The boutique flow · how we run an engagement

A boutique flow, walked end to end. People do the work that needs taste; AI does the work that needs volume.

Three moves make an engagement land: a hand-picked list anchored to real buying context, a first message written with intent, and an AI-assisted cadence that keeps the conversation moving after the opener goes out. This is how each of those moves actually runs inside Pitchgrove — and where the baton changes hands between human and machine without dropping the thread.

Per engagement: 40–80
Per sequence: Opener + 2
Per week: Hot reply

Section 01

Hand-picked prospects per account. With buying context, not guesses.

A boutique list of 40–80 accounts, not a scraped 10k. We sit with your VP Sales or RevOps lead for an hour first; the disqualification list is written before any prospect is touched. Each row carries the buying context we'll anchor the message to — recent hire, posted RFP, a stack change, a funding event.

What the human team does

Lead

Every judgment call on selection, qualification, and ICP-fit comes from a person.

Account selection
Picks 40–80 rows that match the locked ICP — firmographic, technographic, and the triggers that actually predict a closed-won.
Trigger reading
Decides what counts as a real signal before a touch is spent on it. Recent hire, posted RFP, a stack change, a funding event — or nothing.
ICP-lock dialogue
Sits with the buyer's VP Sales or RevOps lead for an hour. Writes the disqualification list before any prospect is touched.

What AI handles

Support

Refresh work that should never have to wait on a person.

List enrichment refresh
Re-pulls accounts, contacts, and role mapping on every cycle — the list doesn't silently go stale between research passes.
Stack hints & trigger surfacing
Continuous scan for buying signals so the researcher walks back to a list with fresh context, not a stale snapshot.

How the handoff keeps the conversation moving

The researcher hands the list — with notes — to the conversation designer. By that point the designer has enough context to write the opener without re-researching the row. Nothing is touched while the list is still in motion; the disqualification list blocks new names from entering mid-cycle, so a message never lands on a row the team hasn't agreed to call.

The handoff

Where the baton changes hands, step by step.

Click a step to see who owns it, what AI supports, and how the handoff keeps the conversation moving without anyone losing the thread.

Prospect

Human-led

Human researchers select accounts, read triggers, and map the surface area. AI refreshes enrichment between cycles and surfaces new triggers as they appear.

Section 02

First message written with voice and intent. Not spun from a template.

A conversation designer drafts the opener, two follow-ups, and the breakup by hand. No templates, no spin-and-swap. Every message references the account's actual surface area — the trigger from the list section above — so the buyer reads a note that was clearly written about them.

What the human team does

Lead

Voice, tone, and the words that show up in the buyer's inbox.

First-touch copy
The opener is written from the buying context on the row, not from a snippet library. Voice is the buyer's voice; words are ours.
Follow-up arc
Two follow-ups that add a new angle each time — never rephrase of the opener.
Breakup & tone review
Closes the sequence cleanly so the row stays warm for a future campaign. Each message gets a tone pass before it ships.

What AI handles

Support

Channel, timing, and pattern-matching the buyer's reply — never the words.

Delivery channel & timing
LinkedIn-safe send windows, channel selection per contact. AI never rewrites a line we sent on your behalf.
Reply pattern-matching
Classifies replies as the sequence runs. Patterns a hot reply in real time so the human team can route it before intent cools.

How the handoff keeps the conversation moving

The designer hands the sequence — opener + two follow-ups + breakup — to the execution layer. The execution layer queues the cadence; replies route back to a human the moment intent is detected, so the buyer talks to a person, not a model, from the first reply onward.

Section 03

AI-assisted cadence for follow-ups and enrichment.

After the first touch, the cadence is throttled by AI on LinkedIn-safe windows across channels. AI refreshes the contact and the account on every touch, so a follow-up still points at the row's real surface area weeks later — not at a stale snapshot from week one.

What the human team does

Lead

Editorial calls and pipeline review on a weekly cadence.

Weekly reply review
Reads the reply themes: which angle is converting, which is dead, which needs a reframe.
Drop & expand calls
Owns the editorial decision on whether the cadence keeps going on a given account, or whether the row moves to a different motion.

What AI handles

Scale

Repetitive, throttled, verifiable. The volume work the human team should not be on.

Multi-channel orchestration
LinkedIn-safe throttled send windows, channel hand-offs, and cadence scheduling across email, LinkedIn, and reply prompts.
Contact & account refresh
Re-pulls contacts and account context on every touch so the follow-up still points at the row's real surface area weeks later.
Reply classification & routing
Intent scoring on every reply. Hot replies land in the buyer's hands in minutes, not at the next morning's standup.

How the handoff keeps the conversation moving

Hot replies go straight to the buyer within minutes. Mid-intent replies get a model-drafted follow-up that the human team edits before it sends. Cold, out-of-office, and auto-replies stay in the cadence — we don't burn human review time on noise. Every reply is qualified by a human before a meeting is booked; we never book a calendar slot from a model's "high-intent" guess.

Read alongside this

The posts that put numbers behind the methodology.

The methodology walks the flow. The blog reads it back through three standalone posts — the AI SDR landscape, the per-tier dollar math, and the pre-ship compliance guardrail a sequence clears before send.

Open the blog →

Run this flow on your own ICP

If this is the model you want running, this is the call.

Send a line about your stage, ICP, and what you'd need to see in the first 60 days. We'll come back with a yes, a no, or a referral — and we'll point you at the kickoff week for the engagement this page walks through.

Prefer email? Reach us at pitchgrove@polsia.app.

Start a conversation

Tell us about your stage, ICP, and first-60-day target.

We'll come back within one business day — yes, no, or a referral.