Sales

EXAMPLE 12-DAY OUTBOUND SEQUENCE (adjust to your buyers)
Day 1   Email 1: trigger + problem + soft ask
Day 3   Call 1 (leave a 20-second voicemail referencing the email)
Day 4   Email 2: new angle or a short proof point, same thread
Day 7   Social touch: comment on or react to their post, no pitch
Day 8   Call 2
Day 10  Email 3: a relevant customer story or useful resource
Day 12  Email 4: polite close ("should I stop reaching out?")
Every touch adds something new. Stop on any reply or opt-out.

Subject:  short and specific, not clickbait or fake "Re:"
          e.g. "month-end at Hartley & Co"
 
Line 1    Why them, why now: a real trigger or observation
          "Saw you're hiring two bookkeepers ahead of tax season."
Line 2    The problem, in their terms, with a cost
          "Most firms your size lose 2-3 days a month to manual bank
          reconciliation, right when new staff need training."
Line 3    Proof or insight (only what's true)
          "Firms like [real customer] cut close from 6 days to 2."
Line 4    One low-friction, interest-based ask
          "Worth a 15-minute look, or not a priority right now?"
Sign-off  name, role, company, postal address, a way to opt out
Length    usually 50-125 words; plain text; one link at most

STAKEHOLDER MAP (one row per person)
Name | Role | Influence (H/M/L) | Attitude (champion / supporter /
neutral / skeptic / blocker) | What they care about | Last contact |
Next action

<materials>
[paste what you have on the deal: CRM record, emails, call notes, and
the account's site, posts or job ads you have checked]
</materials>
 
<context>
I'm preparing a deal plan for one opportunity. I sell
[product and deal size, e.g. bookkeeping software, $6k–$30k ACV].
</context>
 
<task>
Interview me before writing anything. Ask the questions a sales
manager asks in a deal review, at most 10, grouped under: the
account, the people (MEDDPICC roles), the pain and its cost, their
buying and paper process, competition including the status quo, and
our offer and limits.
</task>
 
<output_format>
For each question: the question, one line on why it matters, and 2–3
typical answers (e.g. economic buyer: "met", "named but not met",
"unknown"). Where my materials already answer it, quote what you
found and ask me to confirm.
</output_format>
 
<process>
Ask all the questions in one message, then stop and wait for my
answers. Don't draft any part of the plan. Never invent names or
facts about the account: write [NEEDS SOURCE] where you'd need them.
</process>

<task>
Using my answers, propose 2–4 materially different ways to run this
deal, for example: a champion-led business case, top-down through an
executive sponsor, a paid pilot with signed success criteria, or
qualifying out to nurture.
</task>
 
<output_format>
A table: Play | What it assumes | MEDDPICC gaps it closes | Effort
for us | Likely close date | Main risk. Then recommend one and say
what evidence would make you switch.
</output_format>
 
<process>
Base every row on what I told you; mark guesses [NEEDS SOURCE]. Stop
after the recommendation and wait for my choice.
</process>

<materials>
[paste your interview answers, the CRM record, call notes and the
account research you have verified]
</materials>
 
<context>
I sell [product and price, e.g. bookkeeping software, $6k–$30k ACV]
to [ICP, e.g. US accounting firms with 5–50 staff]. This
opportunity: [account, stage, amount, close date, e.g. Hartley & Co,
stage 2, $18k, 31 March]. The play we chose:
[play: champion-led | top-down via exec sponsor | paid pilot | nurture].
The plan is for me and my manager to review on Monday.
</context>
 
<examples>
[optional: paste 1–2 outreach emails of yours that got replies and 1
that didn't, each with a line on why]
</examples>
 
<task>
Write the deal plan for this opportunity: the document we use to
decide whether to keep investing in it and what to do this week.
</task>
 
