The Messaging tab is where you teach Ava what to say. While the Sequence tab defines the cadence (when and through which channels), the Messaging tab defines the content. Ava doesn’t use all configured components, she selects the most relevant combination for each prospect based on their role, company, and available personalization data.Documentation Index
Fetch the complete documentation index at: https://support.artisan.co/llms.txt
Use this file to discover all available pages before exploring further.
Give Ava context
The first thing Ava needs is background on the company and product. This section accept:- any URL (company website, docs, blog posts, case studies, etc.). We recommend adding at least your website to get started.
- any uploaded files (PDFs, playbooks, one-pagers, sales decks, etc.)
- Ava analyzes the provided links and files to fill the message content.
- This gives users a starting point they can then refine, rather than writing everything from scratch.
Offering & Features
This is the core pitch. It has two parts: One-sentence offering: a concise description of what the company does (e.g., “ACME sets up local entities through which companies can scale operations across 100+ countries”). This anchors every email Ava writes. Pain point / solution pairs: the specific problems the product solves and how.- Each pair has a pain point (the prospect’s problem) and a solution (how the product addresses it).
- Users can add as many pairs as they want.
- Ava selects the single most relevant pair per prospect based on their role, industry, and context. This is how Ava tailors the core pitch to different buyer personas without the user needing to write separate emails.
Social Proof
Credibility signals that Ava can reference to build trust. Three types: Social proof: short, punchy credibility building statements (e.g., “Helped Brex close 14 customers in 5 minutes”). Customers: named company references with the option to add coaching on when to use them (e.g., “Stripe: only reference when lead is in fintech”). This allows you to guide Ava on how she should mention the customers. Case studies: longer form success stories with optional coaching that the user can add to guide Ava on how best to use them. There’s also a global social proof toggle that controls whether Ava can pull from org-wide proof points or is limited to campaign-specific ones.Personalizations
Personalization is what makes Ava’s outreach feel human rather than templated. It’s the message component that makes the prospect think “this person actually researched me.” Three tiers control how deep Ava goes: Lite (3 credits / lead)- Ava picks one relevant personalization per lead from a set of enabled sources. She uses her own judgment to decide which source is most relevant and how to frame it.
- There’s no way to coach individual sources at this tier.
- Users can only toggle sources on or off. Lite is fast and cheap.
- Per-source coaching lets users attach short instructions or coaching to any individual personalization source, telling Ava how to use that specific data point
- e.g., “Don’t mention if the school isn’t well-known” on Mutual school. This is how users fine-tune Ava’s personalization behavior at a granular level.
- Deep is the recommended tier for most campaigns.
- Users get access to using fields from their CRM or any other uploaded sources like CSV / webhook.
- You can added any number of fields and Ava uses them to determine the best, highest fit personalization.
- These aren’t generic web searches, each agent formulates its own research question based on what it knows about the prospect and the campaign context
- Examples include: “What did the lead mention in their recent earnings report?” or “Has this company’s leadership spoken publicly about their priorities for this year?”
- Research goal: What should this agent look for (e.g., “Find recent podcast episodes, conference talks, or panel appearances”). The more specific the research goal, the better the output.
- Coaching points: that tell Ava how to use what the agent finds in the outreach message (e.g., “Mention how you enjoyed their recent talk and connect it to a shared challenge in the space”)
- Optional suggested sources that restrict where the agent searches (e.g., “techcrunch.com, youtube.com”. Leave empty to search the entire web).
Message Coaching
Coaching points are writing rules that Ava follows when composing messages. Each coaching point has a content field (the instruction itself) and a coaching scope section that controls where the coaching applies. The scope has three dimensions:- Channel: which channels this coaching applies to, Email, LinkedIn, or both.
- Sequence steps: which parts of the sequence this coaching applies to.
- Message components: You can granularly coach specific parts of the message too (such as subject, opening, body, etc.)
CTAs
Calls to action are the closing ask in each message. Unlike other messaging ingredients, CTAs are used verbatim, Ava doesn’t paraphrase them. This gives users precise control over the ask. Each CTA can be scoped to apply to specific channels: all channels, emails only, or LinkedIn messages only.Language
The language Ava writes in. A single dropdown with the supported language list (the default is English). We support 40 languages. The choice applies to every AI-generated touch in the campaign, on both Email and LinkedIn. Manual email steps are not affected, since those are sent verbatim.Advanced
Message styling
- Tone of voice: This defines Ava’s writing style and personality.
- Length of messages: The target length of the body of the email.
- Last message’s behavior: What the final email in the sequence does.