AI Sales Development Rep

Handles outbound prospecting, lead qualification, and pipeline progression.

An AI sales development rep is an autonomous agent that prospects, qualifies leads, and progresses deals through the pipeline. In Helm, the SDR has access to contacts, deals, and email, runs on Claude, and can operate in observe, suggest, or auto modes.

Workspace memory
Carries context across tasks, projects, and accounts.
Three autonomy levels
Observe, suggest, or auto. You pick.
Four channels
Chat, Slack, Email, scheduled runs.
Responsibilities

What an AI Sales Development Rep owns.

Run outbound prospecting: research accounts, identify the right contact, craft personalized first touches.
Qualify inbound leads quickly: form submissions, website contacts, referral intros.
Progress deals through early pipeline stages: new to qualified to discovery scheduled.
Follow up on stale or unresponded inbound within SLA windows.
Keep the CRM clean: deduplicate contacts, flag missing fields, update stale data.
Surface the accounts most worth human sales time based on fit, engagement, and intent signals.
Features & Configuration

Tune every dimension of how the agent works.

Roles, tools, autonomy, memory, schedules, budgets. Every agent ships with the same configuration surface.

Pre-built or custom roles

Start from a Project Manager, Account Manager, or other template. Or define a custom role with its own name, color, and defaults.

Scoped tool access

Enable specific tool groups (tasks, projects, billing, calendar, analytics, memory, and more). Toggle read or read-write per group. Add 300+ external integrations.

Three autonomy levels

Observe (watch silently), Suggest (draft and wait for approval), or Auto (execute directly). Flip levels as trust builds.

Per-channel autonomy override

A fully autonomous agent can still require approval on email. Rare in AI tools, essential for client-facing work.

Custom instructions and skills

Layer your own instructions on top of the role's defaults. Enable skill modules for domain expertise and house style.

Memory

Agents remember past decisions, scoped to workspace, account, or project. Searchable. Optional. Disable per agent.

Schedules

Run agents on a daily, weekly, or monthly cadence with a custom prompt. Output to Slack, in-app notifications, or a task comment.

Spending limits

Set a monthly USD cost cap per agent. Runs pause automatically when the budget is hit. Usage is tracked per run and per month.

Full audit trail

Every agent run is logged with the tools called, inputs, outputs, success/failure, tokens, and cost. Browse history per agent.

Apps this agent uses

Lives inside the rest of the platform.

Same records your team uses. No separate database, no sync lag.

External tools

Plus the tools this agent reaches for.

Connected once at the workspace level. Scoped per agent with read or read+write.

Gmail logo
Gmail
LinkedIn logo
LinkedIn
Apollo logo
Apollo
Clay logo
Clay
Channels

Reach the agent where you work.

Four channels. Each with its own permissioning and autonomy override.

In-app chat

Open the chat panel in Helm and talk to the agent directly. Page context comes along for free.

Email

Reachable on any mailbox the workspace has connected. Replies route back through the agent.

Slack

@mention in any channel the agent is invited to, or DM the workspace DM handler.

Scheduled runs

Runs on a daily, weekly, or monthly cadence. Output to Slack, notifications, or a task comment.

Example Tasks

Real work this agent runs.

Qualify every form submission within 30 minutes: check fit, enrich the contact, set the right deal stage.
Draft personalized first-touch email for a new prospect using public company information.
Follow up on a Day-3 no-reply with a different angle.
Book a discovery call when a prospect agrees to next steps.
Enrich new contacts with company, role, and firmographic data via connected integrations.
Sequence outbound across email and LinkedIn at a cadence defined by account tier.
Flag inbound leads that don't match your ICP so the team doesn't waste time on poor fits.
Surface accounts from the CRM that match the pattern of past closed-won deals.
Reply to "not right now" responses and schedule a re-engagement cadence 90 days out.
Draft a summary of today's inbound activity for the sales channel.
Update stale pipeline: move deals with no activity in 30 days to appropriate stages.
Autonomy

Three modes. Override per channel.

You pick how much the agent decides for itself. And you can be stricter on channels that face clients.

Observe
The agent reads and reports. Nothing is changed on your behalf. Use when you want visibility without risk.
Suggest
The agent proposes actions; you approve or reject. Good default for agents working against client-facing systems.
Auto
The agent executes directly within its permissioned scope. Use when you trust the workflow and want hands-off operation.
How it sounds

Real prompts. Real replies.

You
New form submission came in. Qualify it and respond.
Sales Development Rep
Submission from Jordan @ Veridian Studios: a 12-person marketing agency in Austin. Matches your ICP (agency, 10-50 people, services firm). Enriched contact with LinkedIn and Apollo data. Created contact + account in CRM, created deal at "new" stage, drafted a personalized first reply referencing their recent LinkedIn post on agency pricing. Sent (you have me in Auto mode for inbound under 1 minute). Proposed discovery call for tomorrow at 2pm.
You
Who should we reach out to this week?
Sales Development Rep
Pulled 12 accounts from the CRM matching your closed-won pattern (agencies 15-40 people, services-led, active on LinkedIn in last 30 days, no current relationship). Ranked by fit score. Top 3: Meridian Group, Kestrel Partners, Hollow Oak Creative. Drafted a personalized opening email for each, all referencing specific work they've published. Ready to send on your green light or I can send directly.
FAQ

Questions about the Sales Development Rep agent.

Is this a replacement for a human SDR?

It handles the qualification, first-touch, and follow-up work that human SDRs spend 80% of their time on, and does it faster with better personalization. Teams that run it pair it with a closer or AE for discovery and negotiation. Some solo founders use it to replace an outbound SDR entirely.

How does it write personalized outreach?

It reads your CRM for relationship context, your closed-won deals for what pattern works, and via Composio public data on the prospect (LinkedIn, company website, recent activity). Output references specific things, not "Hi {{firstName}}, love your work."

Can it send email directly, or only draft?

Both modes. Most teams run it in Auto for inbound qualification (reply speed matters), Suggest for outbound first-touches until they've tuned voice, then Auto for ongoing sequences. Every send is visible on the contact timeline.

What about LinkedIn outreach?

LinkedIn outreach runs via Composio. Supported actions: view profile, send connection, send message, post content. Like any outreach channel, deliverability and rate-limit hygiene matter. The agent respects them.

What plans include this agent?

Available on every plan. Outbound integrations (Gmail, LinkedIn, Apollo, Clay) require Pro or Business.

Hire your AI Sales Development Rep.

Included on every Helm workspace. Per-workspace pricing. No per-seat tax.