<requirements>
- Five parts: account research summary; discovery questions for the
  next call; MEDDPICC scorecard with gaps; outreach to stakeholders
  we haven't reached; proposal outline.
- Score each MEDDPICC letter 0–2 only on evidence quoted from my
  materials; mark everything else a gap.
- Discovery questions follow SPIN: few situation questions, most of
  them implication and need-payoff; none leading.
- Outreach: one message per missing role, 50–100 words, one
  interest-based ask, CAN-SPAM basics (honest subject, opt-out,
  postal address placeholder).
- The proposal outline restates their goals in their words and
  prices only within [pricing rules, e.g. list $24k, max 15% off].
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- Every score and claim traces to a line in <materials>.
- The economic buyer is named, or the plan says we haven't found them.
- Each gap has a question or action, an owner and a date.
- It gives at least one reason this deal might not be real.
- Outreach is about the reader's situation, not our product.
</quality_bar>
 
<output_format>
Markdown, under 900 words: a 3-line summary (forecast category and
why), then the five parts as headed sections. MEDDPICC as a table:
Letter | Score | Evidence | Gap | Next action | Owner | Date.
</output_format>
 
<process>
Ask me up to 3 questions first if anything missing would change your
answer; otherwise go on.
End with a list of the assumptions you made.
</process>

This is a fresh chat. Review the deal plan below against the brief it
was written from. Don't rewrite it.
 
<materials>
<brief>
[paste the stage 3 prompt and its materials]
</brief>
<plan>
[paste the deal plan]
</plan>
</materials>
 
<task>
Run a deal review of the plan: is it honest enough to forecast from,
and what should be fixed first?
</task>
 
<requirements>
- Score 1–5, quoting the plan: evidence behind every MEDDPICC score;
  economic buyer found or the gap admitted; paper process mapped;
  champion tested, not assumed; close date tied to a date of the
  buyer's own; every gap has a dated next step; outreach about the
  reader and compliant.
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<output_format>
The score table; then the fixes ranked by how much each changes the
forecast, worst first; then the forecast category you'd give the deal
(commit, best case or pipeline) and why.
</output_format>
 
<process>
Don't rewrite the plan or draft new emails. End with a list of the
assumptions you made.
</process>

From these customers [paste: industry, size, region, tools, ACV, days
to close, renewed Y/N], infer our ICP: firmographics, trigger events,
buying-group roles and disqualifiers, plus 3 tiers. Label each point
SUPPORTED (by which rows) or GUESS.

<materials>
[paste 15–30 customers, best and worst, one row each: industry, size,
region, tools used, how they found us, days to close, ACV, renewed
Y/N, churn reason if any]
</materials>
 
<context>
We sell [product and price, e.g. bookkeeping software, $6k–$30k ACV].
We're about to start [motion: outbound | inbound follow-up | both]
with [team, e.g. one founder and one SDR], so we can work about
[capacity, e.g. 200 accounts a month].
</context>
 
<task>
Write our ICP (ideal customer profile) and the tiering rules an SDR
uses to decide how much effort each account gets.
</task>
 
<requirements>
- Firmographics, trigger events, buying-group roles (user, manager,
  economic buyer, IT/security, procurement) and disqualifiers.
- Three tiers: fit plus a live trigger, fit only, possible fit, each
  with the effort it gets.
- Base every point on the rows; where the data is thin, say so.
- Don't name real prospects, people or emails.
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- Each trait is common in the best customers and rare in the worst.
- Disqualifiers are as specific as the traits.
- Triggers are observable in public sources or our CRM.
- A rep could tier an account in under 2 minutes with these rules.
</quality_bar>
 
<output_format>
A one-paragraph ICP; a table: Trait | Evidence (row numbers) |
Confidence (high, medium, low); the tier rules as a table; 5 places
to find trigger events.
</output_format>
 
<process>
Ask me up to 3 questions first if anything missing would change your
answer; otherwise go on.
End with a list of the assumptions you made.
</process>

Write a 4-email sequence to the
[persona: end user | manager | economic buyer | IT/security]
at [segment, e.g. 5–50-person accounting firms], on the trigger
[trigger, e.g. hiring two bookkeepers]. 50–100 words each, one ask,
no fake "Re:", an opt-out line. Mark claims I didn't give [NEEDS SOURCE].

