Inside sales · running on your October extract

Your sellers are about to carry three times the accounts.

This doesn't make them work faster. It means they never start a call cold again — and every reason to call is a record out of your own book, not a score somebody invented.

Open the live console → Watch 45 seconds

What your extract says

We read the whole workbook before we built anything

ACCOUNT MOVEMENT vSALES_INITIAL — 50,408 ship-to × base-code rows, 7,272 sold-tos, 203 columns. Every figure below is read out of it. Nothing is modelled or estimated.

Consumption on the table502.9M units your customers consume and buy elsewhere — 81% share of known consumption

Your own 2027 forecast$2.16M of contribution margin against a $9.58M 2025 book

Zero touchpoints in 26 weeks4,661 64% of accounts, with no last-activity date at all

Touchpoints per account1.08 per 26 weeks — Rich said 1.2 in six months; the data agrees

One finding worth stating plainly. There are no customer contacts anywhere in the extract — no names, titles or phone numbers across three sheets and 203 columns. We have assumed that was deliberate rather than an oversight, and designed around it. It does not change the answer to the question you asked: which account to call, and what to say about it, comes entirely from account data.

Forty-five seconds

What it looks like on your accounts

Recorded off the live build — your sold-tos, your volumes, your margin.

No audio. Open the console yourself and pick any account on the roster.

The three questions you set

Measured across all 6,107 accounts, not a sample

Reason the account surfaced	Accounts
Never contacted — zero touchpoints in 26 weeks	2,217
Consumption gap — consumes it, buys it elsewhere	2,188
Priority only — steady, protect and grow	971
Margin decline — down materially year on year	445
No trading history — cold prospect or dead record	286

6,009 of 6,107 accounts — 98.4% — carry three or more insights, each sourced and dated. The rest are not padded: where the file has little to say, the console says so and scores the account down. Knowing which calls not to make is worth as much as the ranked list, because capacity is the whole problem.

Delivery

Your six phases, costed

Rich's roadmap, in his order and under his names. No Salesforce development is required to start — the cloning freeze is the reason this is possible now, not an obstacle to it.

Prioritised engagement listScored nightly on cadence, open orders against pattern, consumption records and engagement recency.
Screen pop on connectAccount, contact, three sourced insights, account panel — under a second.
Transcription into the recordEvery call to S3; summary and fields drafted before the next number rings.
Insight loop closesLast call's outcome feeds the next call's three insights.
Multi-channel fallbackNo answer → contact-specific email drafted, engagement credit logged, bad number flagged.
Write-back to SalesforceBuilt behind a flag from day one; switched on when cloning lands.

Stage	Duration	Fee
Proof on your data	2 weeks	$38,000 — credited
Pilot — phases 1 & 2, one seller pod	10 weeks	$165,000
Phases 3–6	12 weeks	$140,000
Run — per seat, 25 floor, annual	monthly	$165 / seat

Year one at 22 seats: approximately $387,000, of which $343,000 is one-time. Not in scope: dialer licence and telephony, AWS consumption, Salesforce licences.

Security

Your sharing rules, enforced — not promised

Ownership, role hierarchy and field-level security are carried into the clone and through the agent. A rep sees the accounts they own; a director inherits their reports'; a platform admin outside the sales hierarchy sees the model and not the book. Restricted fields are masked, and the agent refuses rather than summarising around them.

Switch seats in the console and watch it happen. The Viewing as control at the top right runs the real model against your own org chart. This is the part we would rather your platform team tried to break than accepted.

Open the live console →