---
title: "Prospect research and outreach prompts"
description: "Research, insight, message Good outreach follows a simple chain: research (what's true about this company now) leads to an insight (why they might care…"
url: https://optimizeall.com/learn/prompt-engineering-for-content-and-sales/prospect-research-and-outreach-prompts
updated: 2026-10-05
---

Prompt Engineering for Content & Sales · Prompts for sales: outreach, calls, follow-ups and proposals · lesson 8 of 15 · 13 min

# Prospect research and outreach prompts

## Research, insight, message

Good outreach follows a simple chain: **research** (what's true about this company now) leads to an **insight** (why they might care, now) leads to a **message** (short, specific, one easy ask). AI speeds up all three, but the chain breaks if the research is wrong, the insight is generic, or the message is creepy.

## Step 1: account research (public, professional information only)

Use web search, deep research or Perplexity for a one-page account brief:

```text
Company: [name], [country/city], [website].
Using public sources from the last 12 months, give me:
1. What they do, who they sell to, and how they describe themselves (quote their site)
2. Recent news: launches, expansion, hiring, funding, leadership changes (with links and dates)
3. Likely priorities or pains relevant to [what we sell], labeled as hypotheses
4. Two or three specific, non-obvious conversation openers based on the above
Mark anything single-sourced. Company and professional information only.
```

Open the two or three links you'll mention. A wrong "congratulations on your expansion" ends the conversation.

## Step 2: from research to insight

```text
Given this account brief and our context pack (offer sheet, approved proof, customer language),
what is the ONE most relevant reason this company might care about [our offer] now?
Give the reason, the evidence from the brief, the proof point we can use, and a one-line hook.
If there's no strong reason, say so; we'll deprioritize this account.
```

"No strong reason, deprioritize" is a feature: it stops you sending weak outreach.

## Step 3: the message

```text
Write a [LinkedIn message / email] to [role] at [company].
Opening: the insight, specific to them (not about us).
Middle: one sentence on what we do + one approved proof point.
Ask: one low-friction question (not a meeting request in the first message).
Rules: under [80] words; no flattery clichés; no fake familiarity; no invented mutual connections,
results or urgency; plain text; include our opt-out line for email.
Give 2 versions with different openings.
```

## Personalization tiers

| Tier | When | Effort |
|---|---|---|
| **1:1** | Top accounts | Full research, insight and hand-edited message |
| **Segment** | Similar companies (e.g. clinics opening second branches in Riyadh) | One insight per segment, light personalization |
| **Trigger-based** | A public event (new location, new role, funding) | Short message tied to the trigger |

Don't fake 1:1 personalization at scale; recipients notice.

## Compliance essentials

Outreach rules differ by country. Examples: in the **US**, CAN-SPAM requires honest headers and subject lines, a physical address and a working opt-out for commercial email; in the **UK and EU**, GDPR and (in the UK) PECR govern using personal data and sending electronic marketing, with stricter rules for individuals and sole traders; the **UAE and Saudi Arabia** have their own data-protection laws. Research only professional information, honor opt-outs, and follow your company's policy. For deliverability and compliance in depth, see [AI for Sales Teams](/learn/ai-for-sales-teams).

## Worked example (illustrative)

Ayesha sells scheduling software to clinics from Karachi into the Gulf. For a Riyadh dental group, account research finds a public announcement of a second branch opening next quarter (confirmed on the group's own site). The insight: scheduling across two branches is the pain her customers describe most ("patients booked at the wrong branch"). Her 70-word LinkedIn message opens with the second branch, cites one approved result from a similar clinic group, and asks one question: "How are you planning to handle bookings across both branches?" She gets a reply asking for a short demo.

## Hands-on: three accounts in 45 minutes

1. Pick three real target accounts.
2. Run the account research prompt; open and confirm the facts you'll mention.
3. Run the insight prompt; drop any account with "no strong reason".
4. Draft two message versions each; edit the first line by hand.

**Before:** "Hi [Name], I hope you're well! I came across your company and was impressed. We help businesses like yours grow. Do you have 30 minutes this week?"

**After:** "Saw [Company] is opening a second branch in [district] next quarter. Clinic groups we work with say double-bookings across branches are the first headache. One group cut them sharply after moving to a shared schedule [approved proof]. How are you planning to handle bookings across both sites?"

## Pitfalls

- Mentioning "news" you didn't verify.
- Researching personal lives (family, holidays, personal social accounts).
- Invented mutual connections, results or deadlines.
- Automating sends of AI-written messages without review.

## How to measure success

Track reply rate and positive-reply rate per tier and per insight type, not just volume sent. Fewer, better messages usually win.

## Video lecture: Prospect research and outreach prompts

Lecture coming soon · 11 chapters · about 8 minutes. Read the full transcript below.

1. Prospect research and outreach
2. Why fewer, better
3. The doctor's referral analogy
4. Step 1: account research
5. Step 2: the insight
6. Step 3: the message
7. Example 1: segment outreach
8. Example 2: Ayesha's Gulf outreach
9. Watch me do it
10. Compliance essentials
11. Common mistakes and recap

## Lecture transcript

### Prospect research and outreach

How many cold messages have you received this week that started with, I hope you're well, I came across your company and was impressed? And how many did you reply to? Probably none. AI has made it easy to send more outreach, which is exactly why generic outreach is now worthless. In this lecture you'll learn a three step chain, research, insight, message, that uses AI to write fewer, better messages. You'll see how to research accounts responsibly, how to find the one reason a prospect should care now, and how to write a message that earns a reply. You'll follow a Karachi saleswoman's Gulf outreach and watch me work through three accounts.