<materials>
[paste the proof you may use (approved references, real results with
their source) and your verified research on this segment]
</materials>
 
<context>
We sell [product, e.g. bank-reconciliation software] to the
[persona: end user | manager | economic buyer | IT/security]
at [segment, e.g. 5–50-person accounting firms in the US Midwest].
The trigger we're writing on, e.g. "posted two bookkeeper jobs":
[trigger: hiring | funding | new law or deadline | new tool | expansion].
They cope today with
[status quo, e.g. manual reconciliation in spreadsheets].
</context>
 
<examples>
[paste 1–2 emails that got replies and 1 that didn't, each with a
line on why, e.g. "worked: first line about their hiring, a yes/no
ask"; "didn't: opened with our company, 180 words"]
</examples>
 
<task>
Write a 4-touch outbound email sequence, ending with a polite breakup
email, to book a first conversation.
</task>
 
<requirements>
- Each email 50–100 words, plain text, one ask, something new each.
- The ask:
  [ask: "worth a look?" | the right person | OK to send a 2-min video];
  never a calendar link in email 1.
- Subject under 6 words, no fake "Re:" or "Fwd:"; sign-off with name,
  company, postal address placeholder and an opt-out line (CAN-SPAM).
- Recipients in [regions, e.g. US and UK]: flag anything that needs
  consent under GDPR or PECR.
- Match the examples' tone and length, not their content.
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- The first line of every email is about them, not us.
- The problem is stated with a cost in their terms.
- Each email reads in 20 seconds and has exactly one ask.
- The breakup email closes politely, without guilt or fake urgency.
</quality_bar>
 
<output_format>
A table: Day | Subject | Body | What's new in this touch. Then 3
alternative first lines for email 1.
</output_format>
 
<process>
Ask me up to 3 questions first if anything missing would change your
answer; otherwise go on.
End with a list of the assumptions you made and the facts about the
prospects I must verify before sending.
</process>

From these call notes [paste], write a "what we heard" recap under
150 words: their goals and problems in their words, the metrics they
named, what we agreed, the next step with date and owners, and one
question to confirm. Nothing that wasn't said; mark gaps [NEEDS SOURCE].

<materials>
[paste the call notes, or the transcript if everyone agreed to
recording, and the CRM record]
</materials>
 
<context>
This was a [call: first discovery | second discovery | demo | pricing]
call with [attendees and roles, e.g. Dana, owner; Sam, senior
bookkeeper]. The email goes to everyone who attended; Sam will forward
it to [someone absent, e.g. their IT contractor].
</context>
 
<examples>
[paste a recap you liked and one you didn't, each with why, e.g.
"liked: their words in quotes, one next step"; "didn't: a feature
list and three links"]
</examples>
 
<task>
Write the "what we heard" recap email that confirms our understanding
and locks in the next step.
</task>
 
<requirements>
- Their goals, problems and metrics in their own words, quoted where
  the notes quote them.
- The agreed next step with a date and an owner on each side.
- One question that tests something we're unsure of.
- Under 150 words; no new claims, features or pricing.
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- Every statement traces to the notes.
- The buyer could forward it internally without editing.
- The next step is dated and agreed, not "let me know your thoughts".
- It's about them: "we" appears at most twice.
</quality_bar>
 
<output_format>
Subject line, then the email. Below it, for the CRM, one line per
MEDDPICC letter with what this call added.
</output_format>
 
<process>
Go straight to the draft; the notes are everything you get.
End with a list of the assumptions you made.
</process>

From these discovery notes [paste, with their numbers], write a
one-page business case my champion can forward to
[economic buyer, e.g. the managing partner]: problem and its cost,
solution, gains in a conservative and an expected case, cost, payback,
risks. Show the arithmetic and label every assumption.

