cotera
Articles

Salesforce Automated Emails: Beyond Templates and Merge Fields

Ibby SyedIbby Syed, Founder, Cotera
9 min readMarch 6, 2026

Salesforce Automated Emails: Beyond Templates and Merge Fields

Salesforce Automated Emails Guide

Here's a Salesforce email template. You've seen it before. You've probably sent it.

"Hi {First_Name}, I noticed {Company_Name} is in the {Industry} space. Many companies like yours struggle with {Pain_Point}. I'd love to show you how we help teams like yours {Value_Prop}. Would you have 15 minutes this week?"

Every SDR in the world sends some version of this. Every prospect in the world deletes it. The merge fields create an illusion of personalization while producing something that is, unmistakably, a template. Everyone knows it. The sender knows it. The recipient definitely knows it.

Salesforce has had email templates and merge fields for over a decade. They were clever in 2014. In 2026, they're the reason your outbound reply rates are stuck at 5-8%.

The merge field problem

Merge fields pull data from Salesforce records and insert it into email templates. First name, company name, title, industry. Some teams get fancier with custom fields: last product viewed, subscription tier, days since last login.

The problem isn't technical. The problem is that the data available in Salesforce merge fields is the exact same data available to every other company emailing this person. Everyone knows their first name. Everyone knows their company. Everyone knows their industry. Personalizing with universally available information is not personalization. It's mail merge.

Marcus ran our outbound team for two years. He tracked reply rates obsessively. Templates with standard merge fields: 6.2% reply rate across 14,000 emails. He tried everything to improve that number within the template framework. Different subject lines (tested 23). Different CTAs (tested 11). Different send times. Different sequence lengths. The best-performing template variation: 8.1%. The worst: 4.4%. The ceiling on merge-field emails is low and fixed.

Then he did something different. He took 200 accounts and manually researched each one before writing the email. No template. Just a first line that referenced something specific about the company or person. Recent news. A LinkedIn post. A job listing. A product launch. The rest of the email was similar to the template, but that researched first line changed everything.

Reply rate on the 200 researched emails: 22.4%.

Same SDR. Same value prop. Same CTA. The only variable was whether the opening line referenced something the prospect would recognize as genuine research vs. a merge field anyone could pull.

What "researched" emails actually look like

Here's a merge field email:

"Hi Sarah, I noticed Acme Corp is in the fintech space. Many fintech companies struggle with customer onboarding drop-off. I'd love to show you how we help teams reduce onboarding friction. Would you have 15 minutes?"

Here's a researched email:

"Hi Sarah, I saw Acme Corp just launched the instant verification feature last month. That's a big move given how much friction traditional KYC creates for consumer fintech users. We've been helping three other consumer fintech companies optimize what happens right after verification, where most drop-off actually occurs. Would it be worth a conversation?"

The second email took 5 minutes of research. It references a real product launch. It connects the launch to a problem the prospect probably cares about right now. It positions the sender as someone who actually understands the prospect's world.

No Salesforce merge field contains "just launched instant verification feature." That information lives on the company's blog, in a press release, or on Product Hunt. Salesforce templates can't reach it.

The scale problem

The reason merge field templates exist is scale. An SDR needs to send 50-80 emails per day. If each one requires 5 minutes of research, that's 4-6 hours of research and 2 hours of actual sending. The math doesn't work. Research quality and email volume are inversely correlated when humans do both.

So teams make the rational choice: sacrifice quality for volume. Send 80 template emails instead of 15 researched ones. The 80 templates generate roughly 5-6 replies. The 15 researched emails would generate 3-4 replies. Volume wins by a little.

But that calculation ignores downstream conversion. Priya tracked this for a quarter. Replies from template emails converted to meetings at 38%. Replies from researched emails converted at 61%. The prospect who replies to a researched email is already in a different headspace. They're responding to someone who seems credible, not just swatting away another cold emailer.

When you follow the math all the way through: 80 templates generate 5 replies and 2 meetings. 15 researched emails generate 3 replies and 2 meetings. Same output, wildly different input. And the researched meetings tend to be better. Prospects show up warmer. First calls go deeper. Deal velocity is faster.

So neither approach alone works at the scale most teams need. Templates get volume but bad results. Research gets good results but no volume.

The AI research layer

This is where the old template vs. quality tradeoff stops being a tradeoff.

