Question 1 · Exposure
What is at risk if they walk
Revenue and orders at risk per account and per product, priced on the one lever the data actually identifies: how coverage moves volume.
Revenue intelligence · commercial strategy
Know what a customer decision costs before you concede it.
Ask in plain English. Get the dollar figure, the reasoning behind it, and the limit of what it can claim, in the same answer.
Scroll to see it priced
- Book measured
- $15.46B gross revenue
- Accounts priced
- 171
- Median answer
- 10.8 ms
- Levers identified
- 1 of 4
pharmexus.app/app · chat workspace
Answer
based on how volume responds, not assumed
$0.00BGross revenue
0.00MOrders
0.00MUnits
If Meridian Corp dropped all 5 contracted products, that is the loss, measured from how volume responds to account coverage, not assumed.
Contract scenario data
· strong, reliable effect
· 10.8 ms
01
$0.00B
Gross revenue at risk
02
0.00M
Orders exposed
03
0
Accounts priced individually
04
0
Contract changes tracked and accounted for
The whole point
The loss rate is measured from five years of real account history. Not a number someone typed into a slide.
38,494 product-account observations
894 products
52 regions
a strong, statistically reliable effect
Three questions a commercial team actually gets asked.
Each one resolves to a figure backed by real account data, with a plain statement of what it can and cannot claim.
Question 2 · Sequence
What to do first, and who owns it
A ranked action queue. Every row carries an owner, a quantified value, and the reasoning that chose it, never a score with no arithmetic behind it.
Question 3 · Drift
What changed between filings
Every contract move across quarterly filings, ranked by revenue at stake. All 27,672 tier movements are kept as a change log, giving a complete record of what happened, even where the effect on revenue is not yet reliable enough to price.
Four levers tested. One survives.
A lever that is not reliable contributes nothing to a dollar figure. That rule is enforced in the pricing logic, not just stated in a footnote.
| Lever | Effect on revenue | Likely range | Confidence | Priced at | Reliable enough to price |
|---|---|---|---|---|---|
| Coverage breadth | Strong positive effect | Consistently positive | Very high | measured effect | ✓ |
| Approval gate | −6.7% here | Could be flat or negative | Low | −14.8%, from five years of history | — |
| Pricing tier | −0.2% per tier | Could go either way | None | zero | — |
| Onboarding step | +1.9% | Could go either way | None | zero | — |
A single snapshot on its own is not enough to show the approval gate's effect; the five-year history shows it clearly at −14.8%, because it tracks the same product and account before and after the gate appears. Both views agree that coverage dominates and that tier does not matter, so the dashboard shows both side by side rather than quietly picking one.
Move a lever. Watch it price.
Real figures, not a mock-up. Pick an account and a contract move, and it prices the same way the live tool does.
Account
Contract move
Contract loss · gross revenue at risk
$0.00B
$0share of the $15.46B book$15.46B
0.00MOrders at risk
0Contracted products affected
—Effect applied
Allocated by account size, since aggregate data never links a single order to a specific account.
Live scenario pricing· gross spend, pre-discount
Access resolves state by state.
Gross revenue by state, from the geographic extract bundled with this build. Hover a tile.
Low
High
Unshaded tiles are outside this extract, not zero.
Questions that come up in review.
A tier move is priced only when the underlying effect is strong and consistent. The 27,672 observed movements are all recorded, because what an account did is a fact worth keeping, and the pricing model applies the effect only where it holds up under testing.
It is the sum of independent single-account withdrawal scenarios across 171 accounts and 13 client products, against a $15.46B measured book. Read it as the ceiling of the exposure the portfolio carries, which is what tells you how much negotiating capital a given account relationship is worth.
Yes, it comes back with the answer, front and center. A question in plain English resolves against the same business definitions your team already uses, and the tool reads data through a secure, access-controlled path. Answers come back in a fraction of a second, with every one of its built-in test questions answered correctly, including follow-up questions.
It is tested on products it has never seen before, which is a stronger test than holding out random rows, since that would let it get familiar with a product in training and then recognize it again in the test. Held out on a fifth of all products, it still checks out: the coverage effect holds the same direction every time it is tested, all nine business rules hold, and it reconciles against every published total.
The same logic runs on a laptop or in the cloud, so what is tested locally is exactly what reaches production. In the cloud it runs on a managed schedule inside your own environment, and the answer tool runs as a secure, access-controlled service alongside it.
Ready when you are
Ask it something a slide cannot answer.
Open the workspace and type the question you would normally spend a week building a deck around.