### Why fewer, better

Why aim for fewer, better messages? Because volume without relevance is just spam with better grammar. Buyers can spot fake personalization instantly: a first name, a company name, and a compliment that could apply to anyone. What earns replies is relevance: a message that shows you understand something true about their situation right now, and connects it to a problem you actually solve. AI is excellent at the research and drafting that relevance requires. But only if you use it to think, not just to multiply.

### The doctor's referral analogy

Think of a good specialist referral. The specialist reads the patient's file first. The referral states one clear reason for the visit, not ten. And it proposes one clear next step. Outreach is the same. Research is reading the file: what's true about this company now. The insight is the reason for the visit: the one reason they might care about your offer, now. And the message is the referral note: short, specific, with one easy next step. If you can't find a reason for the visit, you shouldn't be sending the referral.

### Step 1: account research

Step one is account research, using web search, deep research or Perplexity. Ask for what the company does and how it describes itself, quoting its own site. Recent news from the last twelve months: launches, expansion, hiring, funding, leadership changes, each with a link and date. Likely priorities relevant to what you sell, labeled as hypotheses. And two or three specific conversation openers. Two rules. Company and professional information only: no family, holidays or personal social accounts. And open the links for anything you'll mention. Nothing ends a conversation faster than congratulating someone on news that isn't true.

### Step 2: the insight

Step two turns research into an insight. Give the assistant the account brief and your context pack, with your offer sheet, approved proof and customer language, and ask: what is the one most relevant reason this company might care about our offer now? Give the reason, the evidence from the brief, a proof point we can use, and a one line hook. And add: if there's no strong reason, say so, and we'll deprioritize this account. That last line is a feature, not a bug. It stops you sending weak outreach just because you did the research. Sometimes the best message is the one you don't send.

### Step 3: the message

Step three, the message. Open with the insight, specific to them, not about you. One sentence on what you do, and one approved proof point. Then one low friction question, not a meeting request in the first message. Rules in the prompt: under eighty words, no flattery cliches, no fake familiarity, and no invented mutual connections, results or urgency. For email, include your opt out line. Ask for two versions with different openings, and edit the first line by hand. That's where your own judgment makes the biggest difference.

### Example 1: segment outreach

A simple example of segment outreach. Instead of fully researching fifty accounts, a sales rep finds five clinic groups in Riyadh that have publicly announced second branches. One insight covers the whole segment: running bookings across two locations is the first headache customers describe. The message template stays the same, and only the first line changes, referencing each group's specific new location. That's honest personalization. It's relevant to each company, and it doesn't pretend to be a handcrafted letter. Save full one to one research for your top accounts.

### Example 2: Ayesha's Gulf outreach

Now a fuller example, with illustrative details. Ayesha sells scheduling software to clinics, from Karachi into the Gulf. For a Riyadh dental group, account research finds a second branch opening next quarter. She confirms it on the group's own website, not just the news site. The insight prompt connects it to the pain her customers describe most: patients booked at the wrong branch. Her seventy word message opens with the second branch, cites one approved result from a similar clinic group, with that client's permission, and asks a single question: how are you planning to handle bookings across both branches? The practice manager replies, asking for a short demo. One message, fully relevant, and it worked.

### Watch me do it

Let me work through three accounts. The first is a logistics firm in Manchester. The research brief shows they're hiring five new dispatchers and launched same day delivery, both on their own careers page and newsroom. The insight prompt connects that to our route planning software: scaling same day delivery with new dispatchers is exactly when routing breaks. It drafts two messages. I rewrite the first line myself to reference the same day launch naturally. Second account: the research finds nothing recent and relevant. The insight prompt says no strong reason, deprioritize. Good. That saves me a weak message. Third account: an insight about a new warehouse, but it's single sourced on a blog. I check the company site. Not mentioned. I leave it out and use a segment message instead.

### Compliance essentials

A quick word on compliance, because AI makes it easy to send more, faster. In the US, CAN SPAM requires honest headers and subject lines, a physical address and a working opt out for commercial email. In the UK and EU, GDPR governs how you use personal data, and in the UK, PECR governs electronic marketing, with stricter rules for individuals and sole traders than for corporate addresses. The UAE and Saudi Arabia have their own data protection laws. So research only professional information, honor opt outs, follow your company's policy, and never let an AI tool send messages without human review. The AI for sales teams course covers deliverability and compliance in depth.

### Common mistakes and recap

Let's finish with the mistakes and a recap. The mistakes: mentioning news you didn't verify, researching people's personal lives, inventing mutual connections, results or deadlines, and automating sends without review. The recap: follow the chain. Research what's true now, from public, professional sources. Find the one reason they'd care now, or deprioritize. And write a short message with one proof point and one easy question. Measure reply rate and positive reply rate, not volume. Here's your try this now. Take three real target accounts and run all three steps from the lesson text. Send only the messages that passed the insight test.

## Key takeaways

- Follow research, insight, message: what's true now, why they'd care now, and one short, specific ask.
- Use web search or deep research for public, professional information only, and verify anything you'll mention.
- Let the AI say 'no strong reason' and deprioritize weak accounts.
- Respect outreach law (CAN-SPAM, GDPR, PECR, local data laws): honest subject lines, opt-outs and no fake personalization.

## Try it

Run research, insight and message for three real accounts; send only those that pass the insight test, and track replies.

- [Previous: Long-form content with prompt chains: articles, newsletters, case studies and landing pages](https://optimizeall.com/learn/prompt-engineering-for-content-and-sales/long-form-content-prompt-chains)
- [Next: Call prep, rehearsal and post-call follow-up](https://optimizeall.com/learn/prompt-engineering-for-content-and-sales/call-prep-and-discovery-prompts)
- [All lessons of Prompt Engineering for Content & Sales](https://optimizeall.com/learn/prompt-engineering-for-content-and-sales)