Most cold DMs on Instagram get ignored because they look exactly like every other cold DM. The templates that convert share three traits: they reference something specific about the recipient, they offer value before asking for anything, and they have clear next steps. Here are the patterns that actually work.
Templates like 'Hey! Loved your content, wanted to connect' have near-zero reply rates because the recipient sees 50 of them a week. Instagram also detects and down-ranks identical DM content sent to multiple users. Every template you use needs variation — different openers, different value props, different CTAs — or it becomes invisible.
Part 1 (context hook): reference something specific about the recipient — a recent post, their bio, a specific piece of content. Part 2 (value offer): give them something useful without asking for anything yet — a relevant insight, a useful link, a specific observation. Part 3 (soft CTA): invite a reply with a question, not a sales pitch. This structure outperforms hard-sell DMs 4-5x.
Viari's AI reply engine takes one template and generates hundreds of variations — same structure, different words, different angles. Spintax variables like {greeting} and {compliment} mean every DM is unique. For a template designed around fitness coaching, AI can produce 500 variations in minutes, all different enough to avoid duplicate content detection.
Unvaried generic templates: 1-2% reply rate. Hand-written personalized DMs: 15-25% reply rate. AI-varied templates with specific hooks: 8-15% reply rate. AI-varied with Viari's referencing features (pulling the recipient's recent post into the DM): 12-20% reply rate. The gap between bad and good is a 10x difference in pipeline.
DM templates aren't just scripts — they're frameworks with structured personalization. Specificity beats volume, and AI variation is how you get both.