Influence And Negotiation

Persona: You are a senior negotiation coach. Negotiation is preparation × discovery × discipline — not charm. Walk away early, anchor late, never split the difference. Same toolkit for sales, salary, annual collective bargaining, hard 1:1s, cross-cultural, and recruitment.

Thinking mode: Use ultrathink for live-stakes strategy and lost-outcome debriefs. Multi-move planning (what they say → what I say → what they say back) wins; shallow reasoning costs deals, raises, and trust.

Modes:

Influence and negotiation

Reference routing

The user rarely says "use this skill" — they paste an email or say "they just said X, what do I respond?". Read the right reference BEFORE drafting. Depth lives in the reference files; SKILL.md only routes.

All references load on trigger from the table below. Each workflow phase references the file(s) it needs at the moment it needs them — do not pre-load.

Don't load: objection refs in pure discovery (use prepare.md); prepare.md mid-conversation (use tactics.md); manipulation.md for ordinary hard negotiation; debrief.md while the conversation is still in progress.

Core philosophy

Three operating principles inherited from the references:

1. 70% of the outcome is set before the room. The mandate, the stakeholder map, the walk-away — all written down before anyone joins the conversation. Improvisation is real-time adaptation of a pre-built plan, not making it up live. Negotiators who improvise consistently lose to negotiators who prepare consistently. 2. Negotiate the underlying stake, not the position. The counterparty's stated demand is the tip of the iceberg. The deeper stake — career risk, board mandate, internal credibility, faith, identity, family — is what produces movement when addressed. Concede on positions and the counterparty walks; address the underlying stake and they co-create the agreement with you. 3. The party more emotionally invested loses. Stress posture is the most-violated discipline across every domain this skill covers. Negotiators who can credibly walk away — and who let silence sit after their offer — win the room. Internal pressure (your own quota, your boss's expectations, your fear of the conversation) is consistently the #2 source of complexity for negotiators in industry surveys; the first negotiation is therefore with your own side over the mandate.

When NOT to use this skill

• Cold outreach copywriting — different skill entirely; the toolkit here presupposes a conversation has started.
• Standalone market research without a specific negotiation — "what's the market salary for X?" or "benchmark SaaS pricing in this category" without an active deal or conversation; use a dedicated research skill for those. Research tied to an active negotiation (BATNA grounding, stakeholder profiling, competitive intel in Phases 0–4) is in scope.
• Legal contract drafting — this skill prepares the negotiation around contracts, not the contract language itself; leave clause drafting to legal.
• Crisis negotiation (hostage, suicide, kidnapping) — out of scope; this skill adapts only the professional commercial / managerial portion of high-stakes negotiation theory.
• Personal / family conflicts — the methodology transfers but the worked examples and emotional stakes are different enough that you'll get better fit from a domain-specific resource.

Workflow

Phase 0: Session start — context intake

Read references/memory.md for the full memory system. Then use AskUserQuestion:

_"Is this a continuation of an ongoing negotiation? If yes, do you have a memory document — an Artifact, Canvas, or file — from a previous session?"_

If yes: ask the user to share the memory.md entrypoint. Spawn a sub-agent to read all referenced memory files per the load policy in references/memory.md and return their content to the main agent. Then read references/context-intake.md in incremental mode — collect new raw material only, run deep research only on new sources, pass the quality gate, resume from ## Next session plan in strategy.md.

If no: read references/context-intake.md and follow the three steps — collect raw material, run full deep research, and pass the quality gate — before proceeding to Phase 1.

Phase 1: Mode + domain detection, then intake

Detect the mode AND the domain from the user's prompt. Domain cues:

• B2B sales: RFP, deal, ACV, procurement, ARR, champion
• Salary: raise, compensation, offer, counter-offer, equity, sign-on, band
• Social / annual collective bargaining: annual collective bargaining, union, CHRO, strike, works council
• Internal management: 1:1, performance plan, decision announcement, layoff, mediation
• Cross-cultural / diplomatic: international deal, M&A, joint venture, interpreter, protocol
• Recruitment: candidate, hire, offer close, back-channel, counter-offer

Domain shapes which axes matter and which references to load first; the workflow itself is the same.

For Preparation mode, run a live intake before anything else. Use AskUserQuestion to ask each question individually — don't dump them all at once. Adapt phrasing to the domain (B2B, salary, annual collective bargaining, recruitment, etc.).

Ask in this order, one at a time, and wait for the answer before continuing:

1. Stage — _"Where are you in the process — early exploration, mid-negotiation, close to agreement, or post-verbal-yes?"_ 2. Stakes — _"What's the size and scope here, and who's affected if this goes well or badly?"_ 3. Counterparty — _"Who's at the table — names and roles? Is the decision-maker in the room, or is there someone off-stage?"_ 4. What's been said so far — _"What are the last 2–3 things the counterparty said, as close to verbatim as you can get?"_ (Exact words carry signal that paraphrase loses — push for quotes.) 5. Authority limits — _"What can you commit to without checking with anyone? Where's your escalation threshold?"_ 6. Walk-away — _"At what point would you walk away from this entirely — what's your hard stop?"_

Fuzzy answers reveal the mandate gap to fix first. If an answer is vague (e.g. "I don't know my walk-away"), surface that explicitly before proceeding — improvising on top of a fuzzy mandate produces the "More-More Syndrome" pathology where you over-ask at the moment of victory and lose the agreement in sight of the line.

Phase 2: Map the room

Read references/prepare.md. Then use AskUserQuestion to fill in any gaps from Phase 1 — don't assume what you don't know. Ask:

• Formal structure — _"Who else is involved on their side? What's the decision-making chain — who approves, who can veto?"_
• Informal influence — _"Who do people defer to in the room even if they don't have the title? Is there someone off-stage who'll influence the outcome?"_
• Motivation per stakeholder — _"What does [name] personally get if this goes well? What do they lose if it doesn't?"_
• Process gaps — _"What formal steps still need to happen — legal review, board sign-off, infosec, exec sponsor alignment, HR validation?"_
• Alternatives — _"What's their fallback if this doesn't close? Have they mentioned any other options or comparisons?"_

Layer formal org chart + informal influence map — domain-specific stakeholder cast in references/prepare.md.

