Close more deals: 10 AI skills for sales conversations
objection-handler
answer the objection behind the objection
How the two work together
Claude thinks it through. Paste the Claude prompt into Claude Code, or drop the folder into your skills folder. Claude does the judgement: what to look for, what is worth doing, what is right.
Codex gets it done. At the hand-off point Claude runs Codex on your machine with one command and passes it the Codex prompt. Codex does the mechanical part and hands the result back. Claude checks it before you see it.
No API key to set up: Claude calls the Codex you already have installed. If Codex is not installed, Claude does that half itself and tells you.
Prompt for Claude
---
name: objection-handler
description: Turns a raw objection into the question to ask first, the real objection underneath, a reply short enough to say aloud, and a way to take the buyer's risk off the table. Use when a buyer pushes back on price, timing or a competitor.
---
# Answer the objection behind the objection
The question to ask first, the objection really in the way, a reply under 21 seconds, and a risk remover you can offer.
## Claude does
1. Collects: the objection in the buyer's exact words, when it landed (call number, stage), what discovery established (problems, metrics, decision process, other stakeholders), the product facts you may state, and the seller's name. Refuses to guess prices, contract terms, reference customers or roadmap dates.
2. Judges indecision first. Of deals lost to indecision, 44% go to a preference for the status quo and 56% to fear of getting it wrong. Says which, quoting the buyer's own words.
3. Writes ONE clarifying question, then tells you to pause. Across 67,149 recorded calls top reps lengthen that pause while average reps shorten it, and answer with a question 54.3% of the time against 31%: a strong answer to the wrong objection makes it worse.
4. Ranks up to three objections underneath, each with the tell that confirms or kills it.
5. Drafts the spoken reply under 60 words. Average reps run knee-jerk monologues of 21.45 seconds. Ends with "Does that make sense?" to close the loop.
6. Adds a recommendation: what you would do in their position, so blame for a wrong choice is shared. Never urgency or FOMO: pushing those backfired 84% of the time.
7. Takes risk off the table using only mechanisms you confirmed exist: pilot, opt-out, staged rollout, proof step.
8. On price, sets the channel: send the number in writing, negotiate by phone. Email only wins 5%, phone only 15%, both 35%.
## Then Codex does
Codex repeats one decided structure across the library: the same objection rewritten for every remaining stage, channel and persona, plus your standing objection list, deduped, word-counted, laid out as cards. Bulk, not judgement.
Claude fills {OBJECTION_LIST}, {PRODUCT_FACTS}, {SELLER_NAME}, {APPROVED_RISK_REMOVERS} first, then run:
```
codex exec --sandbox danger-full-access --skip-git-repo-check -C "<working folder>" - < CODEX.md
```
## Claude checks
1. Counts words on every reply. Over 60 words is over 21.45 seconds spoken. Reject.
2. Counts question-led cards. Below 54.3% of the file, reject.
3. Greps for urgency, deadlines, expiring discounts, FOMO. One hit rejects the file.
4. Checks every price, name, statistic and integration against {PRODUCT_FACTS}. One invented fact rejects the whole file, not the line.
5. Confirms each pricing card says write the number, negotiate by phone.
6. Rerun once with the failures listed. Failing twice, Claude writes the file.
## Rules
- Ask before you answer. Never answer an unconfirmed objection.
- Never argue urgency or scarcity at an indecisive buyer.
- Never negotiate a price by email alone.
- Public information only.
- Never invent a fact, a number or a quote.
- Anything sent in someone's name says whose name it is.
## If Codex is not installed
Claude builds the library itself and says so.
## Built from
- Gong Labs, Handling Sales Objections, https://www.gong.io/blog/handling-sales-objections, 67,149 recorded calls analysed.
- The JOLT Effect, Dixon and McKenna, https://www.jolteffect.com/blog/what-is-the-jolt-effect, research on deals lost to indecision.
