Decision architecture

How THESIVE Thinks

THESIVE does not jump from data to an answer. It builds a decision graph: evidence connects to context, memory, portfolio state, risk and opportunity cost before a proposal is allowed to reach a human.

Ingesting evidence
LIVE DECISION GRAPHExplainable reasoning
Price + volume
Fundamentals
Filings + catalysts
Macro + sector
Portfolio state
Decision memory
THESIVEReasoning
core
evidence in context
Thesis
Timing
Risk
Opportunity cost
ProposalADD · HOLD · WATCH · TRIM · EXIT
HumanAPPROVE · HOLD · REJECT
DECISION STATE
Thesisvalid
Riskbounded
Timingchecking
Authorityhuman
signal movingEvery proposal keeps its evidence + memory trail
The decision loop

From market universe to accountable action.

Strategic thesis state and tactical state are kept separate. Recommendations use persistence and hysteresis, so THESIVE does not flip from BUY to EXIT because a candle changed colour.

01UNIVERSE

Eligible Indian equities, ETFs and portfolio instruments.

02DISCOVER

Proactive scan for credible ideas, not only held names.

03QUALIFY

Thesis, fundamentals, valuation, catalysts and evidence.

04TIMING

Support, trend, confirmation, volume and no-chase filters.

05PORTFOLIO

Sizing, concentration, risk and opportunity cost.

06MEMORY

Compare with prior decisions, reversals and missed entries.

07PROPOSE

ADD, HOLD, WATCH, TRIM, HARVEST or EXIT. Human decides.

Key variables

Many inputs. One explainable judgement.

Valuation

Relative value, earnings context, margin of safety and capital already committed.

Price & technical state

Support-first entry, trend, momentum, volume, compression and confirmation.

Fundamentals

Growth, margins, cash generation, leverage, return ratios and balance-sheet direction.

Catalysts

Results, orders, capex, policy, commodity moves and events that can change the thesis.

Sector & commodity graph

Leaders, laggards, peers, upstream/downstream links, ETFs and shared versus specific drivers.

External conviction

Credit-rating actions and analyst revisions are tracked separately, then compared with price behaviour.

Risk assessment

Invalidation, downside, concentration, liquidity, thesis fragility and transaction effects.

Opportunity cost

Compare holding, adding, trimming or exiting against credible replacements and cash.

Process memory

Rejected ideas, false positives, missed qualified entries, recommendation churn and outcome quality.

Action vocabulary

More useful than BUY or SELL.

Profit managementHARVEST
Risk / thesis repairTRIM
Capital deploymentADD
Default when validHOLD

THESIVE also uses WATCH and EXIT. Price decline alone is never an EXIT reason. Loss-making trims with a recoverable thesis face a high hurdle and are compared against recovery potential, crystallised loss, opportunity cost and portfolio risk.

Cash is valid. Stopping discovery is not.

If no READY trade exists, THESIVE distinguishes NO QUALIFIED BUY from NO CREDIBLE IDEAS and keeps a ranked NEAR_ENTRY pipeline with the missing price, catalyst, pullback or confirmation stated explicitly.

Data sovereignty

Public first. Primary source first. Vendor optional.

THESIVE is designed so no premium commercial data vendor can become a kill switch for the decision brain.

Primary-source intake

Exchange reports, company filings, rating agencies, regulators and government/macro sources wherever legally usable.

Local persistence

Market history and decision evidence are stored with source, event time, ingestion time, revisions and provenance.

Derivation first

THESIVE computes its own support, momentum, valuation context, debt trajectory, sector state and decision variables.

Vendor optional

Commercial vendors may enrich, validate, accelerate or cross-check. They must never be REQUIRED_FOR_DECISION.

Longitudinal memory

Every operating day makes the proprietary evidence base richer, reducing dependence rather than increasing it.

Human authority

Better data does not change the constitutional rule: every trade remains proposed until explicitly approved.