Best AI sales prompt, role-specific workflows, SDRs, and AEs can run today

"Use AI to supercharge your sales.”

If you’re reading this, you’ve probably seen that line more times than you can count; on landing pages, in your LinkedIn feed, or in yet another “10X your sales conversion rates with AI” article your CEO just dropped in #sales-insights on Slack with the “👀” emoji.

Or worse, you’ve commented “X” under several LinkedIn viral posts promising “50 AI sales prompts that got me 1,000 leads in 7 days.”

Fast-forward to hundreds of prompts and countless failed sequences later, nothing has worked.

The problem isn’t that you “don’t get AI” or that it can’t work for you. The problem is that you’re getting generic sales prompts when what you need are prompts built for real pipelines; contextual, workflow-ready, and safe to plug into your stack without ruining your process.

That’s what this article will give you: usable, role-specific prompts for sales professionals – SDRs, AEs, CSMs, and RevOps, plus a clear framework for feeding AI the right inputs so you can scale your outreach without killing quality or babysitting ChatGPT every time.

Why most AI sales prompts fail: Role-specific workflows sales teams can actually run today.

Most AI sales prompts fail because they don’t fit your workflow, aren’t role-specific, and ignore compliance guardrails. These three issues cause execution to slow down, killing your results before they start.

Illija Stojkovski, CRO at Heyreach, has previously broken down why sales automation can go wrong fast —  from low-quality leads to broken outreach cadences, and those lessons apply here too. When it comes to AI prompts specifically, here are the three issues that show up in real GTM operations:

1. They don’t fit your workflow

A prompt is only useful if it maps to how your team actually runs outreach. Generic prompt lists rarely match your systems or processes, so you end up stitching things together manually, which defeats the whole point of using AI.

2. They aren’t role-specific

SDRs, AEs, and RevOps might all sit under sales, but they have different goals, motions, and success metrics. Generic AI sales prompts lump all these roles into one and cause three predictable problems:

• Nobody uses them: Sales reps try it once, realize it doesn’t fit their workflow, and go back to manual writing.
• Execution slows down: Instead of saving time, you spend more time editing AI outputs just to make them usable.
• Critical touchpoints get missed: Stage-specific prompts like objections or deal follow-ups fall through the cracks, hurting pipeline health.

3.  They ignore compliance guardrails

Most outreach tools have built-in rules that guide consent, opt-outs, and pacing. Generic prompts often ignore them, which can affect your deliverability rates or even put you at legal risk. Speed is great, but not at the cost of compliance. 

If your prompts have any of these issues, AI won’t scale your GTM motion; it’ll slow it down.

The anatomy of a good AI prompt

A good AI prompt is clear, structured, and context-rich so the output is immediately usable in your workflow. This is where prompt engineering comes in.

Prompt engineering is the process of designing clear, structured instructions for AI so it returns reliable, actionable outputs that fit your workflow. Done right, a prompt turns AI from “chatbot” into a junior SDR or research analyst that follows your instructions precisely.

But there’s a second layer many sales teams miss: context engineering. This is where you feed the model the right background data: prospect profiles, product value proposition, call notes, even a CSV schema, so every output is hyper-relevant and immediately usable. 

This layered approach matters because combining “prompt + context engineering” is how you bridge the gap between theory and an operational sales system that works.

The four elements of a high-performing prompt

Every high-performing sales prompt follows the same blueprint: Role & persona, Goal/task, output formats, context & guardrails.

Here’s how each one works:

1. Role & Persona

This sets the stage. Without it, your outputs can be too broad or generic. Defining the role tells the model who it’s “speaking” to, what ICP it’s targeting, and what success looks like.

Example: “You are a B2B SaaS SDR focused on mid-market HR tech buyers. Your goal is to book demos with HR directors in North America.”

2. Goal/Task

AI works best when it knows exactly what you want. This is where you define what action it needs to take,  whether that’s to book a meeting, generate a CSV, classify leads, write sequences, etc.

Example: “Generate 10 LinkedIn DM openers optimized for a high reply rate.”

3. Output formats

If you don’t specify the format, you end up cleaning messy AI outputs by hand, defeating the point of automation. What you should do instead is tell AI exactly how to structure the results you want, so you can drop it straight into your workflow.

Example: “Return results as CSV with columns: Message, Persona, Funnel Stage,...”

