Drafts built from your conversations, files, and notes

On a busy day, you can face a 40-page request for proposal that is due tomorrow, a 90-minute call recording that the team wants in two paragraphs, and a 200-message thread that an executive needs to catch up on before they join it. You can end up with time to read only one of them closely.

When you ask Beryl to summarize a file, a thread, or a call recording, it returns the decisions, open questions, next actions, and commitments as separate lists. Each bullet links to the page, message, or timestamp that it comes from, so you can check any point that the team asks about.

Your contacts can write to you in French from Abidjan, in English from Lagos, and in Portuguese from Luanda, while your team writes in one of those languages. For every message in the other two languages, your team will have to use a translator or write it themselves in a language that they use less.

Beryl picks the language for each draft from your last thread with that contact. It also applies the local conventions for tone, greeting, and sign-off, so a message to a contact in Abidjan opens and closes the way that a French-speaking colleague there would write it.

Your Channels reports can show that three subject lines from last quarter earn 40% more opens than the rest. Unless someone works out what those three lines have in common, the campaign team will have to guess when they write the next campaign.

Beryl compares the open and reply rates of your earlier Channels campaigns, then drafts subject lines, body copy, and calls to action modeled on the campaigns that your audience responds to most. Each draft links to the earlier campaign that Beryl bases it on, so the team can see why Beryl suggests each line.