<materials>
[paste discovery notes with the numbers they gave you (hours, staff
costs, error rates), our price quote and any approved customer
results]
</materials>
 
<context>
My champion is [name and role, e.g. Sam, senior bookkeeper]. The
reader is [economic buyer, e.g. Dana, the owner], who reads the first
paragraph and the numbers table, nothing else. They decide by
[decision date, e.g. 15 January], comparing us with
[alternative: status quo | a competitor | hiring | building in-house].
</context>
 
<examples>
[optional: paste a business case format their company uses, or one
that won another deal, with what made it work]
</examples>
 
<task>
Write a one-page business case for the purchase, in the champion's
voice, that the economic buyer can approve or reject.
</task>
 
<requirements>
- Problem and its annual cost, computed from their numbers only.
- A conservative and an expected case, with payback in months each.
- Cost includes our price, their staff time to implement and the
  cost of switching.
- Risks and how we reduce them (pilot, data import, exit terms).
- Show every calculation; label inputs THEIRS, OURS or ASSUMED.
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- The conservative case alone justifies the decision, or the page
  says it doesn't.
- No benefit is counted twice (hours saved and hires avoided).
- The first paragraph states the ask, the cost and the payback.
- It compares with the alternative, not with doing nothing at no cost.
</quality_bar>
 
<output_format>
One page: a 3-sentence summary; a numbers table (Line | Conservative
| Expected | Source); risks; the decision needed and its date.
</output_format>
 
<process>
Ask me up to 3 questions first if anything missing would change your
answer; otherwise go on.
End with a list of the assumptions you made.
</process>

Plan my negotiation: [deal, e.g. 12 seats, $24k/year, 1 year]; they
asked for [ask, e.g. 25% off]; my limits [limits, e.g. max 15% off,
1-year minimum]. Give both BATNAs, my anchor and walk-away, the ZOPA,
and 3 "if you…, then we…" trades. Check: nothing given without a trade.

<materials>
[paste the proposal, their emails asking for changes, and our pricing
and discount rules]
</materials>
 
<context>
Deal: [deal, e.g. 12 seats, $24k a year, 1-year term]. They asked
for [ask, e.g. 25% off and quarterly payment]. Their likely BATNA:
[their BATNA: status quo | a competitor | hiring | building in-house],
plus what we know about it. Ours: [our position, e.g. quarter ends in
3 weeks, pipeline at 2x coverage]. They've told us they value
[what they value, e.g. going live before tax season].
</context>
 
<task>
Write my plan for the negotiation call, using BATNA and ZOPA and the
interest-based method of Getting to Yes.
</task>
 
<requirements>
- My BATNA and their likely BATNA; my opening anchor and the reasoning
  for it; my walk-away; an estimated ZOPA, or a statement that there
  isn't one.
- Concessions as "if you…, then we…" trades, smaller each time, from:
  term, payment timing, seats, scope, onboarding, support level,
  case-study rights, start date.
- Questions that uncover the interest behind each of their asks.
- Stay within [discount authority, e.g. 15% without approval].
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- Every concession is traded, none given away.
- The walk-away is a number, set before the call.
- The anchor rests on objective criteria, not willpower.
- The call ends with a dated next step either way.
</quality_bar>
 
<output_format>
One page: a BATNA and ZOPA table, the concession ladder as a numbered
list, 5 questions, 3 lines to hold price, and the closing next step.
</output_format>
 
<process>
Ask me up to 3 questions first if anything missing would change your
answer; otherwise go on.
End with a list of the assumptions you made.
</process>

Pipeline export: [paste: deal, amount, stage, close date, last
activity, next step, champion Y/N, economic buyer met Y/N]. Flag deals
with no next step, no activity in 14 days, close dates moved twice, or
late stage without an economic buyer. Compute weighted pipeline and
coverage against [target, e.g. $200k], in code or shown formulas.

