Outreach
Icebreaker personalizes your saved message templates for a recipient. It can write a first line if your template is an opener, or a complete message if your template contains the whole outreach message.
You choose the person and the templates, generate, then review the saved messages before sending them yourself.
1. Prepare a template in your Library
Open Library and create a message template with a title and a body. Icebreaker needs at least one saved template before it can generate.
Use square-bracket placeholders for details to personalize, such as [first_name], [company], or [a relevant detail from their profile]. Keep the rest of the wording close to what you would actually send.
For example:
Hi [first_name], [a relevant detail from their profile] caught my attention. How are you handling [a challenge supported by their profile] at [company]?
For help writing the reusable inputs, see Write Better Prompts and Templates for AI Outreach.
2. Choose the recipient
Open Icebreaker. Under Who are you writing to?, choose:
- Saved contact — search by name, company, or LinkedIn URL, then select the person.
- LinkedIn URL — enter a LinkedIn profile URL.
A saved contact with profile details gives the AI more to work with. A URL alone does not fetch the LinkedIn profile in this workflow. Use a contact with captured details or the browser extension on the profile for richer personalization.
3. Choose one template or all templates
Under Template, select a saved template or leave All templates selected to personalize the whole Library in one pass.
Click Generate messages. The messages are saved automatically and appear in the results. Review them and use the copy control when you are ready to send.
Generation does not send a message or complete a campaign step. After sending, log the activity or complete the due step in your queue.
4. Use Advanced options when needed
Open Advanced options to choose a Writing prompt or Target profile. The prompt supplies writing guidance and sender context; the target profile adds audience and positioning context.
Enable Compare models to generate three versions of each selected template using GPT-4.1 Nano, Mini, and Full. Without comparison, Icebreaker uses GPT-4.1 and produces one version per template.
If a model cannot finish all selected messages, the app shows a warning and saves the complete results from the models that succeeded.
5. Use profile context from the extension
The browser extension can send the visible LinkedIn profile details to Icebreaker. This gives the generator more context than a URL alone. Keep templates in your Outreach Library so the generator has messages to personalize.
The extension is maintained separately from this app, so its controls depend on the installed version. You can always use the app workflow above with a saved contact.
6. Review before sending
The generator is instructed to use profile facts and preserve your template's structure. That is guidance, not a guarantee that every sentence is correct.
Check names, company details, claims, and tone. When sender details are missing, the generator may leave square-bracket placeholders for you to fill in. Replace those before copying the final message.
Subscription access
Icebreaker uses your Outreach subscription and does not require a separate API key from you. Generation is subject to the app's rate limit.
The practical routine
Save a template, choose a recipient, generate, review, copy, and send. Then record what happened in Outreach so the next follow-up has context.