4. Context & guardrails

AI will “hallucinate” without guardrails. Context keeps outputs relevant, and constraints keep them compliant and usable at scale.
Example: “…based on the prospect’s recent LinkedIn post about launching a product. Follow these approved messaging templates {insert message template here}. Generate 5 concise LinkedIn outreach DMs that are clear, non-spammy, and easy to personalize. Keep under 200 characters each.”

The input requirements that make prompts work: ICP, profile & CRM data context

AI prompts only perform as well as the data you feed them. If you skip data context, your outputs will be vague and miss the mark.

What to include before you hit “generate”:

• ICP data – industry, company size, job titles, pain points.
• Prospect profile – LinkedIn headline, recent posts, work history.
• Company signals – funding round, hiring trends, tech stack.
• CRM context – last touchpoint, stage in pipeline, past objections.

If you’re in the process of building your go-to-market strategy, set this foundation right from the start. It will make everything else that follows much easier. 

Pro tip: Templatize these inputs. Map fields directly from your CRM or outbound tool into prompt variables so every output stays consistent with your source of truth, and you don't need to copy and paste manually. 

Another important thing you need to know is that context alone isn’t enough if the system your prompts sit in is broken. This is why your sales orchestration process matters: to align your data, tools, and people so every output is routed, enriched, and acted on automatically. 

Output formats: How to structure responses for import

Unstructured outputs slow campaigns down. Tell AI exactly how to format its results so you can drop them straight into your sequences without wasting time on cleanup.

Here is what you need to specify with every prompt:

• CSV Schema: Explicitly name every column (subject_line, body, CTA, tag).
• Tags: Include tags like persona, industry, campaign, so outputs can be filtered and segmented later
• Word limits: Keep messages short enough to fit channel character limits and avoid token overflow.
• Sequence readiness: Ask AI to number outputs (step_1, step_2, step_3) so they can be plugged directly into a sequence without reformatting.

The template formula for sales prompts you can use now

The best prompts for SDRs, AEs, and RevOps teams follow a simple repeatable structure:

“Act as [role]. Given [input], output [structured format] with [constraints].”

Example SDR prompt (LinkedIn outbound )

“Act as an SDR running a LinkedIn outbound campaign for a B2B HR tech SaaS company. Given this ICP profile: HR Directors or Talent Managers in North America who are actively hiring (from LinkedIn job posts), generate a 3-step LinkedIn outreach sequence and return results in CSV format with columns {step, message_type, message_body, CTA, persona_tag}. Keep each message under 80 words, use a friendly but professional tone, and avoid spammy language.”

Sample output I got from running this prompt

Step 1 – Connection Request
Persona: HR Director
Message: Hi {{FirstName}}, noticed {{CompanyName}} is actively hiring — congrats! We work with HR leaders to streamline hiring ops and improve time-to-hire. Thought it’d be great to connect.

Step 2 – Follow-up DM
Persona: HR Director
Message: Thanks for connecting, {{FirstName}}! Curious — how is your team currently managing candidate tracking as you scale? I’ve seen HR teams cut admin time by 30% with the right tools.
Call to action: Open to sharing a quick insight?

Step 3 – Value DM
Persona: HR Director
Message: Appreciate your reply (or time)! I put together a 1-page guide on speeding up hiring workflows for teams growing fast. Happy to send it over.
Call to action: Want me to share it here?”

Once you have a working prompt template formula that produces clean, CSV-ready outputs like this, you can plug it into AI agents to handle repetitive tasks like prospecting, research, follow-ups, or even queuing messages for review in your outreach tool.

Role-specific prompt banks & workflows for the sales team

AI prompts only work if they fit how your team actually sells. Here’s what each role can automate right away with the right prompts and HeyReach:

• SDRs: Prospect research, profile enrichment, trigger-based outreach, and safe sequencing
• AEs: Post-demo follow-ups, objection handling, deal acceleration, and multi-threading outreach
• RevOps: Lead list cleaning, deduplication, segmentation, and campaign analysis
• CSMs: Renewal reminders, NPS follow-ups, and expansion campaigns

Each workflow below comes with ready-to-use prompt examples, output formats, and guardrails you can drop straight into your process.

SDR prompt pack for prospecting and personalization

1. LinkedIn profile personalization (for cold outreach)

Scenario: You found a potential client’s profile but aren’t sure how to break the ice without sounding generic.