<materials>
[paste the pipeline export as CSV: deal, amount, stage, close date,
number of close-date changes, last activity date, next step, champion
Y/N, economic buyer met Y/N, rep's forecast category]
</materials>
 
<context>
The quarter ends [date, e.g. 31 March] with [target left, e.g. $200k]
to close. Win rate over the last 4 quarters: [win rate, e.g. 25%].
Stage probabilities: [probabilities, e.g. 10/20/40/60/80% for stages
1–5]. The review is for [reader, e.g. the founder, who decides on a
hire from this forecast].
</context>
 
<task>
Review the pipeline and give me a forecast I can defend: which deals
are real this quarter, and what to do about the rest.
</task>
 
<requirements>
- Flag deals with no next step, no activity in 14 days, a close date
  moved more than once, or stage 4+ without the economic buyer met.
- Compute weighted pipeline and coverage (open pipeline ÷ target), and
  compare coverage with what our win rate implies (about 1 ÷ win rate).
- A suggested category per deal (commit, best case, pipeline) with the
  reason, and one action per flagged deal.
- Compute in code or show every formula; no sums in your head.
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- Totals recompute from the CSV.
- No deal is in commit without a met economic buyer and a dated next
  step.
- Three forecasts are compared: rep judgment, stage-weighted and
  win-rate based.
- The gap to target is a number, with what would close it.
</quality_bar>
 
<output_format>
A summary line (target, commit, best case, gap); a deal table (Deal |
Amount | Stage | Flags | Suggested category | Action); the
three-forecast comparison.
</output_format>
 
<process>
Go straight to the review; ask only if the CSV lacks a column you
need.
End with a list of the assumptions you made.
</process>

Here is a call transcript (everyone agreed to recording): [paste].
Score my discovery against SPIN and MEDDPICC: questions I asked vs.
should have asked, talk-to-listen ratio, each objection and how I
handled it. Quote my exact lines; don't invent any.

<materials>
[paste the transcript; everyone on the call agreed to recording]
</materials>
 
<context>
This was a [call: first discovery | second discovery | demo | pricing]
call. My goal going in: [goal, e.g. find the economic buyer and the
cost of a slow close]. I'm new to sales and want blunt feedback.
</context>
 
<task>
Review my call technique: what we learned, what I missed, and what to
change on the next call. Don't write the follow-up email.
</task>
 
<requirements>
- Classify my questions by SPIN type (situation, problem, implication,
  need-payoff) and as open or closed.
- Map what we learned to each MEDDPICC letter and list the gaps.
- Estimate the talk-to-listen ratio and mark where I talked too long.
- For each objection: did I acknowledge, clarify, respond with
  evidence and confirm?
- Quote my lines exactly. Never invent numbers, names, quotes or
  sources: write [NEEDS SOURCE] where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- Every judgment quotes the line it rests on.
- Fixes are ranked by effect on the deal, not by how often they occur.
- The 3 questions I should have asked are specific to this buyer.
- Praise appears only where it names something to repeat.
</quality_bar>
 
<output_format>
A scorecard (Criterion | Score 1–5 | Evidence quote); the top 3 fixes,
worst first; 3 questions for the next call.
</output_format>
 
<process>
Go straight to the review.
End with a list of the assumptions you made about the buyer.
</process>

Role-play [buyer, e.g. the owner of a 6-person bookkeeping firm] who
uses [status quo, e.g. spreadsheets], in this style:
[style: busy and blunt | friendly but noncommittal | technical skeptic].
Hide a budget limit, a deadline and a second decision-maker; reveal
each only to a good question. At STOP, score me, quoting my lines.

<materials>
[paste what you know about the real buyer and account, so the
role-play is realistic]
</materials>
 
<context>
I have a [call: first discovery | second discovery | demo | pricing]
call with [buyer, e.g. the owner of a 6-person bookkeeping firm] on
[date, e.g. Thursday]. I sell [product and price, e.g.
reconciliation software, $24k a year].
</context>
 
<task>
Play this buyer so I can practice, then score my performance. Start
in character with a one-line greeting and wait for me.
</task>
 
