One machine that finds the right accounts, finds the right people, writes the right message, and hands a rep a ready-to-work list every morning. This page is how it works, and where I want your read.
The idea in one breath. Clay is an always-on production line for pipeline. The Golden List is the intelligence layer it runs off. Together they turn 5,000 credits into a motion that repeats. The machine does the manual work. People keep the judgment.
Not a spreadsheet. A live account intelligence layer that every campaign, sequence, and agent runs off, and that updates as the market moves.
A record does not just sit there. It gets re-scored when a new signal fires and re-enriched when it goes stale. Six things live on every record, which is what makes one list enough to run the whole motion.
Which companies to prioritize, by fit and intent.
The right people and verified emails, already attached.
Why now, layered in from hiring, tech stack, news, filings.
Each account flows straight into a message set.
Effort lands on the accounts most likely to convert.
Every reply and meeting traces back to the list.
| Before the list | With the Golden List |
|---|---|
| Rep researches accounts one at a time, by hand. | Every account arrives pre-scored and graded. Effort goes to A and B accounts, not guesswork. |
| Each email written from scratch. | Messages are pre-drafted and on-message. A BDR sends and edits instead of writing cold. This is the throughput lever. |
| Sales and marketing work off different lists. | One list for both motions and both teams. Everyone acts on the same trigger at the same time. |
| A good account goes cold before anyone notices. | Signals re-rank the list as news breaks, so timing is built in, not chased. |
Everything that follows, the scoring, the messaging, the Clay machine, runs on this one list.
Every account carries two scores. Fit is how well the company matches who we win. Intent is how ready they look right now, read from live signals. The two combine into a single letter grade.
Fit + Intent → Grade. A = work now · B = strong fit, warm it up · C = keep on the list · D = drop. A rep never has to decide where to start; the grade already did.
We do not chase every plan sitting under 4.0 stars. We rank on the bonus dollars we can actually influence inside the window that is still open, not the whole theoretical number. That cut is what keeps the list credible and the effort aimed.
Figures modeled from public CMS Star Ratings data, ranked on the influenceable 2028 measurement window. First target inside Tier A+B is Humana at roughly $1.08B of that window. These are analytical estimates, not Intradiem-verified customer results.
On top of fit and intent, a tiered signal library decides how loud an account gets. This is what turns the list from a ranking into a clock.
October CMS release · Stars language in SEC filings · earnings-call mentions. Fires outreach immediately with a fresh angle.
New quality leadership · quality hiring clusters · a contract just under 4.0. Raises priority and refreshes the angle.
Lifts priority but does not fire outreach on its own.
For each contact, the engine picks the angle by role and drafts a short, specific message. We write three variants per segment so we test into the winner instead of guessing which line lands. The quality bar is the approved sample already in use.
Subject: closing the gap from 3.91 to 4.0
Hi Kristen, your weighted average looks to be right around 3.91, just under the 4.0 line, though you will know the exact picture far better than I do. For plans sitting this close, call center measures are often where the last few points come from.
If useful, I would value 15 minutes to hear how you are approaching your cliff-edge contracts.
Subject: one cycle of margin
Hi {{first_name}}, missing 4.0 by a fraction is the same financial outcome as missing it by a lot: one cycle of bonus either way. For contracts this close, the last few points usually come from the call center measures.
Worth 15 minutes to pressure-test the numbers together?
Subject: the work off the phones
Hi {{first_name}}, we already work with {{company}} on the front office; what we don't see yet is how the back office ({{function}}) is handling its volume. Most teams have real-time orchestration on the phones and almost none of it behind them.
Worth 15 minutes to hear how you're running it today?
Launch all three variants to matched samples. Measure to qualified reply, not opens. Kill the lowest, scale the winner, and fold the winning angle into the follow-up. The list gets sharper every week because every send teaches it something.
{{fields}} fill per contact from the Golden List. No fabricated stats; any Intradiem metric is pulled from the verified repository before a message can send.
