1. Governed inputs & commercial boundary
Buying Insight does not start from a blank prompt. It starts from Scanner outputs, evidence stores, market context and commercial knowledge—while preserving the rule that this workflow creates a need hypothesis, not confirmed buying intent.
Critical trust idea: this workflow should behave like a translator—company reality → structured signals → testable need hypotheses → safe discovery paths. It must not skip straight to “the customer is buying.”
ΣOpen logicPrimary inputScanner Company Summary
Provides the governed account picture: facts, diagnosis, tensions, financial reality and quality state.
Current realityTraceable evidenceSummary + diagnosis
Σ
Open logic
Key question
What already-established account reality should Buying Insight build on rather than re-invent?
Reasoning logic
Use the Company Summary as the evidence-governed base so commercial reasoning does not restart from unconstrained search or impressionistic interpretation.
Grounding basis
General Summary, AI Diagnosis, financial baseline, market intelligence, evidence IDs and publication state from Scanner.
Reliability control
Buying Insight inherits the upstream evidence boundaries. It may reframe them commercially, but it does not get to change the company facts.
Illustrative thinking sample
“The company summary suggests structural modernization pressure. Buying Insight may turn that into a commercial hypothesis, but not into a claim of active purchase intent.”EOpen logicSupporting inputCompany & Market Evidence DBs
Adds company-level corroboration and market pressure signals to widen the activation picture.
CorroborationPressure signalsFreshness matters
E
Open logic
Key question
What additional evidence strengthens, weakens or contextualizes a possible need hypothesis?
Reasoning logic
Widen the commercial basis without replacing the governed account baseline. Company evidence adds specificity; market evidence adds timing and relevance context.
Grounding basis
Governed company evidence, market evidence, public business signals and external pressure indicators.
Reliability control
Market context cannot be rewritten as company action. Company evidence and pressure evidence keep separate roles all the way to the report.
Illustrative thinking sample
“A domain pressure may increase urgency. It does not prove that the company has recognized the need or allocated budget.”KOpen logicCommercial knowledgeService Portfolio & modernization KB
Supplies solution vocabulary and fit logic, but not evidence of customer need.
Service fit laterNo solution-as-evidenceCommercial framing
K
Open logic
Key question
What response directions could become relevant if a need hypothesis later proves valid?
Reasoning logic
Give the workflow a disciplined path from hypothesis to possible plays and service directions without allowing the offering to bias the initial need formation.
Grounding basis
Service Portfolio KB, modernization tactics, readiness logic and service metadata.
Reliability control
The service layer is downstream. It cannot be used to “prove” the customer has the need simply because the provider offers a related service.
Illustrative thinking sample
“Service fit may be strong, but if the account evidence is weak, the result remains a cautious hypothesis rather than a push toward a solution.”⚑Open logicBoundary ruleNeed hypothesis, not buying intent
The workflow creates demand-creation hypotheses and safe discovery directions—not a claim that a deal is live.
Customer-safeDiscovery-firstNo overclaiming
⚑
Open logic
Key question
What can we responsibly say about possible need without asserting that the customer has already recognized or budgeted for it?
Reasoning logic
Establish the central epistemic rule of the whole workflow. The output is a commercial hypothesis generator, not a detector of confirmed intent.
Grounding basis
Scanner evidence, activation hypotheses, gate-review rules and customer-facing language constraints.
Reliability control
Any wording that implies certainty, project ownership or active purchase intent without support must be downgraded, reframed as a question or removed.
Illustrative thinking sample
“‘This account may face growing pressure to modernize’ is allowed. ‘This account is buying modernization services now’ is not.”2. Canonical signal map
Raw evidence is normalized into stable signal lanes so different report sections do not each invent their own commercial truth.
Account facts→Financial capacity→Pressure signals→Contradictions & gaps
SIGOpen logicDeterministic mapCanonical signal structure
Normalizes the account into reusable lanes that every later section must respect.
