Anatomy of a decision record
One question, three options and a ₹10 lakh budget. This is the sample record from GodEngine's engine documentation, laid out entry by entry.
An illustration, not a customer case. The figures are the documentation's example. They are not advice about any machine, vendor or business, and not investment advice.
The request
Question: “Should we buy a CNC machine, lease one, or outsource machining over the next 2 years?”
| Source | Evidence |
|---|---|
| finance | “We have a 10 lakh budget for this.” |
| vendor | “Lease quote is ₹30,000 per month.” |
| cfo | Value 10, unit lakh |
| Option | Upfront | Monthly | Upside | Risk | Time to start |
|---|---|---|---|---|---|
| Buy outright | 45 lakh | none | 0.9 | low | 2 months |
| Lease | 1.5 lakh | 30k | 0.6 | medium | 1 month |
| Outsource | none | 25k | 0.35 | medium | 0.5 months |
Constraint: budget “10 lakh”, counted by default as the total spend over the horizon.
Recommendation and alternatives
| Rank | Option | Score | Fits the limits | Spend over 24 months |
|---|---|---|---|---|
| 1 | Lease | 0.5756 | Yes | ₹8,70,000 |
| 2 | Outsource | 0.5408 | Yes | ₹6,00,000 |
| 3 | Buy outright | 0.3074 | No: over budget | ₹45,00,000 |
Reasons as the record words them:
- Lease: “Costs ₹8.70 lakh over 24 months, within the ₹10 lakh budget (87% of it).”
- Buy outright: “Costs ₹45 lakh over 24 months, over the ₹10 lakh budget by ₹35 lakh.”
The record also lists its assumptions, for example: “Horizon of 24 months read from the question ("2 year").” Outsource's spend is ₹25,000 a month for 24 months.
The five assessments
Each answers a different question, so the record never blends them into a single figure, and none of them authorises anything.
Confidence
Did it understand the question?
intent: decision, 0.8
domain: financial, 0.71
Coverage
How much was given?
score: 0.758
grid given: 91.7%
cells defaulted: 1
freshness: not assessed
Uncertainty
How settled is it?
calibration: uncalibrated
method: scenario comparison
decisiveness: 0.43
score margin: 0.0348
Feasibility
Does it fit the limits?
status: feasible
fits: Lease, Outsource
Buy: over budget
Authorisation
May anyone act on it?
status: not authorised
executable: no
- Coverage leaves freshness unassessed because the evidence carries no collection time.
- The record's internal model confidence, 0.746, is labelled “an internal diagnostic, not a calibrated probability”.
- On authorisation the record says: “A decision record is advice. Acting on it needs a separate authorization bound to an execution plan; this service grants none.”
Reversal conditions
| If | Then |
|---|---|
| Budget rises to at least ₹45 lakh | Buy outright becomes the recommendation. |
| Budget falls below ₹8.70 lakh | Outsource becomes the recommendation. |
| Outsource's upside is at least 0.46 | Outsource overtakes Lease. |
Read together with the narrow score margin of 0.0348, the third condition shows how close Lease and Outsource are. That is the record's honest answer to “how sure is it?”
What was modelled, and the housekeeping
| Item | Value |
|---|---|
| Upfront cost | ₹1,50,000 |
| Monthly cost | ₹30,000 |
| Total over 24 months | ₹8,70,000 |
| Budget used | 87% |
| Within budget | Yes |
| Dimension | Weight |
|---|---|
| Affordability | 0.2688 |
| Upside | 0.3226 |
| Risk | 0.3011 |
| Speed | 0.1075 |
- Horizon: 24 months, taken from the question. Currency: INR. Budget: ₹10 lakh, counted over the whole horizon.
- Applicability limit: “Holds only for the 3 options given; an option not listed was not considered.”
- Evaluation status: unevaluated, until an outcome is recorded against it.
- AI review: disabled. Engine version: ge-decisions/1.1.0. The record carries a SHA-256 digest of its canonical inputs.