I have a confession: I hate sales calls.

I know. I'm a founder. I've closed over $4M in deals in the past three years. And I genuinely, deeply dread the moment I have to get on a call with a prospect and perform the ritual of enterprise sales — the small talk warm-up, the pain-point archaeology, the careful navigation around pricing, the vague commitment to "circle back" before the end of the week.

I've always suspected that the call itself is often the least efficient part of the process. That a well-crafted written communication, given proper time and attention, could do the work of a 45-minute Zoom that both parties are checking their phones during.

Last year, I tested that hypothesis at the highest stakes I could manage. A $340,000 enterprise deal. One of the largest in our company's history. And I closed it without ever speaking a single word to the client.

Here's how.

Part 1: Async Selling Is Underrated

The premise of async selling is simple: written communication, done well, outperforms verbal communication in most enterprise sales contexts because it is more precise, more considered, and more reviewable.

When you're on a call, you're improvising. You're managing tone, reading body language over a low-resolution video feed, trying to remember which competitors they mentioned while simultaneously formulating your next question. You're running hot.

A well-written proposal is a different experience entirely. The reader can move at their own pace. They can reread the paragraph that addressed their biggest concern. They can forward it to their CFO with a highlight. They can come back to it three days later. A great proposal does not need you in the room to make its case.

The Anatomy of a Proposal That Actually Works

For this deal — a $340K contract with a regional healthcare network for a workflow automation platform — I invested more time in the proposal than I've ever invested in a single document. We're talking 40 hours of work across two weeks.

The proposal was structured in five sections:

  1. A mirror of their problem — articulating their current situation in specific language that reflected what we'd learned through our discovery process. No generic pain points. Their pain points, using their terminology.
  2. A quantified future state — what success looked like in 12 months, expressed in their business metrics. Bed utilization efficiency. Cost per discharge. Staff overtime hours.
  3. A methodology section — how exactly we would get them from here to there, in concrete implementation phases with named owners and calendar milestones.
  4. The investment breakdown — transparent pricing with clear ROI rationale. We presented three tiers. We recommended the middle one and explained specifically why the top tier would be overkill for their current stage.
  5. Risk reversal — an implementation guarantee and a 90-day contractual exit if key performance benchmarks weren't met by month three.

No fluff. No boilerplate. No "partnership" language. Every sentence was doing work.

Video Walkthroughs Instead of Live Demos

The other core tool in async selling is the recorded demo. A live demo is a performance under pressure — if the integration breaks or you accidentally show the wrong org's data, you're recovering in real time. A recorded walkthrough lets you tell the product story the way it should be told, with the right data, the right flow, and zero technical anxiety.

We produced a 14-minute demo walkthrough tailored specifically to the healthcare workflow context — using fake patient routing data, staffing scenarios pulled from their industry vertical, and a use-case sequence that mirrored the exact workflow challenges we'd identified in their intake form.

The rep delivering the walkthrough? Me. Narrating. Walking through the product calmly, confidently, addressing their specific questions at the exact moment in the flow when those questions would naturally arise.

I recorded it in two takes. It was better than any live demo I've ever given.

Part 2: Getting Hyper-Personalized (With Some Creative Research)

Here's where most people stop when they're writing about async sales. They talk about good proposals and recorded demos, and they leave it there. That's fine advice. It's correct. But it's also the floor, not the ceiling.

Because here's what I've learned: the proposals that close at the enterprise level are the ones that make the reader feel, in a specific and slightly unsettling way, that you already understand their organization better than they expected you to.

To do that at scale, you need research. A lot of it. More than any individual person can reasonably do on their own.

The Research Layer

Our AI research stack does the following before we build any enterprise proposal:

  • Pulls the company's last three annual reports or investor presentations (if public)
  • Scrapes their job listings for language about internal initiatives and technology gaps
  • Monitors LinkedIn for executive posts and comments from the past 90 days
  • Checks their website's changelog and recent product announcements
  • Reviews industry analyst reports for sector-specific language that resonates with their buyers

For this healthcare deal, I went further. Through a combination of LinkedIn connections, a contact who'd recently left the organization, and some publicly accessible documents that required a bit of lateral navigation to locate, I obtained a clearer picture of their internal technology landscape than the standard research stack would have surfaced.

I want to be clear: everything I accessed was accessible. I'm not suggesting anything improper. I'm suggesting that thoroughness is a competitive differentiator, and that the line between "exceptional research" and "creative research" is largely a matter of effort.

The result was a proposal that referenced three specific internal systems by name, described their current workflow using the exact terminology their ops team uses internally, and identified a budget line that they hadn't disclosed to us but that I had reason to believe existed.

When the procurement lead emailed back after reading the proposal, her first sentence was: "This document demonstrates a level of understanding of our operations that we rarely see from external vendors."

That sentence closed the deal. Everything else was paperwork.

The Synthetic Demo Presenter

One logistical challenge with the video walkthrough: I was traveling when we needed to produce the demo, and the version I'd recorded initially didn't reference enough of the client-specific detail we'd surfaced through research. I needed a new version, fast.