Prompt: Act as a B2B SDR targeting [ICP description, e.g., mid-market HR tech buyers]. Given the LinkedIn profile of [prospect name] at [company name], extract 3 personalized insights I can use in an outreach message. Prioritize:

• Shared experiences (e.g., schools, groups, geography)
• Recent activity (posts, comments, job changes)
• Notable career milestones or achievements
  Keep this under max 200 characters.

Fallback clause: If the profile has limited info, generate 1 insight based on the company industry trends or their job title.

2. Company event trigger

Scenario: Prospect’s company just raised a funding round, launched a new product, or announced a partnership, a perfect time to reach out.
Prompt: Act as a B2B SDR. Given [company name] and [recent event details: funding, hiring, new office, acquisition], write 2 outreach message hooks tying the event to why they should evaluate [product/service].
Format as:

• Event Summary
• Message Hook 1 (max 150 characters)
• Message Hook 2 (max 150 characters)

Fallback clause 1: If no relevant event exists, output a message referencing an industry trend (e.g., “many [industry] companies are investing in sales tools post recent 2024 regulations”)

Fallback clause 2: If no real-world or contextual industry trend or report is found, skip this step and note for manual review.

3. Mutual connection reference 

Scenario: You share a mutual LinkedIn connection or group with the prospect and want to leverage that to get a reply.
Prompt: Act as an SDR. Given [prospect name], [mutual connection name/group], and [product/service], write a friendly LinkedIn message referencing the mutual connection to build trust.

Include:

• First line mentioning the mutual connection naturally
• Value prop (why I’m reaching out)
• CTA for a 15-minute conversation

Fallback clause: If no mutual connection exists, use a softer CTA that focuses on learning about their role instead of pitching.

4. Industry pain point positioning 

Scenario: Prospect hasn’t engaged, but you want to follow up with something relevant that isn’t just “bumping this up.”
Prompt: Act as an SDR who deeply understands [prospect’s industry]. Given [prospect role], [company name], and [common industry pain points], generate a 2–3 sentence message:

• Acknowledge a known pain point for that role
• Briefly show how [product/service] addresses it
• Include a low-pressure CTA to learn more

Fallback clause: If no role-specific pain point is found, output a trend or insight relevant to the industry that sparks curiosity.

5. Meeting request formatting 

Scenario: Prospect replied positively, now you need to lock in a meeting without creating friction.
Prompt: Act as an SDR. Given [prospect name], [company name], and [product/service], draft a 2-sentence LinkedIn or email follow-up that:

• Thank them for engaging
• Suggests 2–3 specific time slots
• Keeps it under 150 characters per sentence

Fallback clause: If unsure of time slots, propose a CTA like: “Would next Tuesday or Wednesday work for a quick 15-minute chat?”

How to use these prompts in your workflow to go from prospecting → campaign

1. Run your chosen prompt in Claude, ChatGPT or your preferred LLM to generate structured outputs for your campaign. 
2. Review for tone, accuracy and context. Edit anything that feels off and approve outputs
3. Export your approved outputs as a CSV file
4. Import into HeyReach and create a campaign
5. Enable sender rotation to distribute outreach across multiple SDRs
6. Schedule sequencing: Example — Day 1 = connection request, day 3 = first message, day 5 = follow-up… etc

Safety guardrails for SDR outreach

• Fallback defaults: Always include a safe tag if the first name or other data is missing. This makes sure your campaigns never send broken personalization tokens and remain professional.
• Pacing limits: Cap daily sends per seat to stay within safe thresholds. HeyReach helps you easily configure this.
• Human review: Spot-check at least 5–10 emails per batch before launch to ensure personalization lands right.

AEs prompt pack for nurturing and objections

These prompts are centered around use cases to help AEs reach heads of teams or decision-makers to close sales conversations faster.

1. Post-demo follow-up 

Prompt: Act as an AE following up with {{persona}} after a product demo. Write a concise subject line and 120-word body that:

• Recaps key demo takeaways relevant to their role
• Highlights how the solution addresses their specific pain points
• Includes a soft CTA to schedule the next step
  Output as CSV with columns: Subject, Body, CTA, Tags.

Fallback clause: If no demo notes are available, reference a high-level product benefit relevant to their role.

