The Engine

It builds the case. Then makes the call.

Premis works your problem the way a consulting case team does — specialist agents build independent reads, a lead weighs them and commits to one recommendation. The whole chain lands in a single brief — every figure traced to its source.

A Premis brief: an answer-first read, then the real decision, confidence, what would make it wrong, and the fastest test.
A real read — answer first, then the working · click any screen to read it full size
The Case Team

Independent reads on your problem. One call it stands behind.

Specialist agents argue different root causes, each building its own read. A lead weighs them and makes the call the evidence supports. And the engine convenes a case team at more than one point in the brief, so every layer is a verdict — the way a real case team reaches one.

Independent specialist reads
A lead that commits
No averaging, no hedging
A Premis brief disclosure titled How This Read Was Stress-Tested: three competing reads with their own conviction — Demand & Market at 62%, marked Prevailed; Execution & Capability and Structure & Operating Model at 72% — then what decided it, and the strongest argument still standing from the losing reads.
The stress test, on the record — from the example brief · click to read
Real computation

It does the math. And shows the working.

When your problem carries numbers, Premis computes them deterministically — break-evens, runway, sensitivities — from your figures, with the arithmetic in view. Every quantity is checked back to a line you actually wrote; anything it can't source is dropped, so no number is quietly invented.

Exact arithmetic
Anchored to your words
Unsourced figures dropped
Computed figuresFrom your numbers · verified
Break-even revenue
= $48,000 ÷ 0.32
anchored to "overhead runs about $48k a month" · "margins sit around a third"
$150,000
verified
Runway at current burn
= $900,000 ÷ $120,000
anchored to "roughly $900k in the bank" · "burning about $120k a month"
7.5 mo
verified
Intellectual honesty

It tells you what it doesn't know.

Every read names the evidence it was working without, and how that limits the call — then offers to sharpen the moment you add it. The one thing a confident chatbot never does.

Grounded in your own words
Flags the missing evidence
Sharpens on a re-run
A Premis brief disclosure titled Analysis Operating Without: four numbered gaps the analysis does not have, stated plainly because the call rests on them, with an add-and-run-again action.
Analysis operating without — the gaps, named · click to read
What it already knows

It walks in knowing your world.

Premis carries context that has to be earned over time: your organisation's shape and constraints, the outcomes of your last three decisions, and a live read on the operating climate. Every read is grounded in it — and cites it — so the advice fits your situation, and gets sharper the longer you use it.

Standing contextCarried into every read
Org memory · per client
"Twelve people · services-to-product transition · board is risk-averse · two failed pricing changes already tried" — your shape, constraints, and what's been tried, remembered across runs.
Your last three outcomes
"Day-60 check-in — plan on course, diagnosis held: the pricing change lifted win rate" — Premis knows how its own past calls performed for you, and reads the next one against them.
The operating climate
"Capital conditions · sector adoption pressure · hiring market" — the operating environment of the current quarter, read live and carried into every brief.
Two depths

A 60-second read, or the whole brief.

Not every decision needs the full engagement. Premis reads your problem and recommends the depth — you confirm with one click. It scopes the work with you, the way a partner would in the first meeting.

Premis recommends the depth · you confirm
First Read~60 sec
A verdict, fast.
  • One clear call, with the math
  • What would make it wrong
  • One measure to watch
Full Briefthe engagement
The whole case.
  • A staged diagnosis, grounded and checked
  • Three options, one recommendation
  • A 90-day plan — then the room
After the brief

The brief doesn't disappear. It opens a room.

Every Full Brief becomes a live 90-day workspace. Tasks get owners, the quarter unfolds in one view, and at 30, 60 and 90 days Premis comes back to check whether the call held — and feeds what happened into your next read.

A live 90-day plan
30 · 60 · 90 check-ins
Outcomes feed the next read
The Premis Situation Room: a live engagement list, a due day-60 check-in, and the recommendation Selective closure, focused reinvestment.
The Situation Room — the plan, alive at day 63 · click to read
Inputs & reach

Bring anything. Send it anywhere.

Bring in
Voice
Text
Documents
Research
Send out
Linear
Notion
Jira
PDF
Excel

This is what stands behind one brief.