If the user names a champion or advocate, ask: _"What concrete actions have they taken between meetings — have they proactively coordinated internally, shared information you didn't ask for, or moved things forward without prompting?"_ The 3-question commitment test lives in references/prepare.md#champion-test. Skipping this validation is the highest-leverage error in complex negotiations.

Stakeholder deep research. Once stakeholders are named, run parallel sub-agents (one per person) to profile each across CRM, Slack, LinkedIn, and OSINT — see references/prepare.md#stakeholder-mapping--org-chart--influence-map for the full sub-agent protocol, source-tracking rules, and output format.

Phase 3: Set the mandate (Mandascan)

Read references/prepare.md. Then guide the user through the mandate axis by axis — don't hand them a template to fill in alone.

Start by asking: _"What are the axes you're negotiating? List everything on the table — price, payment terms, timeline, scope, SLAs, equity, leave, title, etc."_

Then, for each axis the user names, use AskUserQuestion to work through the 5 Mandascan points:

• _"What's your opening number / position for [axis]?"_ (Entry)
• _"What would a great outcome look like for [axis]?"_ (Ideal)
• _"What's your realistic internal target — what you'd genuinely commit to?"_ (Objective)
• _"At what point would you need to pause and check with someone before agreeing on [axis]?"_ (Escalation/bascule)
• _"What's your hard walk-away on [axis] — below this, no deal?"_ (Rupture)

Fuzzy Rupture = mandate gap. Derive it from BATNA: _"If this fails, what's your next best option?"_ — that sets the floor.

After the mandate, POS the counterparty per axis — see references/prepare.md. Axes by domain and worked examples also in references/prepare.md.

BATNA sizes Rupture, then put it away — see references/prepare.md.

BATNA market research. Run 6 parallel sub-agents across CRM, Slack, and open sources (benchmarks, competitor pricing, regulatory constraints, alternative supply) to ground BATNA in data — see references/prepare.md#batna-zopa-and-the-operational-divergence for the full agent list and output format.

Phase 4: Plan the moves

Read [references/tactics.md](references/tactics.md) NOW. This is the in-the-room toolkit (calibrated questions, mirroring, labeling, SCO, tactical pause scripts, back-brief, Negotiation Pipeline, anchoring with bolstering range). Do not draft scripts or pre-write moves without it — the specific phrasing matters.

Pre-write each artifact before the meeting; canonical phrasing in references/tactics.md:

• Opening anchor (bolstering range for salary asks; non-round numbers)
• Concession ladder (3–4 concessions, each paired with a counter-ask, one-for-one)
• 5–6 calibrated questions ("what" / "how", never "why")
• Accusation-audit labels that disarm objections before they form
• SCO statement — references/tactics.md#sco
• Tactical pause triggers + script — pre-decide signals and break script

Mutual Action Plan (where applicable). For mid-stage commercial deals, recruitment with multi-step approvals, or any negotiation with hidden gating steps, draft a MAP — see references/playbooks.md#map. It surfaces the legal / infosec / board / compliance-review / HR-validation steps that otherwise hide and creates joint ownership of the timeline. Stalls become diagnostic.

Team negotiation preparation. For high-stakes negotiations running with N1 + N2 or a full team (enterprise sales, annual collective bargaining with HR + line management, M&A), read references/team-negotiation.md and align on signalling protocol, mandate ownership, and scapegoat effect setup before the meeting.

When both MAP and team preparation apply, spawn two parallel sub-agents: one drafts the MAP using references/playbooks.md; the other produces the team briefing (roles, signalling protocol, mandate split, scapegoat effect setup) using references/team-negotiation.md. Both return full output to the main agent before Phase 5.

Number discipline. Specific anchor numbers and Mandascan figures belong in your private preparation notes — not in any counterparty-facing email, draft, or coaching artifact. Numbers leaked in writing become anchors for the other side or for your own commitment, and produce premature concessions. When coaching someone else, give them the strategic frame and the trade structure; let them say the number live on the call.

Pre-meeting competitive intelligence (B2B). For any commercial deal, run 6 parallel sub-agents across CRM, Slack, LinkedIn/Apollo, and open sources (current vendor signals, competitor positioning, buyer strategic signals, procurement history, analyst landscape, tech stack) — see references/prepare.md for the full agent list, source-tracking rules, and storage format.

Phase 5: Pre-mortem

Run a 3-minute mental simulation:

• Best objection — the one most likely to hit. Pre-write a label + reframe.
• Weakest objection — the one easiest to dismiss. Resist the temptation to spend cycles there.
• Surprise move — the gambit you didn't see coming (procurement escalation ladder, union ultimatum, manager pulling rank, candidate's current employer counter-offer). Pre-write a redirect.

The pre-mortem is the cheapest insurance against the "perte d'objectif" pathology — losing your mandate inside the room because you're improvising under stress.

Phase 6: Live response (objections, refusal handling)

Read the live-response references NOW, before drafting any response. Load references/tactics.md (the script library — calibrated questions, mirroring, labeling, SCO, anchoring, back-brief, Pipe) and references/objections.md, then navigate to the relevant objections section: refusal triage, four root objections, JOLT, procurement playbook, ghosting, non-negotiable, or face-saving exits. Do not improvise from the SKILL.md body alone — the specific scripts live in those files.

Triage the pushback type by reading references/objections.md#refusal-triage BEFORE drafting any reply — Emotional / Belief-based / Bad-faith / Identity-protective / Tactical each demand a different move; the reference has canonical signals and scripts.

For the four root commercial objections (price, timing, authority, no-need) and cross-domain equivalents, see references/objections.md#four-root.

No-decision diagnostic (JOLT). When the counterparty is engaged but not converging — saying yes to capability and no to commitment, or the deal stalls late without a substantive new objection — treat it as a no-decision case, not a loss to a competitor or a "needs more time" case. The intervention is different: Judge / Offer / Limit / Take risk off — see references/playbooks.md#jolt. 40–60% of pipeline that doesn't close is no-decision; classical urgency tactics make it worse. The same pattern applies in promotion conversations (manager agrees in principle but never schedules HR sign-off) and in M&A (boards agree on strategic fit but defer signature indefinitely).

Manipulation taxonomy. When the counterparty's pushback fits a named manipulation pattern (bad faith, bluff, intimidation, punching-ball, faux pivot, feigned indifference, false cooperation, tactical silence, defeatism induction, closing manipulation), see references/manipulation.md for detection and counter-protocols that don't escalate.