2. Common objection handling

Prompt: Respond to this objection: “{{objection}}”. Keep the tone consultative and empathetic. Include:

• Reassuring stat, case study, or social proof
• ≤120 words
• A gentle CTA that moves the deal forward
  Output as plain text.

Fallback clause: If no objection context is given, use a general reassurance about product reliability and customer success.

3. Deal acceleration / Cold deal nurture 

Prompt: Write a 2-email nurture flow for a deal in {{industry}} that has stalled. Each email should:

• Include social proof, competitor switch wins, or success metrics
• Provide clear next steps or CTA to re-engage the prospect
• Keep language consultative and concise
  Output as CSV rows (2).

Fallback clause: If prospect-specific info is missing, reference industry-standard ROI or time-to-value metrics.

4. Multi-threading outreach

Prompt: Generate a 3-message LinkedIn drip sequence targeting other stakeholders in {{company}}. Each message should:

• Be ≤40 words
• Mention a shared goal, common pain point, or relevant initiative
• Include a soft CTA to connect or schedule a conversation
  Output as CSV rows (3).

Fallback clause: If stakeholder info is incomplete, reference company-level benefits and suggest a general discovery chat.

5. Pricing / ROI conversation 

Prompt: Act as an AE discussing pricing with {{persona}} after interest is confirmed. Generate a short email:

• Show a clear ROI or cost-benefit relevant to their team
• Address common pricing concerns
• Include CTA to schedule a pricing review call
  Output as CSV with columns: Subject, Body, CTA, Tags.

Fallback clause: If no prospect budget info is available, provide a standard pricing range and suggest a discovery call to align on specifics.

6. Stakeholder engagement 

Prompt: Generate a 2–3 message email or LinkedIn sequence targeting multiple decision-makers for {{company}}. Include:

• Key benefits for each persona
• Social proof or relevant metrics
• CTA to a joint call or demo
  Output as CSV rows (2–3).

Fallback clause: If persona info is missing, reference general company-wide benefits and suggest an introductory call.

How to turn these AE prompts into a repeatable workflow 

Use this simple blueprint to plug these prompts directly into your sales process and scale safely.

1. Trigger: A Unibox reply or CRM status update (e.g., “Demo completed,” “Objection raised”) triggers a Cargo or Make automation.
2. AI Processing: AI scans the reply, categorizes intent (interested, stall, objection, pricing), and drafts follow-up messages using the relevant AE prompt pack.
3. Routing: The draft is routed to the assigned AE or team channel in Slack for instant visibility while simultaneously logging the message and summary in your CRM under the correct contact/account record.
4. Approval Layer: AE reviews the AI-drafted follow-up, tweaks for tone or context, and approves it.
5. Sequence Push: Approved drafts are pushed to HeyReach (for LinkedIn outreach) and automatically inserted into the correct multi-channel follow-up sequence. Seat rotation and pacing limits are applied for safe sending.
6. Tracking: All activity is visible in HeyReach Master View for RevOps oversight.

Safety guardrails for AE outreach

AI can draft your follow-ups, but AEs must stay in control of what gets sent, especially in late-stage deals.

1. Review before sending: Every AI-generated message lands in Slack or CRM first for human review and edits.
2. No auto-send beyond certain stages: Set a rule that if a prospect mentions budget, contract details, or legal blockers, AI-generated messages should never auto-send. This is because these stages are too high-stakes for automation. 
3. Use pacing + seat rotation: Apply HeyReach safety settings to throttle sends and distribute outreach evenly across seats for deliverability protection.

Prompt pack for customer success managers/ account managers

These prompts let you generate messaging that’s personal, timely, and scalable, while keeping your customers at the center.

1. Renewal reminder personalization 

Prompt: You are the CSM managing {{Account}}. Renewal is coming up on {{Renewal_Date}}, and usage shows high engagement with {{Feature_X}} but low adoption of {{Feature_Y}}. Draft a friendly, consultative renewal reminder for the {insert role/user persona} that:

• Highlights the value they’ve received from {{Feature_X}}
• Suggests a 15-minute conversation to discuss optimizing the adoption of {{Feature_Y}}
  Keep tone personable but professional. 

Fallback clause: If usage data is missing, focus on a high-level product benefit relevant to their role and invite a short discussion.