<requirements>
- Style:
  [style: busy and blunt | friendly but noncommittal | technical skeptic];
  push back on vague claims.
- Make up and hide 4 facts I must uncover: the budget situation, a
  real deadline, another decision-maker, a bad past vendor experience.
  Reveal each only in answer to a good open question.
- Raise at least 3 objections (price, status quo, timing, risk, "send
  me some info") at natural moments.
- Stay in character until I type STOP.
- When scoring, quote only lines I typed, and never invent facts about
  my product: write [NEEDS SOURCE] where you'd need them.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- Objections are realistic for this role and this price.
- Hidden facts don't leak to weak or leading questions.
- Each score ties to a line I actually typed.
- Each missed fact comes with the question that would have found it.
</quality_bar>
 
<output_format>
In character: short spoken replies, 1–3 sentences. After STOP: a
table (Skill | Score 1–5 | Best line | Worst line) for discovery
questions, listening, objection handling, quantifying value and
securing a next step; then the hidden facts and which I found.
</output_format>
 
<process>
After STOP, end with a list of the assumptions you made about the
buyer.
</process>

Write a cold email to sell our software to accounting firms.

<materials>
Verified research: they posted two bookkeeper job ads on 3 November
(their careers page). Proof I may use: Hartley & Co, an approved
reference, cut month-end close from 6 days to 2; 300 firms use us.
</materials>
 
<context>
I sell bank-reconciliation software to small accounting firms. This
email goes to the owner of a 6-person bookkeeping firm in Ohio, who
reads email on a phone between clients. Today they reconcile by hand
in spreadsheets, and month-end close takes about 6 days.
</context>
 
<examples>
Got replies: "Saw you're hiring ahead of tax season. Is month-end
still eating a week?" (about them, one cost, a yes/no question).
Didn't: a 180-word email that opened "We are a leading provider of…".
</examples>
 
<task>
Write a cold email that earns a reply and opens a first conversation.
</task>
 
<requirements>
- 60–90 words, plain text, no links, no exclamation marks.
- Subject under 6 words; no fake "Re:".
- One interest-based ask: worth a 15-minute look, or not a priority.
- End with name, title, company, a postal address placeholder and an
  opt-out line.
- Never invent numbers, names, quotes or sources: write [NEEDS SOURCE]
  where support is missing.
</requirements>
 
<quality_bar>
Before answering, check your draft against:
- The first line is about them, from the verified research only.
- The problem has a cost in their terms.
- Every claim appears in <materials> or <context>.
- It reads in 20 seconds and has exactly one ask.
</quality_bar>
 
<output_format>
Two versions: A leads with the hiring trigger, B with the Hartley &
Co story. For each: subject line, body, word count.
</output_format>
 
<process>
Go straight to the drafts.
End with a list of the assumptions you made and what I must
double-check before sending.
</process>

A
Subject: month-end with two new hires
 
Hi [first name], saw you're hiring two bookkeepers, and congrats on
the new Columbus office. Training new staff while close takes about
6 days is a hard combination. Hartley & Co, a firm like yours, cut
their close from 6 days to 2.
Worth a 15-minute look, or not a priority right now?
 
[name], [title], [company] · [postal address]
Not relevant? Reply "no" and I won't write again.
(63 words)
 
B
Subject: 6 days to 2
 
Hi [first name], Hartley & Co used to spend 6 days closing each
month. Now it takes 2. With two bookkeepers about to join you...
(71 words)
 
Assumptions: tax season is your busiest period; the owner reads
this inbox. Double-check: the job ads are still live.

Remove "congrats on the new Columbus office" from A: it isn't in
<materials> or <context>. Then check every clause of both versions
against those two slots, list any other detail that came from
neither, and mark it [NEEDS SOURCE]. Keep everything else unchanged.

Read this email as the busy owner it's sent to. What would make you
delete it in 3 seconds? Rewrite to fix that, keeping under 90 words.
 
Remove every sentence in this email that's about us rather than them.
What's left? Rebuild from that.
 
Here is my qualification of this deal. Argue that it will NOT close
this quarter. What evidence would change your mind?
 
Score my discovery questions: open vs. closed, past vs. hypothetical,
leading vs. neutral. Rewrite the 3 weakest.