Stable spineShared truthReduces drift
SIG
Open logic
Key question
What stable signals should define this account before prose generation begins?
Reasoning logic
Convert heterogeneous evidence into a canonical structure so “Account Snapshot,” “Why Now” and “Sales Plays” do not each create their own account interpretation.
Grounding basis
Scanner outputs, financial reality, evidence databases, market signals and contradiction markers.
Reliability control
This is one of the strongest anti-drift controls in the workflow. Once normalized, later sections should reuse the same signal spine instead of improvising new truths.
Illustrative thinking sample
“The account may have modernization pressure, cost sensitivity and uncertain stakeholder ownership. Those become canonical signals reused across the report.”!Open logicUncertainty laneContradictions & evidence gaps
Records uncertainty, opposing signals and what must still be tested through discovery.
Unknowns matterCounter-signalsQuestion fuel
!
Open logic
Key question
What do we not know, and what signals actively weaken or complicate the need hypothesis?
Reasoning logic
Prevent commercial overconfidence by treating contradiction and uncertainty as first-class elements of the reasoning model.
Grounding basis
Weak evidence labels, absence findings, blocked claims, opposing account signals and inconsistent patterns.
Reliability control
These are not side notes. They can reduce priority, change recommended plays and force claims to become discovery questions.
Illustrative thinking sample
“The need may be plausible, but unclear budget signals and weak stakeholder evidence should remain visible because they materially affect how the report can be used.”3. Hypothesis Registry
Named, traceable need hypotheses are created from the canonical signals and linked to parent evidence, contradiction markers and later report sections.
HOpen logicCore reasoning anchorHypothesis Registry
Creates structured, reusable HYP entities so the whole report speaks from one commercial reasoning spine.
HYP IDsTraceable lineageShared truth spineExpanded sample model
H
Open logic
Key questions
What are the top plausible need hypotheses? What supports them? What weakens them? What would validate or falsify them? Which later sections should reuse them?
Reasoning logic
Convert scattered signals into a stable commercial hypothesis structure so the whole report can remain internally coherent even when sections are generated separately.
Grounding basis
Canonical signals, parent evidence IDs, pressure signals, contradiction markers and service-neutral need framing.
Reliability control
Hypotheses remain explicitly testable and falsifiable. They are not allowed to harden into “customer intent” simply because multiple sections reuse them.
Critical note for accuracy
This step is central because it reduces cross-section drift. The same hypothesis wording, support set and uncertainty model travel into later generation steps.
Illustrative thinking sample
“HYP-02: The company may soon need stronger coordination between strategic ambition and execution capability. Support: structural friction + pressure signals. Weakener: limited direct proof of internal ownership. Validate by asking about decision bottlenecks and current transformation priorities.”VALOpen logicValidation logicFalsifiers & validation path
Each hypothesis is paired with what could strengthen, weaken or reject it in real discovery.
TestableNot self-confirmingHuman handoff
VAL
Open logic
Key question
What evidence or discovery outcome would validate, weaken or falsify this hypothesis?
Reasoning logic
Force the workflow to think ahead about what would count as confirming or disconfirming feedback. This reduces the risk of elegant but self-sealing commercial narratives.
Grounding basis
Contradiction markers, weak evidence, stakeholder uncertainty and planned discovery questions.
Reliability control
If a hypothesis cannot be falsified or meaningfully tested, it should not be treated as a strong output. It may remain internal or be removed.
Illustrative thinking sample
“If the account already has a well-governed transformation program, the modernization-pressure hypothesis weakens materially. If decision bottlenecks are confirmed, it strengthens.”4. Bounded section generators
The report is created through separated generation tasks, each with its own role. This limits context mixing and makes section-level correction possible.
Shared section contract: each section must reuse the same canonical signals and Hypothesis Registry, but it is allowed to serve a different reader purpose and certainty level.