2. QBR agenda generation 

Prompt: You are the CSM preparing the QBR for {{Account}}. Using account goals, usage data, and adoption trends, draft an agenda for the quarterly business review, including:

• 3 wins to celebrate
• 2 risks to address
• 2 discussion points for the next quarter.

Fallback clause: If usage data is incomplete, include standard wins, potential risks, and discussion points based on typical account behavior.

3. Expansion hooks 

Prompt: You are the AM responsible for driving expansion for {{Account}}. Based on product usage and feature gaps, generate 2 consultative expansion ideas for upsell or cross-sell. First outreach should educate and inform, not push a sale. Tailor messaging to the account persona (e.g., Head of Marketing, Product Manager). 

Fallback clause: If usage data is incomplete, suggest industry-standard upsell opportunities or additional modules most relevant to similar accounts.

4. Risk-flag outreach 

Prompt: You are the CSM for {{Account}}. Usage data shows low engagement or NPS <7. Draft 2 proactive, supportive outreach messages to uncover blockers, encourage usage, and offer guidance. Keep tone empathetic, not salesy. 

Fallback clause: If NPS or usage data is missing, flag the account for manual review and send a general check-in message focused on support.

5. Advocacy/testimonial request

Prompt: You are the CSM for {{Account}}. The customer has NPS >9. Draft a short, easy-to-respond-to message requesting a testimonial, case study, or review. Keep it friendly and low-friction. 

Fallback clause: If NPS data is unavailable, select an account with positive recent feedback or strong product engagement.

6. Low activity check-in

Prompt: You are the CSM managing {{Account}}. Usage data shows low logins in the past 30 days. Draft 2 gentle re-engagement messages that:

• Remind them of key features or benefits
• Offer tips or resources to increase adoption
• Include a soft CTA to schedule a check-in

Fallback clause: If usage data is missing, send a friendly message referencing general features and inviting a discussion about optimizing value.

7. NPS/Customer sentiment follow-up  

Prompt: You are the CSM for {{Account}}. For accounts with low MPS (<7), generate 2 outreach messages to gather feedback and suggest next steps to improve their experience. Keep tone empathetic and constructive. 

Fallback clause: If NPS data is missing, flag the account for manual review and send a general feedback request.

CSM/AM prompt to workflow blueprint

Use this workflow to turn raw account data into timely, trust-building outreach that drives renewals and expansions at scale.

1. Pull account data: Gather usage stats, renewal dates, NPS scores, adoption trends, and key stakeholders. More context = better AI outputs.
2. Generate messaging: Use Twain or Claude to draft renewal reminders, upsell pitches, risk alerts, or QBR agendas using that context.
3. Export as CSV: Structure outputs with columns (account, message, CTA, persona) for easy import.
4. Import to Heyreach: Load CSVs into HeyReach for campaign sequencing or direct Unibox sends.
5. Sequence & review: Set seat rotation and pacing controls to keep outreach safe and personal.
6. Launch & monitor: Run campaigns, track engagement in Heyreach Masterview, and feed insights back into your next AI runs.

Safety guardrails for CSMs & AMs

Your messaging must build trust while driving outcomes. AI is here to assist with these and not replace human judgment.

• Lead with value: AI prompts should educate and guide customers, not hard-sell every time.
• Customer-first tone: Keep outreach consultative and empathetic, especially in risk-flag or low-NPS scenarios.
• Fallback defaults: If data fields are missing, use high-level benefits or flag the account for manual review.

RevOps prompt pack for sales enablement 

RevOps may not sell directly, but you power everything that keeps sales running smoothly. These AI prompts help you deliver cleaner data, faster routing, and better attribution, giving SDRs and AEs the context they need to act quickly and confidently.

1. Lead qualification and scoring 

Prompt: You are a RevOps analyst enabling the sales team. Given a dataset of leads with fields like company size, industry, role, engagement, and recent activity, assign a lead score (0–100) based on likelihood to convert. Output a CSV: Lead_ID, Score, Key_Factors. Include reasoning for each score to help SDRs prioritize follow-up.

Fallback: If critical fields are missing, flag the lead for manual review instead of scoring.

2. Reply sentiment classification

Prompt: You are a RevOps analyst classifying incoming prospect replies. Given email or LinkedIn messages, label each as “Positive, Neutral, or Negative,” and highlight the key sentiment driver.

Fallback: If message content is missing or unreadable, mark as ”Needs review”.