A Salesforce bulk lead importer agent does the research step before any email gets written. When leads enter Salesforce, either through import, form submission, or manual creation, the agent researches each account automatically. It pulls recent company news, checks LinkedIn for the contact's recent activity, scans job postings for priority signals, identifies the tech stack from public data, and writes a research brief directly to the Salesforce record.

The brief sits on the lead or contact record. When the SDR sits down to write the email, they're not starting from a blank template. They're reading three paragraphs about the company's current situation and picking the most relevant angle for their opening line.

Anya, an SDR on our team, described it: "Before, writing a personalized email meant opening 6 tabs and spending 10 minutes reading. Now I read the brief for 30 seconds and write the first line. The whole email takes 90 seconds."

Her daily output: 45-50 emails per day, all with researched personalization. Reply rate: 18.7%. That's 9-10 replies per day compared to the 3-4 she got from template blasting.

Salesforce email automation: what it should look like

Here's how email automation works in most Salesforce instances today:

  1. Lead enters Salesforce
  2. Lead gets assigned to SDR
  3. SDR opens a template
  4. Merge fields auto-populate name, company, title
  5. SDR maybe customizes one line (usually doesn't)
  6. Email sends
  7. Follow-up emails fire on a schedule with more merge fields
  8. After 4-5 touches with no reply, lead goes to a nurture drip

Here's how it should work:

  1. Lead enters Salesforce
  2. AI agent automatically researches the account (60-90 seconds)
  3. Research brief writes to the Salesforce record
  4. Lead gets assigned to SDR
  5. SDR reads the brief and writes a personalized opening line (60-90 seconds)
  6. The rest of the email follows a loose structure, not a rigid template
  7. Follow-up emails reference different research angles, not the same merge fields repeated
  8. After 4-5 touches with no reply, the agent refreshes the research for a new angle

The difference is step 2. Everything downstream changes because the SDR has actual information to work with instead of {Company_Name} and {Industry}.

Reply rate benchmarks: template vs. researched

We tracked this across 8 months, 6 SDRs, and roughly 28,000 outbound emails.

Template emails with standard merge fields: 6.4% average reply rate. Range across reps: 4.1% to 8.9%.

Template emails with custom merge fields (tech stack, last activity, deal stage): 8.2% average. Custom merge fields helped but the ceiling was still low.

AI-researched emails with personalized first lines: 17.8% average reply rate. Range across reps: 14.2% to 23.1%.

The gap between the best template emailer (8.9%) and the worst researched emailer (14.2%) was 5.3 percentage points. Even the worst-performing rep using researched personalization outperformed the best rep using templates.

Elena, who had the lowest template reply rate at 4.1%, jumped to 16.8% with researched personalization. She wasn't a bad SDR. She was a bad template writer. When she had real information to work with, she wrote emails that sounded like a person, not a mail merge.

What about Salesforce's built-in AI email features?

Salesforce has been adding AI to email. Einstein Email Insights suggests optimal send times. Sales Cloud Einstein can generate email drafts. The newer Agentforce features promise AI-composed emails based on CRM data.

I've tested all of them. The generated emails are better than a blank template but worse than a human with research. They suffer from the same constraint as merge fields: they only know what's in Salesforce. If your Salesforce data says "Company: Acme Corp, Industry: Fintech, Size: 200 employees," the AI will write an email about fintech companies with 200 employees. It can't reference last week's product launch because that information isn't in the CRM.

The AI composition is a better template engine, not a research engine. Better at constructing sentences from available data. Still limited to available data.

Tomás tested Einstein-generated emails against AI-researched-then-human-written emails. Einstein drafts: 9.3% reply rate. Researched emails: 18.1%. The Einstein drafts were better than manual templates but nowhere near the level of emails informed by external research.

The uncomfortable truth about email automation

Most Salesforce email automation exists to help reps avoid doing the hard thing. The hard thing is knowing the prospect's situation before reaching out. Templates, merge fields, and automated sequences are all ways to send emails without doing the hard thing.

But the hard thing is the thing that works. Prospects reply when they feel seen. They feel seen when the email contains information that could only come from someone who bothered to look. No amount of automation on the sending side compensates for skipping the research side.

The fix isn't better templates. It's automated research. Do the hard thing at machine speed, then let humans do what humans are good at: writing an email that sounds like one person talking to another, informed by something real.

Your reply rates will double. Probably more. That's not a guess. We measured it.

Try These Agents

For people who think busywork is boring

Build your first agent in minutes with no complex engineering, just typing out instructions.

Get started