- Gong Labs, negotiation channel study, https://www.gong.io/resources/labs/this-deadly-simple-negotiation-mistake-is-costing-you-deals-heres-how-to-avoid-it/, 37,671 opportunities.
Prompt for Codex
# Answer the objection behind the objection - Codex task
## You are given
- {OBJECTION_LIST}: the objections to cover, each in a buyer's own words.
- {PRODUCT_FACTS}: the only facts, figures, names and integrations you may state. Nothing outside this block is true.
- {SELLER_NAME}: the person whose name the replies go out in.
- {APPROVED_RISK_REMOVERS}: the pilots, opt-outs, staged rollouts and proof steps that genuinely exist.
## Produce
`objection-cards.md`, one card per objection in {OBJECTION_LIST}, each with these fields in this order:
1. **Objection (verbatim)**
2. **Clarifying question** - one question, then the instruction "pause and say nothing".
3. **Likely objection underneath** - up to three, ranked, each with the tell that confirms it.
4. **Spoken reply** - 60 words maximum, plain speech, ending with "Does that make sense?"
5. **Recommendation** - one sentence starting "In your position I would", in {SELLER_NAME}'s voice.
6. **Risk remover** - one item drawn only from {APPROVED_RISK_REMOVERS}.
7. **Channel** - for any price or contract objection, write: send the number in writing, negotiate it by phone.
8. **Word count** - the real count of field 4.
Then `objection-index.csv` with columns: objection, category (price, timing, competitor, authority, status quo, fear of failure), question_led (yes/no), word_count.
## Rules
- Every card must open with a question. At least 55% of the whole file must be question-led.
- No urgency, no deadlines, no expiring discounts, no scarcity, no FOMO language anywhere.
- No fact, price, customer name, statistic or integration that is not in {PRODUCT_FACTS}.
- No risk remover that is not in {APPROVED_RISK_REMOVERS}.
- Never quote a person or a company unless the quote appears verbatim in the inputs.
- Deduplicate: two objections meaning the same thing become one card with both phrasings listed.
- British English. No em dashes.
## Return
Print, in this order:
1. The two file paths written.
2. Total cards, and cards per category.
3. The question-led percentage across the file.
4. A table of every card over 60 words, or the line "no card over 60 words".
5. Every fact you used that came from {PRODUCT_FACTS}, listed, so it can be checked.
6. The line "urgency scan: clean" or the offending phrases with their card names.
Built from the best public work on this
Sources
- **Gong Labs, Handling Sales Objections** - https://www.gong.io/blog/handling-sales-objections - 67,149 recorded sales calls from a database exceeding one million calls. Source of the pause rule, the 21.45 second monologue, the 54.3% question-led response rate against 31% for average reps, and the loop-closing check.
- **The JOLT Effect, Matt Dixon and Ted McKenna** - https://www.jolteffect.com/blog/what-is-the-jolt-effect - published research on deals lost to customer indecision. Source of the 44% status quo against 56% fear split, the 84% backfire rate on urgency, and the shared-blame reason for making a recommendation.
- **Gong Labs, the negotiation channel mistake** - https://www.gong.io/resources/labs/this-deadly-simple-negotiation-mistake-is-costing-you-deals-heres-how-to-avoid-it/ - 37,671 sales opportunities. Source of the 5% email only, 15% phone only, 35% both channels win rates.
Best public prompt we found for this job
https://www.gong.io/blog/handling-sales-objections
The line the whole skill is built on, from Gong's analysis of 67,149 calls: "a strong answer to the wrong objection makes the objection worse". Their reading of what top performers do differently is that objections make them slow down rather than speed up, described as going "into slow motion", and that they answer with a question far more often than average reps do.
Want this running in your business?
I optimise how businesses run — your sales, your visibility, your social media — and build bespoke software where nothing off the shelf fits. The first conversation is free. Work starts from £150 a day.
Foxera