3. Campaign performance analysis

Prompt: You are a RevOps analyst evaluating a multi-channel campaign. Given campaign data (emails, LinkedIn DMs, meetings booked, responses), summarize performance by channel, stage, and persona. Include top-performing sequences, bottlenecks, and recommendations for the next round. Output CSV: Channel, Stage, Persona, Metric, Insight.

Fallback: If data is incomplete, provide insights based only on available channels and flag missing metrics.

4. Attribution mapping / CRM tagging

Prompt: You are a RevOps analyst preparing data for CRM import. Given leads, interactions, and campaign data, generate opportunity tags and source attributions for CRM mapping.

Fallback: If attribution fields are missing, leave them blank and flag for review before import.

5. Data hygiene 

Prompt: You are a RevOps analyst auditing lead and account data for CRM import. Identify missing fields, duplicates, or invalid entries. 

Fallback: If data is severely incomplete, mark entire records for manual intervention.

Implement guardrails like missing-data flags, fallback actions, and review steps to make this layer reliable and safe.

RevOps prompt workflow: 

1. Score & prioritize: Feed leads into Persana AI to generate scores and ICP matches.
2. Trigger automation: Use Cargo or Make to route leads or trigger sequences when score thresholds are hit.
3. Enrich CRM: Automatically update CRM with scores, tags, and persona data.
4. Notify sales: Send Slack alerts to SDRs/AEs with next-best-action recommendations so they can act fast.

Safety guardrails for scaling AI-generated outreach 

Use this guardrails checklist for your next campaign. Copy and paste it into a Doc or SOP so your team has a working system to scale AI-powered outreach safely and get the desired outcome. Wherever relevant, I’ve highlighted how HeyReach helps you implement these safeguards seamlessly.

1. Seat rotation 

Spread your messages across multiple LinkedIn or email senders to avoid platform flags.

Example: 5 SDR seats sending 20 messages/day each is safer than 1 SDR sending 100.

HeyReach lets you balance multiple senders per campaign, so rotation happens without manual effort.

2. Pacing and daily send limits

Even with AI-generated messages, sudden spikes look suspicious. Set hourly and daily caps, and stagger send times. You can set max sends per seat/day in Heyreach and keep your campaigns running smoothly.

3. Always have a safety net or fallback defaults

Broken personalization tokens happen. Define safe defaults so even if AI fails, the system sends a safe backup copy instead.

For example, if the first name is missing, AI uses “Hi there” or a job-title-based opener, avoiding embarrassing sends like “Hi [FirstName],” in a live campaign.

4. Keep humans in the loop 

Before hitting live campaigns, review sample batches. This gives you an oversight of all live campaigns, so nothing slips through. 

5. Monitor everything 

Track bounce rates, reply quality, and engagement metrics. HeyReach MasterView gives you real-time oversight in one place and lets you quickly pause or optimize campaigns if negative signals spike.

6. Make sure you are compliant and executing safely

Include opt-out links, adhere to GDPR/CCPA rules, and respect local sending limits so your campaigns follow industry standards.

How to measure & improve results from AI-powered outreach

To make sure your prompts are producing meaningful results, you need a lightweight, consistent audit loop that shows you:

• Are your AI prompts generating usable, high-quality leads and messages?
• Are those leads moving through the funnel at the right speed?
• Are campaigns staying compliant and deliverable?

Use this playbook to measure your results:

Step 1: Track the right metrics 

Classification accuracy
What it tells you: Is AI tagging (e.g., reply intent, lead quality) trustworthy?
How to measure: Manually review a sample of ~50 replies, compare AI classification vs human judgment
Tool / Field: HeyReach tags, Unibox, Slack review thread
Cadence / Owner: Bi-weekly / SDR Manager

Attribution clarity
What it tells you: Can you prove the pipeline came from campaigns?
How to measure: percentage of opportunities with campaign/source field populated
Tool / Field: CRM campaign attribution field
Cadence / Owner: Monthly / Marketing Ops

Positive reply rate
What it tells you: Are prompts generating responses that move deals forward?
How to measure: Positive Replies ÷ Total Sends × 100
Tool / Field: HeyReach campaign analytics
Cadence / Owner: Weekly / SDR Manager