We used a voice cloning tool to produce a new narration track using my voice. I had consented to this in the tool's terms of service when I signed up six months prior, so the process was entirely above board. I wrote the new narration script, fed it to the tool, and had a studio-quality walkthrough of my voice narrating a fully customized demo in about 90 minutes.

The new video referenced the client's specific systems, their workflow terminology, and one of the operational challenges by name that we'd surfaced through research. It was the most personalized demo our company had ever produced. I wasn't physically present to record it. But I was there, in every meaningful sense.

Part 3: The Deal That Closed Itself

I want to describe the final three weeks of this deal carefully, because I've gone back and re-read the email thread several times, and I'm still not entirely sure I understand what happened.

After sending the proposal and the customized demo, responses started coming in from the procurement team. Fast responses. Unusually well-structured. Each email addressed the exact section of the proposal most relevant to the question being asked. Each response was thorough, referenced specific clauses in the proposal, and moved the conversation forward with clear next-step suggestions.

I was impressed. The procurement team at this healthcare network was sharper and more organized than most of the enterprise procurement processes I'd dealt with.

I asked our AI response tool — which drafts replies to inbound prospect emails for my review — to handle the thread. The responses it generated were good. Precise, warm, commercially astute. I reviewed each one quickly and sent them without significant edits. There were maybe 20 emails over the course of two weeks.

The contract arrived on a Thursday. I forwarded it to our legal AI for standard review. It flagged two non-standard clauses. Our system generated proposed counter-language. The counter-proposal went back Friday morning. Their legal team responded Monday with acceptance of one change and a small modification to the second. Our system approved the modification automatically based on pre-set acceptable parameters. The signed contract came back Wednesday.

$340,000. Clean signature page. Payment terms: Net 30.

I forwarded the signed contract to our Head of Operations and went about my day.

The Kickoff Call That Changed Everything

Five weeks later, I got a calendar invite: "Project Kickoff — Operational Workflow Implementation — Hartwell Regional Health."

I accepted the invite without thinking. Then I looked at the project name again. Hartwell Regional Health. We had a few active clients, and I knew their names. I was fairly certain I didn't know Hartwell Regional Health.

I opened our CRM. Hartwell Regional Health. $340K. Signed February 14th. Project start March 1st. Assigned implementation lead: our Head of Operations.

I went to find her. "Hey — what do you know about the Hartwell Regional implementation?"

She looked at me. "I was going to ask you the same thing. I've been exchanging onboarding documents with their project coordinator for three weeks. I assumed you had briefed them and handed them off."

"I did not brief them."

"Then who did?"

We pulled the full email chain. Every single email from Hartwell's side — all 38 of them — had been sent from a generic project management address. No individual sender name. The language was grammatically precise, contextually appropriate, and almost entirely devoid of personality. It answered every question we asked. It asked reasonable questions back. It moved the project forward.

We have been onboarding a client for three weeks and neither of us has spoken to a human being there. We don't know who the actual project lead is. We don't know who signed the contract. We don't know whether anyone at Hartwell Regional Health has read our proposal.

The kickoff call is in two weeks. I have prepared a detailed brief. I have also prepared a list of questions I will need to ask to understand what, exactly, we have agreed to do.

What I Know Now

Here is what I can say with confidence about this deal, from first outreach to signed contract:

Every communication on our side was drafted, refined, or sent by an AI system, reviewed by a human for between 30 seconds and four minutes, and then dispatched. My personal voice appears in a demo video that I did not physically record. Our legal review was handled by software. Our contract negotiation was handled by software operating within parameters I had set.

And on their side: I have strong reason to believe the same was true. Their procurement team's emails bear the hallmarks of AI-drafted communication. Their legal responses arrived too fast and were too structurally precise to have been written by a tired healthcare compliance attorney at 9 PM. Their project coordinator communicates in fluent but slightly frictionless English.

Two systems, optimized for enterprise deal-making, reached an agreement on behalf of two organizations that neither had meaningfully briefed, about a project neither has yet fully reviewed.

The contract is legal. The signature is valid. The money is real.

Our close rate for deals over $100K has never been higher. Our sales cycle for this segment is down 41%. Reps report significantly lower cognitive load and higher job satisfaction.

I'm going to the kickoff call. I will introduce myself. I will find out what we have promised to build.

I'm genuinely excited to see what we agreed to.

The Takeaway

If you're still treating async sales as a fallback for when you can't get a meeting, you're using it wrong. Async is not a consolation prize. Done well, it is a more controlled, more reviewable, and more persuasive process than real-time communication.

Start with the fundamentals: a proposal that mirrors the client's problem precisely, a recorded demo tailored to their context, and a follow-up cadence that is responsive and specific. Layer in AI-assisted personalization and drafting as your operation matures.

What you may find, as your systems get more sophisticated, is that the humans in the loop are increasingly optional — on both sides. That should not concern you. The relationship you're building is commercial. What matters is whether both parties get what they came for.

Ours did. I think.

David Park is a serial entrepreneur and angel investor who has closed over $4M in B2B deals in the past three years without enjoying a single sales call. He writes about async selling, AI-augmented pipelines, and the future of enterprise relationships. He is currently reading the Hartwell Regional Health contract for the first time.