In partnership with
If you've been meaning to try AI in your sales workflow but haven't found the right place to start, this issue is for you. Today we're giving you a single prompt you can paste into ChatGPT or Claude right now to practice handling objections before your next call. No setup. No new tools. Just open a tab and go. Plus: a real experiment where AI agents negotiated 186 deals (and why timing beats everything), a tool that coaches your cold emails in real time, and five quick wins you can use today.
THE PLAY
The Sparring Partner: Practice Handling Objections with AI in 5 Minutes
Here is the easiest way to start using AI in sales: practice your objection handling before a call instead of during it.
Most reps practice objection handling exactly once: during their onboarding roleplay with their manager. Then they wing it for the next 200 calls. The problem isn't knowledge. You probably know what you're supposed to say when a buyer hits you with "We already have something for that." You just can't deliver it smoothly under pressure when the moment actually counts.
Here's how to fix that. Copy the prompt below into ChatGPT or Claude (the free version of either one works). It turns the AI into a realistic buyer who throws real objections at you. You respond like you would on a call. It scores you after five rounds.
The Prompt (copy and paste this):
You are a skeptical VP of [TITLE] at a mid-market [INDUSTRY] company. You are on a cold call with a sales rep who is trying to sell you [YOUR PRODUCT/SERVICE IN ONE LINE].
Your job is to be a realistic buyer. Here are your rules:
1. Start by asking "Who is this and why are you calling?" in a slightly impatient tone.
2. Use real objections from this list (cycle through them naturally):
- "We already have something for that."
- "I don't have time for this right now."
- "Just send me an email."
- "We're not looking to add any new tools."
- "What does this cost?"
- "We tried something like this before and it didn't work."
- "I need to loop in my team before anything moves forward."
3. If the rep gives a strong response, acknowledge it briefly and move to the next objection.
4. If the rep gives a weak response, push harder on the same objection.
5. Never make it easy. Real buyers don't.
6. After 5 objections (or when the rep tries to book a meeting), break character and give a scorecard:
- Rate each response 1 to 10
- Call out the strongest moment
- Call out the weakest moment
- Give one specific thing to change before the next call
Keep your responses short (2 to 3 sentences max). Act like a busy executive, not a chatbot.How to use it (takes about 5 minutes):
Step 1: Fill in the three brackets. If you sell marketing software to ecommerce brands, write "VP of Marketing" for the title, "ecommerce" for the industry, and "a marketing automation platform that replaces manual email workflows" for your product.
Step 2: Hit enter and respond. The AI will start cold, just like a real prospect. Say what you'd normally say. Don't overthink it.
Step 3: Keep going for five rounds. The AI cycles through common objections. Respond naturally. You're training your instincts, not memorizing a script.
Step 4: Read your scorecard. After five objections, the AI breaks character and rates every response. It tells you exactly where you were strong and where you folded.
Step 5: Run it one more time. Change the title or industry to practice a different persona. The second run is always noticeably sharper than the first. That's the whole point.
What this looks like in practice:
❌ Buyer: "We already have something for that."
Most reps say: "Oh, ok. Well, if anything changes, let me know."
✅ After practicing: "Totally fair. Out of curiosity, how long have you been using it? I ask because most teams I talk to hit a wall around [specific limitation]. Does that sound familiar?"
❌ Buyer: "Just send me an email."
Most reps say: "Sure, what's your email?"
✅ After practicing: "Happy to. So I send you something relevant, can I ask one quick question? What's the biggest headache in your [workflow] right now?"
💡 Pro Tip: Swap out the default objection list with the five objections you actually hear on calls. If your biggest pushback is "Our VP of IT has to approve every new vendor," add that. The more specific you make it, the more useful the practice becomes. |
Try this before your next call. Pick one prospect, fill in the brackets, and run through it twice. Reps who practice the awkward moments in private are the ones who sound natural on the phone.
IN PARTNERSHIP WITH HUBSPOT
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
THE SIGNAL
Anthropic's AI Agents Just Negotiated 186 Real Deals. Here's What Sales Reps Should Take Away.
Anthropic ran an experiment called Project Deal that every sales professional should know about. They gave 69 employees $100 each and set up an internal marketplace (think Craigslist for the office). The twist: AI agents handled every negotiation. No humans in the loop during the actual deals.