01Open logicSection generatorAccount Snapshot & Buying Thesis
Defines the account frame and the primary need hypothesis in a compact sales-readable form.
SnapshotPrimary thesisQualification logic
01
Open logic
Key question
What concise account picture and primary need thesis should open the report?
Reasoning logic
Give the reader immediate commercial orientation while keeping the thesis visibly hypothetical and bounded by evidence.
Grounding basis
Account facts, top hypotheses, qualification logic and contradiction markers from the Registry.
Reliability control
The opening cannot sound more certain than the underlying hypothesis. Weak support remains visible or is converted into discovery language.
Illustrative thinking sample
“This account may face increasing modernization coordination pressure. Priority looks moderate-to-high, but stakeholder ownership remains uncertain.”02Open logicSection generatorWhy Now / Why Soon
Explains timing relevance and the conditions that could accelerate or weaken urgency.
Timing logicPressure horizonConditional urgency
02
Open logic
Key question
Why might this account care sooner rather than later, and what could change that timing?
Reasoning logic
Translate pressure signals into a timing hypothesis suitable for strategic outreach or discovery preparation.
Grounding basis
Pressure signals, company events, financial reality and contradiction / disqualifier markers.
Reliability control
Urgency is always framed conditionally unless the account evidence directly supports stronger timing claims.
Illustrative thinking sample
“Pressure appears to be increasing because market and internal coordination signals point the same way—but urgency remains conditional until the account confirms its current agenda.”03Open logicSection generatorPain-to-Solution Map
Connects a testable pain hypothesis to an outcome direction without overstating solution certainty.
Problem framingOutcome directionDiscovery-led
03
Open logic
Key question
If the hypothesis is true, what kind of problem-to-outcome conversation becomes relevant?
Reasoning logic
Help the seller move from abstract pressure to a practical conversation frame without pretending a specific solution path is already agreed.
Grounding basis
Hypothesis Registry, pain signals, commercial patterns and service-neutral outcome thinking.
Reliability control
This section cannot diagnose the customer’s exact pain as fact; it proposes a conversation map to test.
Illustrative thinking sample
“If coordination friction is real, a conversation about decision speed, ownership clarity and service-model coherence may be relevant.”04Open logicSection generatorRecommended Sales Plays
Builds activation plays and later enforces a primary play plus explicit guardrails.
Activation optionsService fit laterGuardrails neededExpanded sample model
04
Open logic
Key questions
Which commercial plays could credibly test the top hypothesis? What is the objective, trigger, disqualifier and service relevance of each? Which one should become the primary play?
Reasoning logic
Translate need hypotheses into actionable next-step options while preserving evidence boundaries, disqualifiers and service realism.
Grounding basis
Hypothesis Registry, signal map, service KB, readiness logic and contradiction / risk indicators.
Reliability control
Plays must stay hypothesis-led and can be downgraded or removed if service fit is weak, evidence is thin or later gate review finds unsafe certainty.
Critical note for accuracy
This section is especially exposed to solution bias. The correct sequence is hypothesis first, then possible play, then service fit—not the other way around.
Illustrative thinking sample
“Primary play: exploratory modernization conversation focused on coordination bottlenecks. Disqualifier: if the account already has a clearly owned transformation program, deprioritize or reframe.”05Open logicSection generatorStakeholder Lens & Validation Path
Suggests likely roles and discovery questions without turning titles into buying ownership claims.
Role lensDiscovery questionsRole ≠ intent
05
Open logic
Key question
Who may care, what might they care about and what should be asked to validate the hypothesis safely?
Reasoning logic
Make the report useful for human outreach planning by translating account logic into role-oriented validation paths.
Grounding basis
Stakeholder evidence, account structure, role patterns and hypothesis-specific discovery questions.
Reliability control
Titles are not treated as proof of ownership. The section suggests probable relevance and safe questions, not declared stakeholder facts unless evidence exists.