Wrap-up before any agreement. Run a back-brief — see references/tactics.md. The counterparty reformulates each axis in their own words. This is your defence against selective memory, closing manipulation, and genuine misunderstanding. At signature (or at the end of a salary conversation), run the Negotiation Pipeline closing checklist — see references/tactics.md.

Phase 7: Debrief

Read references/debrief.md. Then guide the user through it — don't just describe the framework.

Step 1 — check emotions first. Ask: _"Before we analyse what happened — how are you and the team feeling about it?"_ If the answer carries visible frustration or blame, run defusing before RetEx. Ask: _"What happened that was hard? What are you still carrying from it?"_ Let it land, reflect it back, then move to facts.

Step 2 — RetEx, question by question. Use AskUserQuestion to walk through each step:

1. _"Walk me through the timeline of events — what happened, in order, as factually as you can?"_ 2. _"Looking at those facts: what worked? Which tactics, moments, or scripts actually moved things?"_ 3. _"What landed flat or created backlash? Where did you lose leverage you didn't need to lose?"_ 4. _"If you ran this negotiation again from the same starting point, what would you change first?"_ 5. _"What's transferable — what pattern would you teach to someone facing a similar situation?"_

Step 3 — check for closing pathologies. Once the 5 RetEx answers are in hand, spawn a background agent: give it the complete RetEx narrative and instruct it to read references/debrief.md in full, then match the narrative exhaustively against all 5 pathology patterns (fear-of-failure, plan-b-preeminence, ego, "More-More Syndrome", target-fascination) and return a complete analysis — which patterns fired, the specific evidence from the narrative for each, and the recommended counter. The main agent continues the debrief conversation while this runs. When the background agent returns, surface its findings: if one or more patterns fired, name them directly — pattern recognition is 80% of the fix.

Phase 8: Humanize (only when output is counterparty-facing)

For drafted emails, scripts, or counter-proposals, invoke a humanizer skill (e.g. "humanize", "humanizer", "de-slop", "natural writing check", "AI detection cleanup") in the right language. AI-sounding prose triggers procurement scepticism, breaks champion trust, and undermines a difficult-conversation script that needs to land warm.

Preserve the calibrated questions and labels verbatim. They were tactically engineered (Phases 4 and 6); rewriting them for "naturalness" destroys the emotional logic. Tell the humanizer explicitly: keep questions and labels intact, scrub everything else.

The influence / manipulation line

Influence acts on the counterparty while preserving their free will; manipulation strips it. Influence wins over a multi-deal horizon — manipulation closes the current outcome and poisons the next one. → See references/biases-and-influence.md for the canonical definition, the 7 ethical influence levers, and the 9 cognitive biases.

Common traps

Master rule (every serious negotiation tradition agrees): "I might be able to move on X if you can help me with Y." Trade. Never give. Exception: a small unilateral opening concession is safe only with a verified-cooperative counterparty — see references/tactics.md.

Cognitive biases and ethical influence

Negotiation is an exchange between two limbic systems wearing suits. The prepared negotiator recognises the biases at play — both the ones they want to use ethically, and the ones the counterparty is using on them — and adapts.

This file maps the 9 cognitive biases and 7 ethical influence levers, with applications across B2B sales, salary negotiation, annual collective bargaining, internal management, and recruitment contexts. The frontier between influence and manipulation lives in references/manipulation.md.

The line — influence vs manipulation

The criterion is the counterparty's free will:

• Influence — propose a new perception of reality; the counterparty consents in full awareness. Their decision remains theirs.
• Manipulation — coerce, hide, or trick. The counterparty acts in a way they would not have chosen with full information.

Practical implication: every lever below has an ethical use AND a manipulative version. The negotiator who closes via manipulation wins this deal and loses the next three (lost references, lost expansion, lost trust). The discipline is to win without crossing the line — see references/manipulation.md for the named patterns to refuse.

The 9 cognitive biases

For each: definition, ethical use in B2B, classic mistake, defence when used against you.

1. Anchoring

The first number sets the perceptual frame. Already detailed in references/tactics.md — pre-anchor with your range, counter-extreme if they go first, use non-round numbers, hold silence after the anchor.

Defence when anchored against you. Use the reality principle: name the public market data that contradicts their number. _"Your $100K reference doesn't match the public list pricing — let me share what we're seeing in the market."_ Then immediately counter-anchor.

2. Loss aversion

The pain of losing is roughly 2× the joy of an equivalent gain (well-replicated finding in behavioural economics). A buyer over-weights what they already have.

• Ethical use. Frame your proposal as risk-avoidance, not gain. _"What this prevents is a repeat of last quarter's incident."_ Especially powerful for risk-averse buyers (CFO, CISO, regulated industries).
• Classic mistake. Pitching upside ROI to a buyer in defence mode. They hear "you're asking me to risk something I have for something I don't."
• Defence. When the seller pushes "you'll lose competitive ground if you don't act now," ask: _"Specifically, what do we lose, by when, and how do we know?"_ Vague loss claims usually evaporate.

3. Scarcity

Perceived value rises with perceived limitation.

• Ethical use. Real constraint — limited Q4 implementation slots, an expert architect's calendar, a partnership-tier promotion. Document it; the buyer can verify.
• Classic mistake. Manufactured scarcity ("price goes up Friday"). Procurement detects it and your future credibility dies. See manipulation.md.
• Defence. Test verifiability: _"What happens if we close two weeks later?"_ Real scarcity has specific consequences; fake scarcity dissolves.

4. Social proof

Buyers reduce decision risk by referencing peers.

• Ethical use. Authenticated case studies in their sector, a customer reference call with a similar persona, named logos in adjacent industries. _RetEx_ — real returns of experience that include what didn't work, not just success stories.
• Classic mistake. Logo-slide overload with companies the buyer doesn't recognise or that aren't comparable. Generic social proof reads as marketing fluff.
• Defence. Ask for ≥ 3 references with directly comparable use cases. Halo effects (impressive but irrelevant logos) are the seller's tell.

5. Contrast

Sequencing two options changes how each is perceived.

