You hired a sales development rep to book meetings. That's the job. But if you track where their hours actually go, you'll find something uncomfortable: most of the week isn't selling. It's researching.
Tab open: LinkedIn. Tab open: Crunchbase. Tab open: the company's homepage. Google the VP of Sales, check if they've posted anything recently, try to figure out what their actual tech stack is. Write a half-personalized email. Send it. Start over.
Multiply that by 15 prospects a day. Then wonder why pipeline is thin.
The Manual Research Loop
Here's what a typical SDR research workflow looks like, broken down honestly:
- Finding the right person at a company: 8–12 minutes. The VP of Sales from LinkedIn might have left six months ago. The actual decision-maker might be a Director of Revenue Ops nobody writes about.
- Gathering context (recent news, funding, product launches, headcount changes): 10–15 minutes per company if you're doing it right. Less if you're copy-pasting generics.
- Writing a personalized email that actually references what you found: another 8–12 minutes. The difference between a 1% and 8% reply rate is specificity. Specificity takes time.
That's 26–39 minutes per prospect. For 10 prospects, you're looking at 4+ hours. For a rep working an 8-hour day with meetings, admin, and CRM updates, prospecting research consumes the majority of available time.
That number — 65% — has been stubbornly consistent across surveys for years. The tools have gotten shinier. The process hasn't changed much.
The Math Nobody Wants to Do
Let's run the numbers on what this actually costs.
An SDR at quota touches roughly 50 new prospects per week. At, say, 30 minutes of research per prospect (being generous), that's 25 hours. In a 40-hour week, you've already burned more than half the week before they've sent a single cold email.
Except it gets worse. Most of that research time is producing good-enough outputs, not great ones. Reps under time pressure skip the deep context and default to semi-personalized templates. "I noticed you're scaling your sales team" gets applied to 20 companies. The reply rate tanks. The manager pulls the rep into a training on "better personalization." The rep stresses about quota. The cycle continues.
"The irony of modern sales is that the tools designed to scale outreach have mostly just scaled mediocrity."
The 80% of research that's routine — company description, headcount, recent news, finding a plausible contact — is exactly the kind of work that doesn't require a human. The 20% that actually requires judgment (why does this specific company need what you're selling, right now?) gets squeezed out because there's no time left.
What If AI Handled the 80% That's Routine?
This is the bet behind AI-assisted prospecting tools like Crescendo. Not to replace SDR judgment — but to eliminate the part that's just pattern matching and data assembly.
Describe your ideal customer: industry, company size, target role. The AI finds real companies that match, pulls context on each one, constructs a plausible contact, and drafts a personalized cold email referencing specifics about that company. In under 60 seconds.
What that does to the research loop:
- Contact identification: handled
- Company context and research summary: handled
- First-draft cold email with specific hooks: handled
- What's left for the SDR: review, adjust tone, verify the contact is real, add personal context if they have it, send
That review step takes 3–5 minutes, not 30. Your SDR still makes the call on whether to send. They're not outsourcing judgment — they're outsourcing assembly.
A Real Example: 10 Prospects in 60 Seconds
Input: B2B SaaS company selling sales coaching software. Target: VP of Sales or Director of Sales at mid-market SaaS companies (50–500 employees) in the US.
Output: 10 companies with real names, websites, estimated headcount, contact name + title + email, a 3-4 sentence research summary with specific context, and a personalized cold email referencing that context. Total time: 47 seconds.
What the SDR does: Scans each prospect card, swaps any email that looks off, adjusts tone on two of the drafts, exports to their outreach sequence. Time spent: ~12 minutes for all 10.
That's not a theoretical productivity gain. It's a real shift in how the day is structured. Instead of researching 10 prospects before lunch and sending emails in the afternoon, an SDR can work through 40–50 prospects in the time previously spent on 10.
The quota math changes. Not because the rep got better at selling — but because the friction before selling dropped by 80%.
What Actually Changes (And What Doesn't)
AI prospecting tools don't fix a bad value proposition. If your product doesn't solve a real problem, faster outreach just means reaching more people faster with something they don't want.
They also don't replace the SDR's role in the conversation. Once a reply comes in, that's human territory. The AI gets them to the starting line. What happens next is still sales.
What does change: the amount of pipeline that gets worked in a week. And in a world where pipeline is the constraint, that matters a lot.
If your reps are currently researching for 4 hours to send 10 emails — and you could get the same output in under an hour — the question isn't whether AI prospecting is worth it. The question is how fast you can move.