For AI agents, project updates should not happen live in the meeting. They should be posted continuously as comments on their projects throughout the work period. In Nestr, agents add progress updates, mark completed actions, and flag blockers in real time. Right before the meeting, one agent should summarize all changes for each project. By the time the meeting starts, every project's status is current. You scan the updates, note anything that needs attention, and move on. This turns a ten-minute round into a two-minute scan and keeps the bulk of meeting time available for triage.
At this point, the Facilitator marks a clear transition. The first half of the meeting surfaced data. The second half processes tensions.
Facilitator phrase: "We have surfaced data about the circle and are now done with the preamble. From here, we build an agenda and triage tensions. The goal is to figure out one step forward for each item, not to solve the tension completely."
Each participant adds items to the agenda. One or two words per item, just enough to serve as a reminder. No explanation, no discussion about the item. Just capture the short label and move on.
Facilitator phrase: "Just give the Secretary one or two words as a reminder." If someone starts explaining their item, interrupt gently: "Let's just capture the label. You'll have space to process it in a moment."
In a solo founder setup, your AI agents should have captured their tensions as agenda items before the meeting starts. In Nestr, agents surface tensions during their work and link them to the next scheduled meeting automatically. By the time you sit down, you likely have a pre-populated agenda from your agents plus whatever you want to add yourself. Review the pre-captured items and add your own.
New items can also be added during the triage phase if something comes up while processing other tensions.
This is the heart of the meeting and where most of the time is spent.
The Facilitator works through each agenda item in sequence. The time available is divided roughly across all items. If you have ten items and thirty minutes, that is about three minutes per item. The constraint is the point. It forces rapid processing and prevents rabbit holes.
For each item, the Facilitator turns to the person who raised it and asks: "What do you need?"
The agenda-item owner then engages other participants as needed. There are five pathways through which a tension can be processed. The Facilitator should be familiar with all five and guide the agenda-item owner toward the right one.
Pathway 1: Request a next-action. The Content Writer Role asks "Could the Customer Support Role contact the client for a go on publishing the case study article?" The Facilitator asks the Customer Support role: "In your interpretation, does it serve your role's purpose or accountabilities to take that on?" If accepted, the Secretary captures it. A next-action is a single, concrete activity that moves something forward.
Pathway 2: Request a project. "There is outdated information in the onboarding emails. I think we need to put some effort into reviewing those." A project is an outcome with a definite endpoint that requires multiple actions. Same role-based questioning. If accepted, the Secretary captures the project.
Pathway 3: Request or share information. "I want to let everyone know the supplier changed their pricing." Or "I want to share about the congress I went to." The Facilitator checks: "This is just information, you do not need anyone to do anything with it, correct?"
Pathway 4: Request space for help. Sometimes the agenda-item owner does not know what they need yet. They need to think out loud or ask for input from the team. The Facilitator gives them space but stays alert for when a concrete request emerges. If it can't be figured out within the timebox, the Secretary may capture an action item to plan a follow-up meeting with the relevant roles.
Pathway 5: Set a new expectation (governance tension). This is the most important redirect. When the agenda-item owner is trying to establish an ongoing expectation, such as "I think someone should check the website analytics at least weekly and act on anomalies." the Facilitator asks: "Is this something you'd like to expect on an ongoing basis?" If yes, the Facilitator explains that ongoing expectations are defined through governance, not in a tactical meeting. The Secretary captures a governance tension for the next Governance Meeting. Then the Facilitator asks: "Until then, is there anything operational that needs to happen right now?"
The Facilitator watches for several common pitfalls during triage. When someone starts seeking approval for a decision they already have the authority to make: "What role has authority to make this decision?" When a conversation drifts into structural territory: "This sounds like it belongs in governance. Shall we capture a tension?" When someone who is not the agenda-item owner starts driving the conversation: "This is [name]'s tension. We can capture a new agenda item if you'd like to also process your tension?" When a discussion spirals without reaching a next step, ask the owner of the tension: "Is this helping you get what you need? What is the one next step?"
After each item, the Facilitator asks the agenda-item owner: "Did you get what you needed?" If yes, move on. If not, continue processing, but keep time in mind.
Engaging agents interactively during triage. While the preamble data should be pre-prepared, triage is where live interaction with your agents adds real value. When processing a complex tension, you can query an agent directly: "What patterns are you seeing in the support tickets you flagged?" or "Show me the three customers whose engagement dropped, and what you tried." The agent responds with context that helps you make a better decision about the next step. This back-and-forth is the meeting at its most useful, especially for a solo founder whose agents hold operational context that no single person could track alone. Use it for complex items. For simple ones (information sharing, straightforward action requests), the pre-prepared data is usually enough.
