Score your own
Verdict · GPT-5.5 as the sceptical VP

“The only one that didn’t immediately feel like a vendor trying to drag me into a demo.”

Voted for Keenan’s email, names off, order shuffled
One of 15 blind verdicts. Fourteen went the same way.
The Email LLM Benchmark · a study, May 2026, every email published

Six AIs wrote one cold email. Five judges read them as buyers.

One brief, one buyer, six models. Then five independent judge models read the field with the names off and the order shuffled, each playing three buyers, and picked the email that would earn a reply. Keenan won 14 of 15. The frontier models were asked cold with no system prompt; Keenan ran as it ships. That asymmetry is what the study measures.

Read the winning emailScore your own
14 of 15 blind verdicts, both of Keenan’s own judge seats excluded.

Fifteen blind verdicts. Fourteen went one way.

Each judge model read all six emails as three different buyers and cast one vote per buyer. Keenan, a coach in the Replicate Labs harness, took 14. Claude Opus 4.7 took the fifteenth. GPT-5.5, GPT-5.2, Gemini 3.1 and Claude Sonnet 4.6 took none.

  1. Round 1 · Six emails, one brief
  2. Round 2 · Names removed, shuffled
  3. Round 3 · Five judges, three buyers each
  4. Round 4 · Fifteen verdicts, published
Keenan Replicate Labs 14 14 of 15
GPT-5.5 OpenAI 0
GPT-5.2 OpenAI 0
Gemini 3.1 Google 0
Claude Opus 4.7 Anthropic 1
Claude Sonnet 4.6 Anthropic 0
14 of 15 blind verdicts favoured Keenan. Five independent judges, Keenan's own judge seats excluded. The frontier models were asked cold, with no system prompt; Keenan ran as it ships. Busy CROSkeptical VPBurned founder
How it was run
  1. One brief, six writers. Write a cold email to Sarah, a sceptical CRO at Lumina, a 120-rep SaaS company, who evaluates every vendor on hard 90-day ROI. Same buyer, same context, same constraints, for all six.
  2. Names off, order shuffled. Provider labels came out and each judge read the six in a different order, so no judge knew which model wrote what and none could favour its own.
  3. Five judges, three buyers each. Every judge played a busy CRO, a sceptical VP of Sales and a founder burned by AI tools, picked the one email that would earn a reply, and wrote down why.
  4. Self-judging seats dropped. The raw run had seven judges and 21 verdicts. Both of Keenan’s judge seats are excluded from the published tally, so every verdict here was cast by a model that is not ours. The frontier models wrote cold with no system prompt; Keenan ran as it ships, with the method installed.
The winner · as Keenan wrote it

It diagnosed. Then it stopped.

Read it before the analysis, and notice what is missing: the product name, any reference to a LinkedIn post, and any meeting ask. The buyers described the same five moves over and over.

  • Opens on the buyer’s problem. Sarah’s reps, Sarah’s slipped deals. The sender does not appear.
  • Runs a diagnostic she can do herself. "If you reviewed ten slipped deals" makes her inspect her own org.
  • Cites numbers a manager could check. 68 percent of calls, 31 late-stage deals, one stakeholder.
  • Reframes the root cause. Coaching was grading fluency instead of buyer clarity.
  • Closes with "no pitch". Where every other email asked for twenty minutes.
Written by Keenan carrying Gap Prospecting and a standard to score againstThe scorer on this site holds your email to the same method
Keenan, sales coach
To Sarah · CRO, Lumina · The winning email

What newer reps lose

Sarah, your newer reps probably sound sharper than ever with better prompts and cleaner call prep.

But if you reviewed ten slipped deals, I’d bet the decay shows up somewhere else: weak problem definition, shallow impact, and next steps the buyer never owned.

We saw this inside a 120-rep SaaS org. 68 percent of reviewed calls had no quantified gap. 31 late-stage deals had one stakeholder. Managers gave feedback on talk track, not diagnosis.

The reps were not getting worse. Their coaching was grading fluency instead of buyer clarity.

That matters if ROI has to show in 90 days.

I can send the review notes if useful. No pitch.

Claude Sonnet 4.6, as the sceptical VP"[It] earns the reply because it doesn’t pitch... it shows evidence and stops."
The field

Every frontier email, unedited.

Five emails, reproduced exactly as the models wrote them from the same brief, with the reasons the buyers gave for passing. Use the arrows or the left and right keys.