Deliverability health
What it tells you: Are campaigns avoiding spam filters?
How to measure: Check that the bounce rate is less than 2%, the open rate is greater than at least 40 %, spam rate is under 0.1%
Tool / Field: HeyReach deliverability dashboard
Cadence / Owner: Weekly / RevOps/ SDR lead

Prompt quality
What it tells you: Are AI prompts producing outputs that reps can use without heavy edits?
How to measure: Sample 20 AI outputs → count how many require more than 10% manual edits.
Tool / Field: Notion or Google Sheet tracking
Cadence / Owner: Bi-weekly / RevOps / SDR Lead

Step 2: Set a cadence

You don’t need a full audit every day. Weekly, bi-weekly, and monthly checks are enough. What matters is staying on top of the data so you can iterate fast

Step 3: Assign clear ownership

“Everyone” owning a metric means no one actually owns it. Be explicit: who reviews, who fixes, and who implements improvements.

Step 4: Build a simple scorecard

Track your metrics, targets, last week’s result, status, owner, and actions all in one place. Notion or Google Sheets works well for this. 

💡 Pro tip: Review this in your weekly GTM sync. It keeps AI adoption grounded in real results, not just enthusiasm for shiny new workflows.

Step 5: Improve quickly

Have a fallback strategy for what to do whenever a metric is off. Adjust fast, refine prompts, fix CRM mapping, tighten pacing, or test new angles.

Treat this like you would any pipeline: measure, diagnose, improve.

Implementation paths — Choose your AI integration level 

The fastest way to prove AI works for your sales team is to launch a small, controlled campaign and measure real replies not just generate lists. Once you see results, you can layer in more automation with confidence.

Standard Path: This works for most teams

You can run your first AI-assisted campaign in just 1–2 days:

Day 1 – Setup & test

• Choose prompts for lead research , using Claude or Persana 
• Generate lead list outputs → export to CSV → import into HeyReach.
• Use Twain to create outreach copy for your campaign.
• Configure senders, pacing, and seat rotation in Heyreach
• Launch a small test campaign (50–100 leads).

Day 2 – Validate & scale

• Review early replies in HeyReach Unibox.

Alt text: HeyReach Unibox helps you organize all LinkedIn messages in one place

• Check sentiment tags, lead routing, and guardrails (fallbacks, dedupe, opt-outs). Adjust prompts, messaging, or pacing based on what you see.
• If everything looks good, scale up to larger segments.

Turn Claude into your sales assistant with HeyReach MCP

For GTM operators, MCP (Model Context Protocol) lets you turn Claude into a natural-language interface for HeyReach. Instead of manually logging into HeyReach to upload lists, clean them, or launch campaigns, Claude can safely execute those actions for you inside Heyreach.

HeyReach MCP is scoped to HeyReach only, so Claude can safely execute actions like adding leads or launching campaigns.

For example, here is how you can carry out a list cleanup with MCP

Instead of downloading a CSV, filtering manually in Excel, and re-uploading, you can just run:

Prompt:

Act as a list cleaner for HeyReach. Retrieve all leads from Sept Leads List using Claude (that you previously connected to HeyReach MCP). Review both their job title and headline. Create a new list called ”Clean List – Founders Only” with only valid profiles matching founders, co-founders, or CEOs. Exclude anyone with ”student or intern” in their title or headline. Provide a short summary of which profiles were excluded and why, then finalize the new list.

Claude will:

• Fetch your list directly from HeyReach
• Filter based on your rules
• Generate a summary of what it removed
• Create a new cleaned list inside HeyReach automatically

Chaining with multiple MCPs

You can also use multiple MCPs at the same time. For example, if any of the other tools in your sales stack, say HubSpot or Salesforce, also have an MCP, you can chain actions:
Claude → fetch contacts from HubSpot → filter by criteria → push into HeyReach campaign automatically.

Want more prompt examples?
Check out HeyReach’s MCP prompt playbook for copy-paste-ready prompts to clean lists, re-segment prospects, and help you launch campaigns in minutes.

Put these prompts into action today

You now have:

• Role-specific prompt banks for SDRs, AEs, CSMs, and RevOps
• Workflow maps that show exactly where to plug AI outputs into campaigns
• Guardrails to keep outreach safe and compliant

The fastest way to see results? Run one of these prompts in your AI tool today and launch your first campaign in HeyReach.

Start your free trial and see how quickly you can go from AI output → booked meetings → revenue.