Illustrative thinking sample
“A strategy or transformation role may care about coordination pain, but the report should still ask who owns the issue rather than assuming that ownership.”5. Structured-output hygiene & fallback
Malformed output, broken JSON or missing fields are treated as recoverable technical failures—not permission to publish weak reasoning.
Schema generation→Parse / required fields→Repair / salvage→Section fallback if needed
↺Open logicHygiene loopStructured-output repair
Repairs malformed or incomplete section output while keeping the underlying reasoning scope bounded.
JSON repairField completenessBounded fix
↺
Open logic
Key question
Can the section be made structurally valid without changing what it is allowed to say?
Reasoning logic
Stabilize the output format first, then return the section to downstream gate review and contract enforcement.
Grounding basis
Existing section schema, parse checks, repair routines and deterministic fallback logic.
Reliability control
Technical validity does not equal trustworthy content. Parse quality and evidence quality remain separate concerns.
Illustrative thinking sample
“If the section is malformed, repair its structure—but do not treat the repair as evidence that the reasoning itself was strong.”FOpen logicFail-safe logicDeterministic fallback
If generation is weak or broken, the workflow can produce a safer bounded version rather than a fluent but unsafe one.
Fail-safeCautious defaultVisible limits
F
Open logic
Key question
What is the safest report behavior when generative quality is insufficient?
Reasoning logic
Prefer a cautious, more deterministic section over a persuasive but unreliable output. Failure should reduce confidence, not be masked by language polish.
Grounding basis
Section requirements, canonical signals, Registry content and deterministic fallback rules.
Reliability control
The fallback preserves the contract and keeps quality notes visible. It is designed to degrade gracefully, not fake completeness.
Illustrative thinking sample
“If the section cannot safely generate a nuanced sales play, produce a simpler cautious play with explicit uncertainty rather than an overconfident recommendation.”6. Independent semantic Gate Review
After prose is generated, claims are decomposed into review packets and assigned a customer-use boundary.
Build claim packets→Gate decisions→Repair / downgrade→Remove unresolved blocked wording
GOpen logicGate reviewSemantic claim gate
Assigns each claim a customer-use status: fact, hypothesis, question, internal-only or blocked.
Fact / hypothesis splitClaim packetsDowngrade path
G
Open logic
Key question
How should each generated claim be treated before the report reaches human or customer-facing use?
Reasoning logic
Turn commercial prose into reviewable units with explicit customer-use boundaries rather than trusting paragraph-level tone.
Grounding basis
Claim packets linked to hypothesis IDs, parent evidence, section paths and output text.
Reliability control
Unsafe claims can be rephrased as hypotheses or discovery questions—or blocked entirely if they exceed the evidence boundary.
Illustrative thinking sample
“‘Likely buying soon’ may be blocked. ‘May soon face stronger pressure and is worth discovery-led outreach’ may be allowed as a bounded hypothesis.”↯Open logicCritical safety logicRepair, downgrade or remove
Flagged wording is not merely noted; it is actively repaired, downgraded or stripped from public sections.
Active enforcementNot passive reviewAudit trail
↯
Open logic
Key question
What should happen when a claim is commercially appealing but evidentially unsafe?
Reasoning logic
Operate only on the flagged claim set and force a concrete action: rephrase, downgrade, questionify, internalize or remove.
Grounding basis
Gate decisions, repair instructions, repeated field actions and claim-level metadata.
Reliability control
Any unresolved blocked wording is removed from reader-facing sections. The workflow does not rely on readers to infer that the claim was weak.
Illustrative thinking sample
“If the report says ‘the CFO is the buyer,’ but stakeholder evidence is weak, convert this to ‘a finance leadership role may be relevant to validate’ or remove it.”7. Deterministic sales post-processing
The generated sections are normalized back to the hypothesis contract, service-fit rules and final sales-safe wording requirements.
SVOpen logicAlignment stepService-fit validation
Checks whether proposed plays and solution directions genuinely match the controlled service knowledge.