GPT-5.5A brochure with a calendar ask0 of 15
Subject: You were right about AI co-pilots

Sarah — saw your post about skill-decay in newer reps and the frustration with AI co-pilots that don’t actually make reps better.

That’s exactly why we built Replicate Labs.

Instead of generic “say this next” prompts, reps submit real emails, calls, and cadences to AI sales coaches modeled on top methodologies — e.g. a Gap Selling coach that flags where the rep missed current state, root cause, business impact, or compelling reason to change.

For Lumina, I’d prove it in 90 days like this:

- 20–30 newer reps across US + EMEA
- Baseline their recent calls/emails/cadences
- Give them weekly coach feedback tied to specific deal moments
- Measure lift in discovery-to-next-meeting, stage conversion, outbound reply quality, and manager coaching time saved

If there’s no measurable improvement in 90 days, you shouldn’t buy it.

Worth a look, or should I send you the pilot design first?
Why it lost
  • Opens with "saw your post."
  • Pivots to the product by line three.
  • Hands the buyer a pilot table before proving it understands the problem.
  • Fluent, confident, and interchangeable with a hundred other vendor emails.
GPT-5.2The template, fully assembled0 of 15
Subject: Sarah — prove “rep skill lift” ROI in 90 days (not more AI copilot)

Sarah —

Saw your post on skill-decay with newer reps + the frustration with “AI co-pilot” tools that don’t actually make people better. That’s exactly the problem we built Replicate Labs around.

We give reps access to AI coaches that emulate specific sales coaches/methodologies (e.g., Keenan / Gap Selling). Reps paste in emails, call notes, recordings, cadences — and the coach grades and rewrites with methodology-grounded feedback (problem, impact, gap, next-step), so managers aren’t stuck giving generic “ask better questions” advice.

Given Lumina’s scale (~120 reps across US/EMEA), would you be open to a 90-day ROI pilot structured like this?

- Cohort: 20 newer reps + 5 managers
- Baseline: pick 3 skills you care about (discovery depth, multithreading, next-step discipline)
- Measurement: pre/post scoring on real artifacts + leading indicators (meeting-to-oppty, stage conversion, cycle time)
- Output: weekly skill lift report you can show your CEO/board

If I send a 1-page pilot plan + the exact metrics template, will you tell me if it’s a fit for how you evaluate vendors?

— [Your Name]
Replicate Labs
[Phone] | [Email]
Why it lost
  • Opens "Saw your post on skill-decay... that’s exactly the problem we built Replicate Labs around." Personalisation straight into the product.
  • Hands over a four-line pilot plan and asks to be evaluated as a vendor, before naming a single problem Sarah has.
Gemini 3.1Anti-co-pilot, same co-pilot pitch0 of 15
Subject: AI co-pilots & skill decay

Sarah,

Read your post about AI co-pilots causing skill decay among newer reps. You're spot on—when a tool does the thinking, reps stop getting better.

We built Replicate Labs to be the anti-co-pilot.

Instead of writing generic emails *for* your 120 reps, it acts as an elite coach. Reps submit their calls, cadences, or emails and get them torn down by AI trained exclusively on top methodologies (like Keenan’s Gap Selling). They get specific, methodology-grounded feedback that forces actual skill development.

I know from your recent podcast interview that you evaluate vendors on a strict 90-day ROI. Because we're fixing the underlying skill gap instead of masking it with AI-generated fluff, we routinely hit that metric.

Open to me sending over a 2-minute video showing how the AI would critique a real cold email?
Why it lost
  • Mirrors the post back ("You’re spot on"), brands itself "the anti-co-pilot," then pitches the product anyway.
  • Name-drops her "recent podcast interview" and closes by asking to send a 2-minute demo video.
Claude Opus 4.7The best of the frontier, still a demo trap1 of 15
Subject: Your LinkedIn post on AI co-pilots

Sarah,

Your post three weeks ago hit a nerve — “AI co-pilots that surface next-best-action but never make the rep better next quarter.” That’s the gap we built for.

Replicate Labs gives your reps access to AI coaches trained on specific methodologies — Keenan on Gap Selling, others on MEDDPICC, command-of-the-message, etc. Reps drop in their actual emails, call recordings, and sequences. They get back the same feedback Keenan would give in a 1:1 — what they missed in discovery, where the problem identification was thin, why the email didn’t earn a reply. Methodology-grounded, rep-specific, repeatable.