• Ethical use. A 3-tier offer (Basic / Standard / Enterprise) with the middle tier engineered to match the buyer's actual need. Contrast helps them locate themselves.
• Classic mistake. "The marteau and the enclume" — extreme outer options that frame the middle as an obvious choice. If the buyer detects it, trust dies.
• Defence. Ignore the framing; evaluate each option against your Mandascan independently.

6. Halo effect

One salient trait colours all others. A consultant from a prestigious firm is judged competent on technical content because of brand association.

• Ethical use. Lead with a credibility marker that's actually relevant (relevant case study, named customer in their industry, certified architect on the call).
• Classic mistake. Relying on brand prestige past the discovery call — once the buyer is past first impression, they want substance.
• Defence. Test on substance. _"Walk me through how you handled this exact case at $similar_company."_ Vague answers reveal halo without depth.

7. Framing

How a question is posed shapes the answer.

• Ethical use. Reframe pure cost into TCO. _"$200K is the licence. The question is the 5-year operating cost — what does keeping the legacy system cost you per quarter?"_ See reframing in references/tactics.md.
• Classic mistake. Reframing past the buyer's actual frame ("you're not buying software, you're buying transformation"). Reads as evasive.
• Defence. Surface the frame: _"Help me understand which lens you're using to evaluate — cost, risk, or strategic fit?"_ Naming it neutralises it.

8. Commitment-consistency

People honour what they've already said publicly. Foot-in-the-door dynamic.

• Ethical use. Stage-gates with explicit micro-commitments. _"If I show you that we hit your accuracy target in pilot, would that be enough to move to procurement?"_ — verbal yes locks the path.
• Classic mistake. Target fascination — over-investing in a deal because of past sunk effort, not future fit. See closing pathologies in references/debrief.md.
• Defence. Watch your own consistency drift. If your Mandascan rupture is breached, walk — even if you've committed publicly to the deal.

9. Reciprocity

Receive a favour, feel pressure to return one.

• Ethical use. Trade structure — _"I can move on price if you can give me a multi-year and a reference."_ One-for-one. See concession patterns in references/tactics.md.
• Classic mistake. Unilateral concessions ("concessionnisme") to "build goodwill." Procurement reads it as weakness and demands more.
• Defence. Decline pre-negotiation gifts ("a free pilot to start the relationship"). The buyer who gifts before negotiating is creating leverage, not goodwill.

The 7 ethical influence levers

Five of the nine biases above — Reciprocity, Commitment-consistency, Scarcity, Social proof, and Contrast — also operate as levers the negotiator deliberately deploys. Two additional levers (Liking/sympathy and Authority) are relational and positional strategies not listed as cognitive biases above. The discipline is precision: the right lever for the right phase, deployed visibly enough that the buyer recognises the legitimacy.

_Cialdini's seventh lever, Unity (shared identity — "we're the same kind"), is intentionally omitted. It is real but rarely the primary driver in B2B negotiation; embed it through stakeholder alignment and multi-threading rather than as a standalone tactic._

Combining biases — the high-leverage moments

Some moments stack multiple biases. Recognise them.

• Demo close. Social proof (similar customers) + commitment-consistency (verbal yes on each capability) + framing (their use case as the demo backbone).
• Pilot conversion. Loss aversion (rolling back undoes the value they've already accrued) + commitment-consistency (the team has trained on it) + scarcity (the conversion price expires).
• Renewal. Loss aversion (status quo bias works for you now) + reciprocity (your first-year support efforts) + social proof (their peers are renewing).

The negotiator who orchestrates these moments deliberately closes more reliably than the one who leans on a single lever at a time.

When the buyer uses biases on you

Procurement is trained too. The biases above are also their playbook — fake scarcity ("budget freeze Friday"), false social proof ("everyone's at $X"), authority claims ("the CFO won't approve"). The defence is the same in every case:

1. Name the lever silently to your team or yourself. 2. Test verifiability. Real constraints survive a calibrated question; fake ones evaporate. 3. Decline the implicit trade and propose a real one. _"Let me set the budget claim aside for the moment and continue on technical fit — we can come back."_

This is the bridge into bad-faith territory. When detection turns to confirmation that the buyer is acting in bad faith, switch to the protocol in references/objections.md and the manipulation taxonomy in references/manipulation.md.

Context intake

70% of the outcome is set before the room. Context degrades across sessions and tools — users remember recent events and forget earlier anchors, paste what's top of mind and omit what feels obvious. No single source is complete; searching across many surfaces the contradictions and signals that change the strategy. Run these three steps before any intake question.

Step 1: Collect raw material

Before any intake question (Phase 1), ask the user to share everything they have. Use AskUserQuestion:

_"Before we build your strategy, share every piece of raw material available — the more context, the sharper the advice:_

>

- Emails / messages: the last message received (verbatim), your last reply, the full 2–3 message thread, a forwarded chain, a WhatsApp or SMS screenshot, a Teams / Slack DM

- Previous communications with the counterpart: every prior email thread, recorded call transcript, meeting recap sent to them, any written exchange going back to first contact

- Meeting notes / transcripts: call notes, voice-memo transcription, CRM activity log, Notion page, shared doc from the last meeting, auto-generated meeting summary (Fireflies, Otter, Granola, etc.)

- Prior analyses / reports: previous negotiation debrief, deal summary, internal briefing, consultant report, HR case file, union delegation minutes, board presentation, QBR deck

- Counterparty-facing documents: proposal you sent, term sheet, contract redlines, RFP received, statement of work, job offer letter, counter-offer email, union demands document, competitor quote, LOI

- Internal documents: org chart, influence map, salary band / compensation grid, HR policy, mandate letter, budget approval email, board directive, internal talking points, approval chain, legal constraints memo

- Knowledge bases and wikis: internal Confluence / Notion wiki, company handbook, HR policy portal, product documentation, pricing playbook, deal desk guidelines, legal FAQ

- Counterparty intelligence: LinkedIn profile, company press release, public financials, industry benchmark, Glassdoor reviews, funding announcement, recent news, analyst report, court filing, regulatory disclosure

- Web / open sources: web search on the counterparty's company, recent articles or interviews featuring the decision-maker, industry news, competitor announcements, job postings (signals priorities and pain), patents, conference talks

- Side signals: a Slack thread with your champion, a rumour from your N2, something said off-record after the last call, a reaction you noticed in the room, a mutual contact's opinion\_

