How sales reps and SDRs use KlipPack outreach prompts to skip prompt engineering and ship better outbound copy in their existing AI tool, without another subscription.
Last revised May 11, 2026
The cold-outreach world is in a strange place right now. Reply rates are at multi-year lows (Gong's 2025 outbound report tracks the slide), every prospect's inbox is buried, and AI-generated outreach has trained buyers to spot template language inside the first six words. The reflex among many teams is to write more email. That's the wrong move.
The teams I see actually winning outbound right now do three things differently. None of them involve sending more.
They write less, but each one is research-grounded
The dominant pattern I keep seeing on winning sequences: cold emails that open with a specific observation about the prospect's company, not a general flattery line. "Saw your new pricing page lost the contact-sales option last Tuesday, was that on purpose?" beats "I noticed your team is doing impressive work in [industry]."
This isn't novel. Predictable Revenue's playbooks have argued this for a decade. What is new is that AI has finally made per-prospect research fast enough to do at scale. Twenty minutes of human research × 50 prospects used to be a full day. Now it is twenty minutes of model-assisted research × 50 prospects, with the model doing the boilerplate while the human does the judgment.
The structure I use:
Read this prospect's company page, their LinkedIn post from the last 90 days, and their last earnings call (paste here). Pull out three signals a senior account executive would notice. Rank by which is most likely to be a current pain point. Write a single-sentence opening line that references the strongest one without flattery.
The model gives you raw material. You write the line.
They follow up like a human
Most automated sequences die at message three with "Bumping this to the top of your inbox." That message implicitly says: I have nothing new to say but I want a response. Buyers ignore it.
Every follow-up should add value the prospect didn't have before opening it: a relevant data point, a comparison, a link to a specific resource, or a better question than the one in the first email. Sahil Mansuri's writing at Bravado makes this argument well, follow-ups are a chance to trade information, not to nag.
They protect the ask
The biggest mistake I see in AI-assisted outreach is letting the model write the call-to-action. Models default to "Are you free for a quick 15-minute call next week?" which is the most-asked question on planet earth. Replace it with something specific the prospect can answer in two sentences: "Are you the right person to ask about this, or should I be talking to someone else on your team?" That single line outperforms generic meeting-asks consistently in the data I have seen.
Becc Holland's frameworks on cold-call openers translate directly to email CTAs: lower the activation energy of the reply.
A small workflow that scales
A sequence I have run for B2B SaaS teams:
- Research pass (model-assisted, human-edited): pull 3 signals per prospect, score them.
- Open with the strongest signal as a question. No introductions. No "I hope this finds you well."
- Two-sentence body with one concrete claim. A number, a comparison, or a named example.
- Single low-friction CTA. "Right person? Or should I find someone else on your team?"
- Three follow-ups across two weeks, each adding new information.
Sequences built this way land in the 8-12% reply range in my data, against an industry baseline closer to 1-3% per Lavender's 2025 benchmark report.
What AI is actually good for
It's good for the parts of outreach humans hate: scanning ten LinkedIn posts to find a pattern, drafting three variations of a single line so you can pick the strongest, summarizing a 60-page S-1 into the three points relevant to your offer. It is bad, still, at deciding which prospect deserves your time today. That judgment still has to be yours.
If you want the prompt scaffolds I use day to day, our Sales AI Prompt Pack collects them. But the prompts are the easy part. The hard part, the judgment about who to write to and what to actually say, that stays with you.