Skip to main content

Customizing Your Prompts in Skyp for Better Email Performance

Learn how to adjust campaign-level prompts, global prompts, guardrails, and models in Skyp to get better, more on-brand email performance.

Alexander Shartsis avatar
Written by Alexander Shartsis
Updated over a month ago

Why customizing prompts matters

Skyp’s default prompts and models are tuned to work well out of the box.

But as you learn more about what works for your audience, you may want tighter control over:

  • How subject lines are framed

  • How the body of the email is structured

  • What the goal of each campaign really is

  • What is never allowed to go out (stylistically or legally)

Skyp gives you three main layers of control:

  1. Campaign-level prompts

  2. Global (account-wide) prompt

  3. Guardrails + model selection


1. Editing prompts inside a specific campaign

Use this when you want to adjust how one particular campaign behaves.

Where to do it

  1. Go to Campaigns.

  2. Open the campaign you want to adjust.

  3. Navigate to the page where you can see the generated emails.

On this screen, you can edit everything that affects the AI for this campaign only.

What you can edit

You can customize:

  • Subject prompt – how Skyp should think about writing subject lines.

  • Body prompt – how Skyp should structure and tone the email body.

  • Goal – the underlying intent of the campaign (e.g. “book meetings”, “get replies”, “drive webinar registrations”).

The Goal field will pull in the goal you originally set during setup so you can literally revise it instead of starting from scratch. Useful if the emails aren’t quite doing what you wanted.

Extras on this screen

While you’re here, you can also:

  • Edit the signature used in the emails

  • Send a test sequence to see how the emails look and behave in your inbox

This layer is ideal when:

  • One campaign is underperforming

  • You’ve changed your mind about the strategy for that specific motion

  • You want to experiment without impacting your whole account


2. Tuning your global prompt (advanced)

For more advanced users, Skyp allows you to customize a global prompt in your account settings.

Where to find it

  1. Go to Settings.

  2. Look for the global prompt / AI configuration area (advanced section).

What it controls

The global prompt:

  • Applies to every email sent from this Skyp account

  • Defines the base instructions the AI follows across campaigns

You can customize it however you like to:

  • Reflect your company’s philosophy on outreach

  • Set overall rules for structure and tone

  • Shape the “personality” of your emails at a system level

Important caveats

With more control comes more room to break things:

  • Skyp uses CTAs and internal logic behind the scenes.

  • Over-editing the global prompt can unintentionally affect how those CTAs and flows perform.

If something feels off after changes:

  • You can always click Reset to default.

About resetting to default

  • Skyp periodically improves the base master prompt as it learns.

  • Your account will never be changed automatically.

  • When you click Reset to default, you pull in the latest master prompt from Skyp’s systems.

This gives you:

  • A safe experimentation space

  • A guaranteed “known good” baseline you can always return to


3. Using guardrails for style and compliance (very advanced)

Guardrails are an extra layer on top of prompts, ideal for teams with strict style or regulatory requirements.

What guardrails do

Guardrails:

  • Read every email before it goes out

  • Enforce rules you specify

  • Act as a final validator to catch anything that shouldn’t be sent

Examples of stylistic guardrails

You can define rules like:

  • “Don’t use em dashes.”

  • “Don’t use any kind of dash.”

  • “Avoid exclamation marks.”

Some of these constraints can live in your main prompt, but guardrails double-check them on the outgoing text.

Examples of compliance guardrails

If you’re in a regulated industry (e.g. healthcare), you can use guardrails to enforce rules such as:

  • “Do not make any healthcare efficacy claims.”

  • “Do not reference outcomes or cures.”

The more precise your guardrail instructions, the more effective they’ll be.

Guardrails are especially useful when:

  • Legal/compliance has a list of “never say this” patterns

  • You need to ensure every email meets policy, even if prompts change


4. Choosing models for your workflows

Skyp lets you choose which AI models to use for different tasks, with defaults chosen to work well for most teams.

Default behaviors and options

  • Claude is great for editing:

    • It tends to adjust what needs changing instead of rewriting everything.

    • Ideal when you want small, precise edits rather than wholesale rewrites.

  • GPT-4o mini is:

    • Very predictable

    • Strong at writing good, consistent emails

You’ll see a variety of model choices you can fine-tune as you get more comfortable.

Suggested approach

  • Start with the defaults – they’re tuned to perform well for typical outbound use cases.

  • Once you’ve run real campaigns and know what you want to change, experiment with:

    • Different models

    • Small prompt tweaks

    • Combined use of prompts + guardrails

Did this answer your question?