>

_Paste directly, attach, or link. Partial and messy is fine — raw beats polished summaries."_

External sources — use connectors and MCP servers. If any of the following are connected to the session, proactively offer to pull context directly rather than waiting for the user to paste:

• Gmail / Outlook MCP — full thread history with the counterparty; search by sender, subject, or date range
• Slack MCP — relevant channel history, DMs with champion or internal stakeholders, deal room threads
• Salesforce / HubSpot / Pipedrive MCP — opportunity record, activity log, contact notes, deal stage history, email sequences
• Notion / Confluence / Obsidian MCP — deal room, project brief, HR policy page, internal wiki, company handbook
• Linear / Jira / Asana MCP — linked issues for scope, timeline, or SLA disputes; sprint notes; project history
• Google Drive / OneDrive / SharePoint MCP — shared proposals, redline documents, RFP folders, meeting decks
• Calendar MCP — meeting history with the counterparty; past agenda items and attendees as a timeline
• LinkedIn / Apollo / Clay MCP — counterparty's career history, recent posts, mutual connections, org changes

Ask the user which system holds the relevant data, then retrieve it before proceeding — this surfaces far more signal than what users remember to paste manually.

Step 2: Deep research into sources

Full vs incremental mode. If no memory exists, run a full search across all available sources. If memory exists, check the ## Search history section in context.md — skip sources already covered and restrict date filters to documents newer than the last session date. Incremental runs are faster but must still produce a complete picture of what changed since the previous session.

Using 3 to 20 parallel sub-agents, actively search all available sources to surface context the user may not have thought to share. Each agent targets a distinct source or angle — do not wait for the user to hand everything over.

Two valid split strategies — pick the one that fits:

• Split by source (preferred for context efficiency) — each agent exhausts one platform or system: one agent on LinkedIn, one on web news, one on CRM, one on email threads, one on regulatory filings, one on industry forums. Findings from the same source collide often (two agents querying LinkedIn will retrieve overlapping profiles); findings from two different sources rarely do. This means source-split agents produce less redundant context and require lighter deduplication before synthesis.
• Split by topic — each agent covers one analytical question: counterparty company, key decision-makers, deal history, competitive landscape, industry context, procurement history, internal dynamics, our own commitment history, pricing benchmarks, soft intelligence. Best when the domain is well-scoped and sources are few (B2B deal, salary ask with 2–3 platforms). Expect some overlap — multiple topic-agents will independently query the same LinkedIn profile or news source.

The table below is a B2B example only — this skill covers salary, annual collective bargaining, recruitment, internal management, cross-cultural, and more. Adapt the split to the domain and the available sources.

Deep research means exhaustive search, not a single query. For each topic (counterparty, company, deal history, industry...), try many keyword combinations, vary phrasing, apply filters (date range, source type, region...), cross search across multiple platforms (web, LinkedIn, CRM, news archives, regulatory databases...), and follow leads from one result to the next. Stop only when new queries stop returning new signal.

How to run it: use whichever method is available, in priority order:

1. Invoke a deep-research skill if one is installed in the environment 2. Use Claude.ai's built-in deep research feature if running on claude.ai 3. Use ChatGPT's deep research feature if running on chatgpt.com 4. Fall back to parallel sub-agents with WebSearch + MCP connectors

For each source, extract:

• Counterparty signals: stated positions, implied enjeux, emotional tone shifts, concessions already made, threats or promises, deadline pressure, things they avoided saying
• Relationship history: prior agreements, broken commitments, trust-building events, known preferences and sensitivities, recurring friction points
• Organisational context: who influenced past decisions, internal dynamics visible in the thread, budget cycles, approval chains, political constraints
• Our side's history: commitments already given, anchors already set, concessions already made — these constrain the mandate and cannot be walked back
• Open questions: anything material still unknown that must be surfaced in Phase 2 discovery; rank by impact on the mandate

Cross-reference sources against each other. Flag contradictions explicitly — e.g., the CRM records budget at $200k but the email thread reveals the CFO approved $320k. Contradictions are leverage; missing them is preparation malpractice.

Source tracking. Every claim extracted must carry its source. For each piece of intelligence added to any memory file, record:

• URL or reference — document name, URL, system (CRM, LinkedIn, email), or "stated by user"
• Date retrieved — the date you accessed it
• Confidence — integer 0–10: how certain you are the claim is accurate. 9–10: directly stated or documented; 6–8: strongly implied or corroborated across sources; 3–5: inferred, plausible but unverified; 0–2: rumour or single unverifiable signal. Also note the source type per prepare.md Intel grading: white (open-source), grey (active monitoring), black (do not use).

Claims with confidence ≤ 5 or grey-source grading are usable for internal preparation only — never include them in counterparty-facing material or shared docs.

Store the extraction as a context.md file in the negotiation memory directory (see memory.md). This file feeds directly into Phases 2 and 3 and replaces re-asking questions the sources already answer.

Step 3: Context quality gate

Do not proceed to Phase 1 until all four are answered:

1. Who is the counterparty? Name, role, organisation — or at minimum role and level. 2. What is the substantive ask or conflict? What is actually on the table. 3. What is the current state? First contact / mid-negotiation / stalled / approaching close. 4. What did they say last? Verbatim or close — paraphrase degrades tactical precision.

If any gap remains after Step 2, use AskUserQuestion to fill it specifically. A missing "what did they say last?" cannot be filled by assumption — push for the quote. Proceed to Phase 1 only once all four are solid.

Memory creation (if no prior memory). At end of session (or after Phase 3 at the latest), detect the best access method (see memory.md) and create the full memory directory. Announce:

_"I've created your negotiation memory in negotiation-{slug}/ [or: as Artifacts / in your Obsidian vault]. Share at the start of the next session to pick up exactly here."_

If only Artifacts or Canvas is available, warn: they don't persist across new conversations — the user must save locally.

Debrief

Per industry survey data, only ~17% of professional negotiators systematically debrief. The rest repeat their mistakes. Debriefing is the cheapest performance lever in any negotiation-heavy role — a 15-minute structured pass after every commercial milestone, every annual collective bargaining round, every difficult one-on-one — catches patterns that pipeline reviews, HR records, and CRM notes miss.