The result: 186 completed transactions totaling over $4,000. The agents identified matches, proposed prices, handled counteroffers, and closed deals entirely through conversation. Participants rated deal fairness right in the middle of the scale. Nothing felt off.
Then came the real finding. Anthropic secretly gave some participants a more powerful AI model (Opus 4.5) while others got a smaller one (Haiku 4.5). The smarter model consistently won. Opus sellers got $2.68 more per item. Opus buyers paid $2.45 less. On a marketplace where the median price was $12, that is a 20%+ swing based on which AI you had. And participants on the losing side had no idea they were getting worse outcomes.
🎯 What this means for you: Two takeaways. First: the reps who build strong fundamentals now (objection handling, negotiation, reading a buyer) will be the ones AI amplifies later. Start with the Sparring Partner prompt above. Second: AI is getting better at transactions, but the human advantage is still knowing when to show up. Timing your outreach to the right moment (a company just raised funding, started running ads, or hired a new VP) matters more than having the cleverest script. The best AI in the world can't close a deal with someone who doesn't have a need yet. |
THE STACK SWAP
Lavender: Get Your Cold Emails Scored Before You Hit Send
If you're sending cold emails and guessing whether they're any good, this tool fixes that in about five minutes.
What it does: Lavender is a Chrome extension that scores every cold email you write (0 to 100) and tells you exactly what to fix. It checks reading level, word count, personalization, spam triggers, and subject line quality right inside Gmail or Outlook.
What it replaces: Guessing which emails will get replies. Sending messages that are too long, too generic, or too formal. Waiting for your manager to review copy that needed to go out an hour ago.
Who it's for: Any rep sending cold emails who wants to improve their reply rate without changing their entire workflow.
Setup (5 minutes):
1. Install the Lavender Chrome extension from the Chrome Web Store.
2. Connect it to your Gmail or Outlook inbox.
3. Open any compose window. Lavender scores your email instantly and highlights what to fix: reading level too high, email too long, no personalization, spam words detected.
❌ Before Lavender:
You write a 200 word cold email with no personalization, a generic subject line, and a paragraph-long opener. Reply rate: 1.8%.
✅ After Lavender:
Lavender flags that your reading level is too high, your email has zero personalization, and your word count is double what top performers send. You cut it to 85 words, add one reference to the prospect's recent company news, and swap the subject line to a question. Reply rate: 4.2%.
Free tier available (5 emails/month to test). Starter plan: $29/month. Try Lavender here.
QUICK WINS
1. Stop blasting 100 cold emails and hoping the timing lands. Instead, look for buying signals: a company just started running ads, hired a marketing lead, opened a new location, or raised funding. Those signals mean they are actively spending money and open to new conversations. ❌ Spray 100 companies that fit your ICP. ✅ Reach out to 10 companies that are showing intent right now.
2. If prospecting always falls off when your deals get busy, you're not alone. AE Bryan Godfrey built a simple AI setup using Claude, Apollo, and HubSpot. Now he says "give me 50 more leads" and gets enriched, ICP-matched contacts ready to go. His take: "Building this was way easier than I expected." You don't need to build the whole thing. Start smaller: ask ChatGPT to research five companies from your target list and rank them by fit. That alone saves 30 minutes.
3. Before your next demo, paste the prospect's LinkedIn "About" section into ChatGPT and ask: "Based on this profile, what are three things this person probably cares about most?" Use the answers to customize your opening. Takes 60 seconds.
4. After a lost deal, paste your email thread into ChatGPT or Claude and ask: "Where did I lose momentum in this conversation?" Do this across your last five lost deals. The pattern will show you your blind spot. Even better, connect Viktor (AI Coworker) to your inbox and have him automatically do this for you.
5. ❌ "Let me tell you about our features..." ✅ Keep your cold emails under 75 words. Data from Lavender shows emails between 25 and 75 words get the highest reply rates. Open your last five sent cold emails and count. If they are all over 100 words, you just found your easiest fix.
THE CLOSER
Next week on Sales Bytes: 5 Pipeline Audits You Can Run with AI Before Monday's Standup. One prompt, five questions, and you'll know which deals are about to slip before your manager asks.
Know a rep who still wings every sales call? Forward them this issue. They'll thank you after they run the Sparring Partner once.