Clay is built from tables that feed other tables, like stations on an assembly line. A record only moves to the next station if it passes a check at the current one. Nothing is manual once it is wired, and the whole line refreshes itself. It runs eight stations deep. Click a station.
Public CMS Star Ratings data for new-logo targets, and our own customer list from Salesforce for the back office. A record only enters if it qualifies, for example a plan sitting under 4.0 stars that is not already a customer.
Every arrow is an automatic Clay action, not a copy-paste. Reference tables (approved claims, the persona rulebook, the message library) sit alongside the line and are read by the stations that need them.
Clay runs on a fixed budget of credits we spend on enrichment. The plan is built to spend them on the right rows only, show the results, and use that record to renew. Three guardrails keep the spend honest.
A row is only enriched if that field is still empty. Once we have someone's email, we never spend a credit re-finding it.
Out-of-window and out-of-profile rows are dropped before any enrichment runs. We never pay to research a company we would not pursue.
Low-cost company details first; premium data (intent, hard-to-find emails) only on rows already worth it.
On top of that, a running ledger tracks every batch: credits in, contacts produced, messages sent, qualified replies, meetings. That ledger is the renewal case, told in receipts rather than promises. A checkpoint at 60% of the budget forces a review before we spend the back half.
The first three figures are confirmed on the onboarding site. Every dollar and ROI figure the engine generates is labeled as a modeled estimate until an Intradiem-verified number replaces it.
The fastest place to show impact, and it is Mandate 3: 200-plus back-office contacts.
Our front-office customers cannot hand us leads into their own back offices, so we source those contacts ourselves through the same engine, then run campaigns inside accounts where we are already trusted.
We are already in the building on the front office. The back office is white space next door, not a cold start.
The target functions are known: claims, enrollment, billing, payments, document processing. We point the machine at a list, not the open market.
"We run your phones, we don't yet see the work behind them" is a clean, honest opener only we can send.
The back-office target profile gets ratified with Scott Kemme first. His edits are the profile. We review the first fifty contacts together, then the list-build runs against the agreed definition and paces to the Back Office Optimizer launch, not a dump. This keeps the 200 contacts credible instead of a raw count.
Sales Navigator fits the engine best as a feed into the machine we already have, not as a separate tool a rep works by hand. It plugs in at three points.
Its filters (title, seniority, function, geography) feed the Contacts station directly. It is the cleanest way to find the exact buying committee, and it is especially strong for the back-office personas, where public CMS data gives us the account but not the people.
It surfaces job changes, new-in-role moves, and hiring, which are exactly our priority triggers. A new quality leader or a claims-ops hire fires a fresh, timely angle instead of a generic touch.
LinkedIn is already the second touch in every sequence. A light automation layer runs the connection, follow, and message steps in step with the email, so the same contact gets one coordinated push.
Bring it in as a data and signal source first. That upgrade needs no new behavior from reps and immediately makes the engine's contact and timing quality better. Layer the outreach automation second, once deliverability is green and the profile is ratified, so we stay coordinated rather than spraying.
One guardrail to plan for. LinkedIn automation has account-safety limits, so we pace it under daily caps and keep the same human-approval gate we use on email. Same discipline, one more channel.
This is the path I'd take. I'll scope the specific tool and its cost so we're deciding on a real number, not a maybe.
The engine is built and the messaging is drafted. A handful of things sit with you rather than with the build, and I'd rather shape them with you than assume. Here is where your read moves this forward.
Where this is heading. Every one of the new DWO products, Back Office Optimizer and Queue Optimizer first, needs pipeline aimed at the right buyer at the right moment. This engine is that front end. It's built and aimed at the Star Ratings play first, with the back-office motion the next thing we point it at. It's designed to get sharper with every campaign rather than needing a rebuild. That compounding is the point, and it is the part that keeps paying off as the product line grows.