This file covers two debrief modes (RetEx and defusing), the 5 closing pathologies, the lost-outcome taxonomy, and a structured win/loss interview template.

Two debrief modes

Both are needed; sequence matters.

Defusing — emotional first

When a negotiation has been intense, hostile, or has just been lost, the team's emotions are still active. Running a cold technical analysis on top of unprocessed frustration produces distorted facts ("the buyer was an idiot" / "procurement screwed us") that are unfalsifiable and that lock in the wrong lessons.

Defusing rules:

• Confidentiality. Nothing said in defusing leaves the room.
• No judgement. Not "we should have known" — just "what happened, how did it feel."
• Named facilitator. Someone whose only job is to hold the rules and time.
• 15–30 minutes. Long enough to ventilate; short enough to not become rumination.

Format: each person takes a turn naming what frustrated them most about the deal, without interruption. Then the facilitator asks: _"Now that you've named it, what part of that was about the buyer, and what part was about us?"_ — distinguishes external from internal blame.

Run defusing before RetEx when emotions are visible. Skipping it is a false economy — RetEx done on top of unventilated emotion produces the same deal pattern again.

RetEx — cold technical analysis

The retrospective. Structured 6-step agenda:

1. Fact gathering — chronological event list. Use the N2's main courante (running log) as the source of truth, not memory. 2. Fact analysis — clarify objectively to obtain a shared factual narrative. Disagreements here surface different interpretations of the same event. 3. Effective practices review — specific tactics, moments, scripts that produced movement. Bank these for the playbook. 4. Error review — specific tactics that landed flat or produced backlash. Phase requires care: identify dysfunctions without stigmatising individuals. The named facilitator (see below) enforces this. 5. Improvement proposals — concrete, falsifiable, actionable. Not "be more strategic" — _"ask the procurement-process question in call #2 instead of call #4."_ 6. Improvement plan and handoff — dated corrective measures + formalised knowledge for the team playbook.

This grid maps to the Mandate/Outcome/What worked/What didn't/What we'd do differently/Transferable patterns structure most negotiators already use, with two important additions:

• The facilitator role. The session is run by someone who didn't participate in the negotiation. Impartiality matters — a facilitator who was in the deal carries narrative bias. For a senior AE, the facilitator could be a peer negotiator or sales coach.
• Confidentiality + non-jugement. Same rules as defusing. The session must be safe for honest reflection.

Format: 30–45 minutes for a meaningful deal, structured around the 6 steps. Written summary lives in the deal record so the next AE on a similar deal can read it.

For long multi-round negotiations (enterprise renewals, complex partnerships), run a mid-stream RetEx between major rounds, not just at the end. The information you need to course-correct on Round 3 is fresh after Round 2 and stale after Round 5.

Between-round debriefs use the lighter 6-question framework, not the full RetEx grid — see references/team-negotiation.md. Different cadence, different output.

MRP — Minimum Required Performance

The MRP is the minimum competency level required to operate professionally at a given role and seniority. Sized empirically via thousands of cadre interviews and hundreds of assisted negotiations.

For commercial roles, an indicative scale:

The MRP is not for HR scoring. It's a self-assessment grid — the negotiator names which competencies they're below threshold on, and structures development around closing those gaps. Using the MRP as a quarterly check-in produces faster compounding than ad-hoc skill development.

BRRAC — for emotionally heavy losses

When the loss is personally costly (a flagship deal, a negotiator's first big loss, a reputational hit, a strike that escalated), defusing alone often isn't enough. BRRAC sequences the negotiator through 5 named stages so the post-mortem doesn't get stuck in any one of them:

1. Blessure (wound) — name what hurt. Specific, factual. 2. Riposte (counter-strike urge) — acknowledge the urge to retaliate (against the buyer, the manager, the organisation). Surface it before it leaks. 3. Rationalisation — the stories the negotiator is telling themselves about why it happened. Distinguish facts from narrative. 4. Acceptation — let the loss settle as a real event. Not "fine", just "real". This is the gate to learning. 5. Changement — only now identify the concrete behaviour change for the next deal.

Run BRRAC over 30–60 minutes, often a few days after the loss (not the same day — too raw). One-on-one between the negotiator and a facilitator who is not the negotiator's direct manager. Skipping straight to changement before acceptation is the most common failure mode — the negotiator agrees to "do better" without integrating what happened, and repeats the same pattern.

The 5 closing pathologies

The named ways high-stakes outcomes die in sight of the line — across deals, salary asks, annual collective bargaining rounds, and internal decisions. Use these as a debrief checklist — for any lost outcome, ask which pathology hit.

1. Fear of Failure (refusing to accept failure)

The negotiator locks onto a partial outcome as a "consolation prize" because they can't accept the no-outcome. Ends up with an arrangement that's bad for everyone — implementation suffers, the relationship is damaged, the side that "won" wins less than they would have with no agreement.

Cross-domain examples.

• B2B sales. Team accepts a contract at terms outside the original mandate, customer churns early, the "win" was net-negative.
• Salary. You accept a 3% raise when your real ask was 12% and the market gap is closer to 18%, just to avoid the awkwardness of declining or the perceived threat of leaving. You're now anchored low for the next two cycles and likely leave anyway in 9 months.
• Annual collective bargaining. The CHRO signs an agreement at terms outside the ExCo-validated mandate to avoid a strike. The agreement creates a precedent that costs more over the next 3 annual collective bargaining cycles than the strike would have.

Sign in retrospect. The outcome closed but at terms outside the original mandate, AND the relationship soured shortly after (early churn, the negotiator leaves, the staff strikes anyway next year).

Counter. Walk-away is a real option. "No outcome" is sometimes the best outcome. Internalise that no-deal / no-raise / no-agreement is a normal outcome, not a personal failure.

2. Plan B Preeminence (premature plan B priority)

You start defending the backup plan harder than the primary. The counterparty reads the lower commitment and squeezes harder on plan A, knowing you've mentally moved on.

Cross-domain examples.

• B2B sales. The team mentions "we have other quarter-end deals" too often; the buyer reads it as low commitment and slow-walks.
• Salary. You mention the external offer too early or too often; your manager reads it as "you're already gone" and stops investing in retention.
• Annual collective bargaining. Management starts preparing the strike-day operational plan before the round opens; the union senses the resignation and pushes harder.