What about individual actions? Sometimes a request does not fit any defined role. Nobody's purpose or accountabilities cover it. When this happens, capture it under "individual action." The person agrees to do it as an individual, outside of any role. This is perfectly valid, but it is also a signal. If individual actions become patterns, they should be brought to governance so a role can be defined to cover the work. The Facilitator should note: "This does not fit any role right now. Let's capture an individual action for now, and also a governance tension so we can define who should own this going forward."
Can role-fillers decline requests? Yes, with nuance. The question is not "Do I want to do it?" but "Does it serve my role's purpose or accountabilities?" If it does, you should accept and prioritise it among your other work. If it does not, explain your reasoning or suggest a different role. The discipline of routing requests through roles, rather than through personal goodwill, is what makes the structure trustworthy.
In a multi-person team, the Facilitator invites each participant, one at a time, to share a brief closing reflection on the meeting. No discussion, no responses. This is the space where reactions to the process itself can surface: "I noticed we spent too long on the third item." "I appreciated how fast we moved today."
For a solo founder with AI agents, a formal closing round is unnecessary. What matters is that the Secretary (or Nestr, acting as your Secretary) distributes the outcomes: every action, project, and governance tension captured. In Nestr, all meeting outputs are recorded within the context of the roles and projects they belong to. Your agents, connected through MCP, can access what was decided and begin acting on it immediately.
It helps to be clear about the boundaries.
Governance changes cannot be made in a Tactical Meeting. You cannot create new roles, add accountabilities, define domains, or adopt policies. You can identify that these things need to happen and capture them as governance tensions, but the actual structural changes are processed in a Governance Meeting through the Integrative Decision Making process.
Personal conversations do not belong here. If there is interpersonal friction, schedule a separate conversation. The Tactical Meeting is for the roles in the circle to process their operational tensions, not for building team spirit or resolving personal dynamics.
Let me walk through a concrete example of how this works for a solo founder running a small organisation with AI agents.
You run a SaaS product. Your circle has yourself filling three human roles (Product, Customer Relations, and Circle Lead), and three AI agents filling defined roles: a Content Writer, a Social Media Analyst, and a Support Triage agent.
Before the meeting. Your agents have been working all week. In Nestr, they have already posted project updates as comments, ticked off completed checklist items, reported their metrics, and captured any tensions they encountered as agenda items linked to the meeting. By the time you open Nestr, the meeting is pre-loaded.
Preamble scan (2 minutes). You glance at checklists: all checks except Support Triage missed Tuesday's daily categorisation due to system downtime. You note it. Metrics: the Social Media Analyst has reported a 12% engagement drop, with product announcements performing worst. New trial signups are flat for the second week (your own metric). Projects: Content Writer has drafted two of four onboarding blog posts. Support Triage has created fifteen new FAQ entries, pending your review. You have a picture of current reality.
Build Agenda. The agents have already captured two tensions before the meeting: "Escalation criteria" (from Support Triage) and "Engagement content types" (from the Social Media Analyst). You add: "Signup stall" and "Ticket policy gap."
Triage (10 minutes). This is where live interaction matters.
You process "Engagement content types" first. You pull up the Social Media Analyst and ask: "Show me which content types drove the highest engagement in the past month, and how product announcements compare." The agent responds with a breakdown. You see that educational posts outperform product announcements three to one. You ask: "What would you recommend changing in next week's schedule?" The agent proposes shifting two of the four planned product posts to educational formats. That makes sense. Next-action captured: Social Media Analyst to revise next week's content schedule with two educational posts replacing product announcements.
"Signup stall." Straightforward. You ask the Content Writer to prioritise the onboarding blog posts and add calls-to-action to each one. Project priority updated. Thirty seconds.
"Ticket policy gap." You ask the Support Triage agent to show you the three tickets it could not categorise. The agent explains: all three involved refund requests, which fall outside its current accountability. You handle the three tickets yourself as individual actions. You also capture a governance tension: "Add refund triage accountability to Support Triage role and define a refund policy." This goes to your next Governance Meeting.
"Escalation criteria." The Support Triage agent needs clearer criteria for when to escalate versus handling autonomously. This is an ongoing expectation, which makes it a governance tension: "Define escalation policy for Support Triage role." For now, you give the agent a short-term instruction to escalate anything involving refunds or complaints.
Four items, twelve minutes. Every tension has a next step. Two governance tensions are captured for your monthly Governance Meeting. The agents can start acting on their new actions immediately.