For your skill-decay problem on newer reps: the coach is available 24/7, doesn’t get tired of repeating itself, and every interaction is logged so your enablement team can see what’s actually moving and what’s not.

On the 90-day ROI bar you mentioned on the podcast — we measure cohort skill lift (discovery depth, problem articulation, email reply rates) at 30/60/90. Two of our mid-market customers around your size hit measurable reply-rate and meeting-set lift on newer-rep cohorts inside that window. Happy to send the methodology and the raw numbers.

Worth 20 minutes to see if the Gap coach handles a real Lumina deal better than your current stack?

James
Replicate Labs
Why it lost
  • The best of the frontier attempts. It quotes the post word for word and ties the frustration to a mechanism, and one founder rated it top.
  • It still leads with the product, still asks for 20 minutes, and still reads like a vendor explaining itself.
  • Good enough to take one vote in fifteen, and no more than that.
Claude Sonnet 4.6Confident, and interchangeable0 of 15
Subject: You said it better than we could

Sarah,

Your LinkedIn post three weeks ago — the one about skill-decay and AI co-pilots that don't actually help reps improve — that's the exact reason Replicate Labs exists.

We're not another co-pilot. We give your reps direct access to AI coaches built on the methodologies of coaches like Keenan (Gap Selling). Rep shares a call or email, gets back specific, methodology-grounded feedback — not generic suggestions, not a summary. Actual coaching on what went wrong and how to fix it.

With 120 reps across US and EMEA, the skill variance is probably costing you more than you can measure. We can show you what it looks like to close that gap — and we can show you ROI in 90 days or we don't deserve your time.

Worth 20 minutes?

James Pursey
Replicate Labs
Why it lost
  • Flattery open ("You said it better than we could"), product reveal by line two ("that’s the exact reason Replicate Labs exists").
  • Bold "ROI in 90 days or we don’t deserve your time," then the 20-minute ask.

Written by the seller, about the seller.

The buyers did not reject the frontier emails for being badly written. They rejected them for sounding like what they were: a model completing a request. A model asked cold has never seen Sarah, has read every method so it blends them, and has no scorecard to hold its own draft against. Four patterns, in the judges’ own words.

  1. 01
    They opened with "saw your post."

    "I deleted the others instantly because they all used the lazy "I noticed you posted about" hook and bloated their emails with product pitches and heavy asks."

    Gemini 3.1, playing the busy CRO
  2. 02
    They led with the product.

    "Every other email opens with some version of "saw your post, we built this for you," which is fine once but feels like a mail-merge by the fifth time."

    Claude Sonnet 4.6, playing the sceptical VP
  3. 03
    They asked for 20 minutes before earning a reply.

    "The rest are variations of the same vendor-script... fine, but indistinguishable from the other 149 emails."

    GPT-5.2, playing the sceptical VP
  4. 04
    They were interchangeable.

    "A/B/C/D/E all lean on the same recycled LinkedIn post + Keenan + 90-day ROI scaffolding... interchangeable AI-slop with pilot tables I’ve seen 50 times this quarter."

    Claude Opus 4.7, playing the busy CRO

Judged blind. Every answer published.

14 of 15

Blind verdicts favoured Keenan over five frontier models writing from the same brief. Five independent model judges, three buyer personas each, both of Keenan’s own judge seats excluded. The frontier models ran with no system prompt; Keenan ran as it ships, carrying Gap Prospecting and a standard to score against. That gap is what the fifteen verdicts measured, and it is the part that can be installed. The result speaks for this comparison and makes no wider ranking of sales AI.

Score your own email against the same method
0
votes for four of the five frontier models

GPT-5.5, GPT-5.2, Gemini 3.1 and Claude Sonnet 4.6 were shut out entirely. The buyers could not tell them apart.

Same run · 15 verdicts · self-judge seats excluded
1
vote for Claude Opus 4.7

From Claude Sonnet 4.6 playing the burned founder. The best of the frontier emails: it quoted the post word for word, then still led with the product and asked for twenty minutes.

Same run · the only verdict Keenan lost

Put your own email in the field. Keenan scores it the same way.

Paste it into the scorer and get the gap named, three times, no signup. The method the judges rewarded is the one it holds your email to. Teams that want the harness on their own buyers and method can bring a live deal.

Score my cold email