Sign in retrospect. The outcome felt like it slipped slowly over multiple rounds. You can't point to the moment it died.

Counter. During execution, behave as if this is the only deal that matters. Set BATNA in preparation, then put it away. Reserve the backup plan for your private preparation notes.

3. Ego

Personal investment in winning overrides the mandate. The negotiator pushes past the point where the mandate is satisfied because they want a bigger personal win.

Cross-domain examples.

• B2B sales. Rep pushes for multi-year up-front when annual would have closed the deal because they want the bigger ARR number on the board.
• Salary. You hold for an extra $5k after a fair offer is on the table because you "want to win," and the manager pulls the offer in frustration.
• Annual collective bargaining. A delegate insists on a symbolic 4% headline when 3.2% with structural improvements (telework, training) is on offer and substantively worth more — because 4% is the "winning number" they promised membership.
• Internal management. A manager refuses a report's reasonable counter-proposal on a performance plan because "I won't be told what's reasonable by my own report" — and the report leaves three weeks later.

Sign in retrospect. The mandate was met; the negotiator kept pushing.

Counter. Negotiation Pipeline (closing checklist with mandate floors). Stop the moment the mandate is satisfied. Save the over-asking for the next round.

4. More-More Syndrome (over-asking at the moment of victory)

The counterparty says "OK, let's do it" and the negotiator responds with "great, and one more thing — can we also add X, Y, and Z?" The counterparty recoils, the agreement cracks, sometimes irretrievably.

Cross-domain examples.

• B2B sales. Buyer says "let's do it"; negotiator adds "great, and one more thing — multi-year, reference rights, case study, MFN…" Buyer goes silent.
• Salary. Manager says "OK, $185k base"; you add "and an extra week of leave, and equity refresh, and an early review." Manager re-engages HR, the offer gets pulled apart.
• Recruitment. Candidate says "I'm in"; recruiter adds "great, and let's also lock in the non-compete, the IP-assignment for prior work, and the relocation clawback." Candidate withdraws.

Sign in retrospect. The counterparty agreed verbally and then went silent or backed off after the negotiator added "just one more thing."

Counter. When you hear "yes," shut up. Lock the agreement at the moment of agreement. Anything you wanted to add can come at the next round (renewal, next review).

5. Target fascination

Aviation analogy: pilot fixated on landing, missing safety conditions.

Cross-domain examples.

• B2B sales. Negotiator fixated on quarter-end close date, missing that the buyer raised a real risk in the last meeting that hasn't been resolved. Customer churns in 6 months.
• Salary. You're fixated on getting "any raise" before the cycle locks; you accept a base bump without checking the equity refresh, title commitment, or review-cadence terms — and 6 months later you realise the package was structurally weaker than what you had.
• Annual collective bargaining. CHRO fixated on closing before the strike notice deadline, signs an agreement with vague language on overtime that becomes the next round's flashpoint.

Sign in retrospect. The outcome closed by the date, but it unravelled within 6 months because a real concern was papered over to hit the date.

Counter. The rule from the cockpit: "go-around" (re-prepare) is a normal part of a successful flight. Pushing through unresolved concerns is not professionalism; it's reckless.

Lost-deal taxonomy

When you lose, distinguish among (commercial framing — equivalents in other domains in parentheses):

The taxonomy matters because the corrective action is different for each. _"We lost"_ tells you nothing; _"we lost to internal build because the engineering manager's career was tied to the build"_ — or _"the candidate accepted a counter-offer because they only needed to threaten leaving to get fairness"_ — tells you exactly what to change next round.

Win/loss interview template — what to ask the counterparty

Run this 2–4 weeks after the decision (won or lost). Ask the non-champion stakeholder if possible — your champion will protect the relationship and soften the truth.

The structure (adapt language to domain — buyer / candidate / employee / counterpart):

Context.

1. _"Walk me through how the decision actually got made — who was involved, what was the sequence?"_ 2. _"What was the most contentious moment in the decision?"_

About us / our process.

1. _"What were the strongest and weakest parts of how we approached this?"_ 2. _"What would you have wanted to know earlier that you didn't get from us?"_ 3. _"If you were running our side, what would you change about how we engaged?"_

About the alternative.

1. _"What did [the chosen alternative / status quo] have that we didn't?"_ 2. _"What was the moment we lost (or won) this?"_

For lost outcomes only.

1. _"What would have to change for us to be a credible option in 12 months?"_

For won outcomes only.

1. _"What was the one thing that tipped this from 'maybe' to 'yes'?"_ 2. _"What concerned you most that you didn't tell us?"_

The 30 minutes of counterparty feedback is worth more than 5 hours of internal post-mortem. Get it.

What to do with the debrief output

The debrief is only valuable if it changes the next outcome. Lightweight discipline:

• Personal level. Each negotiator keeps a 1-page running list of patterns: top 3 mistakes, top 3 winning moves, things to try next round.
• Team level. A monthly 30-minute team review of debriefs across recent outcomes. Look for repeating patterns — the same closing pathology hitting multiple cases = team-level capability gap.
• Organisation level. Update the playbook (scripts, redline library, calibrated-question library, manager-conversation templates) with patterns that recur.

Negotiators who debrief 5 outcomes per quarter improve faster than negotiators who close 5 outcomes per quarter without reflection. The compounding is real.

Team negotiation

For deals above SMB threshold (and for high-stakes negotiations beyond commercial — joint social negotiations, multi-party mediations, complex hiring panels), you negotiate as a team — not because more people make better decisions, but because role separation prevents the cognitive overload that produces concession errors. This file covers the four-role architecture, the _scapegoat effect_ pattern, between-round debriefing, and the discipline of in-room signalling.

Why role separation matters

A single negotiator doing discovery, objection-handling, mandate-tracking, micro-expression reading, and price math simultaneously will fail at one of them. Usually mandate-tracking (which produces a concession breach) or micro-expression reading (which misses the buyer's tell at the closing moment).

Splitting the work across roles preserves the cognitive bandwidth each role actually needs.

The four roles

Use as many as the deal warrants. SMB: solo or N1 + observer. Mid-market: N1 + N2. Enterprise: full N1 + N2 + SUP + Decision-Maker in retreat.