Fit checkReadiness bandSolution-bias control
SV
Open logic
Key question
Do the recommended plays align with available capabilities, and are they appropriate for the account hypothesis?
Reasoning logic
Translate hypothesis-led plays into realistic commercial options rather than generic recommendation boilerplate.
Grounding basis
Service Portfolio KB, modernization tactics, fit scoring, readiness logic and account-specific constraints.
Reliability control
This step validates fit; it does not validate need. That separation is essential to avoid solution-led hallucination.
Illustrative thinking sample
“A play may fit the provider’s capability set, but if account evidence is weak, the play must still remain exploratory rather than assertive.”1POpen logicNormalization stepOne primary play & guardrails
Enforces one primary play, consistent readiness logic and reusable caution fields across the report.
One primary playConsistencyGuardrail buckets
1P
Open logic
Key question
How do we make the report operationally clear without losing nuance and caution?
Reasoning logic
Normalize the report so a reader gets one primary commercial direction, explicit disqualifiers and consistent labels rather than a vague list of possibilities.
Grounding basis
Registry values, gate decisions, service fit results and deterministic field-completion rules.
Reliability control
Blocked or weak claims propagate into “use only as question”, “do not say” or weak-evidence buckets. Caution travels with the artifact.
Illustrative thinking sample
“Primary play selected: discovery-first modernization conversation. Guardrails: do not imply funded initiative, do not assign stakeholder ownership as fact, keep urgency conditional.”8. Final QA, reliability status & validation boundary
The final artifact organizes what to test, why it may matter, who may care and which conversation path is safest. It becomes actionable only when human discovery validates or rejects the hypothesis.
A/B/COpen logicPublication controlReliability status & final QA
Converts evidence sufficiency and gate results into a visible use status, from strong discovery basis to do-not-use.
Use statusNot confidence theaterOperational clarity
A/B/C
Open logic
Key question
How usable is this report for discovery, and how much caution should the reader apply?
Reasoning logic
Express report strength in operational terms rather than abstract model confidence so a seller knows whether to use it as a strong basis, a cautious basis or a question-led draft.
Grounding basis
Evidence sufficiency, gate-review coverage, unresolved issues, report completeness and public-language safety.
Reliability control
The workflow is allowed to produce a cautious, question-led or do-not-use outcome. That is a major trust feature, not a weakness.
Illustrative thinking sample
“The report may be commercially useful, but if stakeholder evidence is weak and urgency is mostly contextual, treat it as a cautious discovery basis rather than a strong activation basis.”?Open logicHuman boundaryValidation path & what not to say
The human seller validates owner, pain, urgency, consequence and fit—or rejects the hypothesis. Unsafe claims remain blocked.
Discovery questionsHuman judgmentCustomer-safe boundary
?
Open logic
Key question
What should a human now test in conversation before treating the hypothesis as commercially meaningful?
Reasoning logic
Turn the report into a disciplined discovery agenda rather than a substitute for validation. Make negative boundaries visible so misuse becomes less likely.
Grounding basis
Role lens, disqualifiers, validation questions, blocked-claim categories and customer-use rules.
Reliability control
The workflow ends in a human validation boundary by design. It explicitly blocks unsupported certainty, ungrounded stakeholder ownership and claims of active buying intent.
Illustrative thinking sample
“Allowed: ‘This may be a useful discovery angle.’ Blocked: ‘This stakeholder is the buyer’ or ‘the company is already seeking this solution’ unless supported by evidence.”Account Snapshot & Thesis
Concise account frame and primary need hypothesis.
Why Now / Why Soon
Timing logic built from pressure signals and account context.
Pain-to-Solution Map & Plays
Conversation directions, primary play and disqualifiers.
Guardrails & Validation Path
Question-only claims, role lens, reliability status and next questions.
Core boundary: “Need hypothesis and demand-creation potential, not confirmed buying intent.”