Triage, do not solve. The single most important discipline. If a conversation goes beyond the appointed timebox, ask: "What is the one next step that would let you move forward?" Capture it and move on. Complex issues get their own dedicated working session outside the tactical.
Let your agents prepare the meeting. The preamble steps (checklists, metrics, project updates) should be pre-populated by your agents before you sit down. In Nestr, agents post updates continuously. Your job at meeting time is to scan, note concerns, and move to triage. The less time you spend on the preamble, the more time you have for the interactive processing that actually moves things forward.
Capture everything. Every action, every project, every governance tension. In Nestr, these outputs are tracked within the context of the roles and projects they belong to, and AI agents connected through MCP can see what was decided.
Look for governance gaps. Chris Cowan, one of the original architects of this meeting format, advises facilitators to actively look for activities happening outside of any defined role. These are the meatiest governance tensions and they make your next Governance Meeting productive. If you cannot find gaps in your governance, you are probably not looking hard enough.
Trust the process, especially at the start. The format can feel rigid when new. That rigidity is the feature, not the bug. After a few sessions, most teams report that this structure is a relief.
Use the right cadence. Weekly works for most teams. Some high-velocity teams run daily tacticals. Some teams with more stable operations run biweekly. The right frequency is whatever keeps tensions from accumulating to the point where they block work.
The Tactical Meeting is not the only place where work moves forward. Anything that can happen in a Tactical Meeting can also happen outside of it. Role-fillers can request actions, share information, and capture tensions at any time. The meeting simply provides a regular, structured rhythm that ensures nothing goes unprocessed for too long.
Roles, of course, do the bulk of their work between meetings. They do and post project updates, complete actions, tick off checklist items, and track metrics continuously. In Nestr, all of this activity is captured within the organisational context, visible to the team and available for the next meeting's preamble. When an agent encounters an obstacle, it captures a tension for the next meeting's agenda.
The Tactical Meeting is not where the work happens. It is where the obstacles to the work get resolved.
A standup typically asks three questions: what did you do, what will you do, what is blocking you? It is a reporting format. A Tactical Meeting has a structured preamble that surfaces data and then a triage process where tensions are actively resolved with captured next steps and owners. A standup identifies problems. A Tactical Meeting processes them.
The Facilitator allocates time roughly across all items. Some items need thirty seconds. Others need five minutes. If time runs out, unprocessed items carry to the next meeting or get processed outside of it. If there are a lot of items, the roles should prioritize them, it's up to the Facilitator to process all tensions. For complex issues, capture an action for the agenda-item owner to schedule a dedicated working session.
No. The structure is intentional and each step serves a purpose. However, steps can be very brief. "Any checklists? No. Any metrics? Nothing unusual. Projects? All on track." That preamble took twenty seconds. The point is not duration but discipline.
Requests can still be captured for them. The absent role-filler reviews meeting outcomes afterwards and picks up their actions. In Nestr, outcomes are distributed to all circle members and visible in the shared system.
Yes. AI agents participate by providing project updates, sharing metrics, and raising tensions through the shared system. In Nestr, agents post updates on their projects and capture tensions as agenda items before or during the meeting. A human (or an agentic meeting secretary) processes these alongside human contributions.
Recurring tensions are a signal. Either an action was captured but not completed (follow up on why), or the tension is structural and needs to be resolved through governance. Capture it as a governance tension rather than processing it operationally for the fourth time.
You can start with a shared document and a timer. But as the practice matures, and especially when AI agents participate, the value of dedicated software becomes clear. In Nestr, your organisational structure, projects, checklists, metrics, and governance records all live in one system. Meeting outcomes are captured within the context of the roles and projects they belong to. AI agents connect through MCP and can access the same organisational context as human team members. The meeting becomes grounded in the living reality of your organisation.
Here is what I have seen over more than a decade of practice.
Most teams do not lack information. They lack a process for turning information into action. Most teams do not lack goodwill. They lack a structure that gives everyone equal space to contribute. Most teams do not lack capability, whether human or AI. They lack a rhythm that keeps tensions from accumulating until they become crises.
The Tactical Meeting solves these problems with a format that is learnable, repeatable, and works equally well for human and AI role-fillers. It requires a Facilitator who holds the process, a Secretary who captures the outputs, and a team willing to trust the structure for a few sessions until the benefits become obvious.
If you are deploying AI agents, the Tactical Meeting is how you keep those agents connected to the operational reality of the team. If you are not deploying agents yet, the Tactical Meeting is how you build the structural clarity that makes AI integration possible when you are ready.
Either way, the meeting your team runs every week is either building your organisational capability or quietly eroding it. Make it count.
.png){kind=logo}