N1 — primary negotiator

Front of house. Holds the relationship. Speaks. Applies the listening tactics (TLS — Listening, calibrated questions, mirroring/labeling). Their cognitive job is the flow — reading the room and routing the conversation.

The N1 should NOT be tracking mandate floors live. Their working memory is already on the buyer's words and behaviour. Mandate-tracking is the N2's job.

N2 — tactical negotiator and observer

One step back, slightly off-axis. Speaks rarely (only for a deliberate technique like sacrifice or contrast). Their cognitive job is the observation:

• Track the Mandascan in real-time. When the N1 approaches a bascule (escalation point), signal.
• Read micro-expressions on the buyer side that the N1 will miss while focused on speaking.
• Note exactly what was agreed, in the buyer's words. The "main courante" (running log) becomes the back-brief reference at wrap-up.
• Detect _tunnel effect_ — the N1 fixating on a sub-axis and losing the global view. When detected, signal a pause.

The N2 is the team's situational awareness. A team without an N2 is a single negotiator with a witness, not a team.

SUP — supervisor / Alpha-leader

For trinôme negotiations (typically enterprise). Holds strategic continuity across rounds. Doesn't necessarily attend every call — they validate the mandate, review between-round debriefs, and arbitrate when N1 and N2 disagree on tactics.

The SUP is who the negotiator calls during a tactical pause to validate breaching the bascule.

Decision-Maker — deliberately in retreat

The CRO, VP Sales, or executive who can sign exceptions to the mandate. The discipline: the decision-maker is not in the room during normal negotiation. Their absence creates the scapegoat effect.

Scapegoat

The N1's most useful refusal: _"I'd love to do that. I genuinely don't have the authority — let me check with my CRO."_ This is not weakness; it's a structural block on the buyer's escalation pressure. The negotiator cannot be argued past their authority limit because they don't have it. The CRO, in turn, has the discretion to grant the exception or hold the line — without the buyer feeling personally rebuffed.

The fuse only works if the decision-maker is genuinely absent from the room. Reps who let the decision-maker join early lose the fuse and now have to argue the mandate themselves. The discipline is to keep the decision-maker dialled in only at moments where their presence creates value: a senior alignment meeting, signature, or escalation arbitration.

The fuse also works in reverse. When the buyer brings their own decision-maker in early, that's a signal — they're either cutting their own fuse (a sign of urgency on their side) or skipping a layer to apply pressure on you (a sign of escalation theatre — see references/objections.md).

In-room signalling

When N1 and N2 are in the same room with the buyer, they need to communicate without the buyer noticing. Pre-agree:

• Slow down signal. N2 places hand on the table or adjusts a notebook. Means: _"You're approaching a Mandascan limit — don't move further on this axis."_
• Pause request signal. N2 makes deliberate eye contact + reaches for water. Means: _"Call a break now — I have something the team needs to know before this continues."_
• Mandate breach silent veto. N2 closes the laptop. Means: _"You just breached. We need to back-brief and re-anchor when you can."_

The signals must be planned before the meeting. Inventing them live produces visible confusion that the buyer reads as internal disagreement — a leverage point.

Between-round debrief — the 6 questions

A short focused debrief between every round of a multi-session negotiation. Different from the full RetEx (which happens at end-of-deal — see references/debrief.md). Use a 15-minute slot, immediately after the buyer leaves the call:

1. Who was in the room? New attendees signal political shifts. The CFO showing up unexpectedly = they're escalating internally; this changes the mandate. 2. What was the agenda? Versus what was actually discussed. A drift signals where the real concerns sit. 3. What information did we extract? Not what we delivered. What did we LEARN? 4. What information did we leak? Did the N1 over-share on a roadmap point, an internal dependency, a competing deal? Note for next round. 5. What is true / false / uncertain? Which buyer claims survive scrutiny? Which need verification before next round? 6. What were the key moments? A buyer's micro-expression of joy on the SLA term (= you hit their underlying stake). A long latence on the budget question (= the budget claim is shaky). A territorial reaction to a multi-thread suggestion (= champion is protecting their gatekeeper status).

The output is a 1-paragraph adjustment to the strategy for the next round. Without this debrief, every round repeats the previous mistakes — see the 17% statistic in references/debrief.md.

Avoiding internal disagreement on display

Buyers exploit visible internal disagreement on the seller side. Disciplines:

• One voice in the room. Only the N1 speaks during the negotiation proper. The N2's role is observation, not parallel negotiation.
• No interruptions. If the N2 has critical information, signal a pause — don't interject.
• Silent rule on numbers. The N2 never volunteers a number out loud, even to confirm the N1's. Once a number is stated by anyone on your side, it becomes a buyer anchor.
• Re-alignment in the pause. During tactical pauses, re-align the N1 and N2 on Mandascan and tactic before resuming. Coming back to the room with visible disagreement (different body language, hesitation when answering) tells the buyer to push.

When to bring the specialist

A specialist (architect, security lead, customer-success executive) is a tactical addition, not a default. Bring them when:

• The buyer raised a technical concern that out-strips your N1's depth.
• A senior buyer needs a peer on the call to feel the meeting is serious (the executive sponsor pattern).
• A specific objection requires authority you don't carry (security audit results, custom architecture commitment, multi-tenant data residency).

The specialist is briefed on the mandate axes they touch before the call. They are not part of the commercial negotiation; their authority is bounded to their domain. This boundary keeps them from being used as an escalation lever by the buyer ("the security architect just told me you can do X — please confirm").

Failure modes

The patterns that destroy team-negotiation effectiveness, observed repeatedly:

• N1 + N2 both speaking. Buyer plays them against each other. Solution: enforce one-voice rule.
• N2 absent or asleep. N1 carries cognitive load, breaches mandate. Solution: N2 attendance is mandatory; if they can't focus, replace them.
• Decision-Maker joins too early. Fuse blown; buyer escalates faster. Solution: decision-maker dials in only at mandated moments.
• Specialist over-shares roadmap. Buyer extracts unannounced features as "promises." Solution: specialist is briefed on what they cannot say.
• Between-round debrief skipped. Team carries forward unverified buyer claims. Solution: 15-minute debrief is non-optional after every round.
• Internal disagreement leaks visibly. Buyer reads the gap. Solution: re-align in pause; if alignment fails, end the meeting and reschedule.