LEAF Advisory · Intelligence Unlocked
Ask the question. Trace the answer to the evidence.
Decisions become stronger when the evidence behind them remains visible. Intelligence Unlocked works with your teams to investigate a specific question using authorized public, licensed, specialist and client-provided sources.
AI helps search, structure, translate and connect the material. Analysts decide what matters, verify key findings and turn the evidence into an assessment that states its sources, confidence and gaps.
What it is
Sourced evidence, turned into a decision-ready assessment.
Intelligence Unlocked is the decision intelligence capability of LEAF Advisory. It is not an OSINT feed or a monitoring tool: it is research around a question that matters to a decision.
Every engagement begins with a specific question and an agreed purpose. Sources, entities, time horizon, permitted source categories and expected outputs are defined before collection starts.
Start with the question. Keep the evidence attached.
Start with the question
Collection follows scope, not curiosity.
The question decides what is collected, from where and for how long. Illustrative examples:
- Should we work with this partner?
- What is changing around this organization?
- What evidence supports this market claim?
- What technologies is this competitor investing in?
- Where does this narrative originate?
- What can we substantiate about this exposure?
Illustrative questions, not fixed engagement scopes.
Agreed before collection
- Purpose
- Entities
- Time window
- Source categories
- Legal basis
- Retention
- Outputs
Nothing is collected before these are agreed with your team.
Four source categories
Not everything is open-source intelligence.
Each finding is labelled with its source category, because reliability, terms of use and how the content may be shared differ between them.
- Public
Public OSINT
Publicly accessible websites, news, public registers, corporate filings, publications, patents, standards and public social content where relevant to the agreed question.
Visible online does not mean fair game: source terms, purpose and legal basis still apply.
- Licensed
Licensed sources
Commercial databases and other information legitimately licensed for the engagement and used within their contractual terms.
Never presented as public or open sources.
- Specialist
Specialist threat-intelligence feeds
Feeds from specialist providers, used when the question concerns security exposure, threats or incidents.
Only when provider access and use are authorized for the engagement.
- Client
Client-provided sources
Documents and datasets your organization is authorized to provide for the agreed purpose.
Access, retention and scope are defined per engagement.
Acquisition and use depend on authorization, licence conditions, source terms and the applicable legal basis. Access to closed databases or breach data is not promised.
How the evidence is built
Every material finding carries its own provenance.
Findings are not a pile of search results. Each material finding is recorded with the elements needed to judge it.
A material finding carries
- Claim
- What is being asserted.
- Source
- Where it comes from.
- Source category
- Public, licensed, specialist or client.
- Date
- When it was published or collected.
- Confidence
- A stated, qualitative level, never a percentage.
- Corroboration
- Whether independent sources support it.
- Known gaps
- What could not be established.
Confidence is an analyst judgement recorded with its reasons, not a universal score applied to every raw source.
AI output is a starting point for analysts, never a finding in itself.
A model can surface, translate and connect material. It cannot decide what is true.
Conflicting evidence is a finding too.
When credible sources disagree, the disagreement is preserved and reported instead of being forced into a single narrative.
Absence of information is not evidence of absence.
Not finding something in the sources investigated is reported as a gap, never as proof that it does not exist.
How we work
Scope. Gather. Analyze. Verify.
01/ Scope
Frame the question together
Agree the purpose, entities, decision, legal basis, permitted source categories and outputs with the team that owns the question.
02/ Gather
Collect what is relevant
Retrieve information only from the agreed source categories, recording its source and when it was obtained.
03/ Analyze
Structure the material
Use AI-assisted methods to normalize, translate, identify entities, cluster, compare and summarize, so analysts can work across more material.
04/ Verify
Test the important findings
Corroborate key claims, compare independent sources, assess reliability, identify conflicts and state uncertainty, then discuss the implications with your team.
AI assists with
- search
- translation
- entity extraction
- clustering
- classification
- summarization
- relationship discovery
Analysts remain responsible for
- relevance
- verification
- source assessment
- interpretation
- conclusions
Provenance
A finding you can inspect, a map you can trace.
Two common ways findings are shared: a finding card with its sources, confidence and gaps, and an entity map where every link points back to evidence.
Finding
Organization A has announced a second production site for Technology X.
Sources
- Source 01 · company filing
- Source 02 · patent record
- Source 03 · trade-press interview
Organization A
- supplierOrganization B2 sources
- executivePerson in a public role1 source
- patentTechnology X3 sources
Questions we help answer
Where decision intelligence is used.
Competitive & market context
How competitors and new entrants position themselves, what they announce, hire for and patent, and what that suggests about their direction.
Due diligence support
Research support for evaluating organizations such as partners, suppliers or investment targets, proportionate to the decision. Not formal financial or legal due diligence.
Technology & research monitoring
Following publications, patents, standards and early products in a defined area, with defined entities, cadence and sources.
Reputation & narrative context
How an organization, product or topic is discussed in public sources, and where that discussion originates.
Security exposure context
Supporting your existing security team by combining authorized threat-intelligence feeds with public information to understand an exposure.
Strategic research
Structured, sourced research on a specific topic for leadership, strategy or innovation teams that need evidence rather than opinion.
Illustrative patterns, not descriptions of specific client engagements.
From evidence to decision
Evidence is useful when it changes the decision.
The work does not stop at collection. The assessment connects each important finding to what it means, the options it opens or closes, the questions that remain open and the next step your team can take.
- Finding
- Implication
- Option
- Open question
- Next step
Boundaries
Proportionate, organization-focused, defensible.
Intelligence work is only useful if it can be defended. These boundaries apply to every engagement.
Personal data
Intelligence work involving personal data must remain necessary and proportionate to the agreed purpose and supported by an applicable legal basis, assessed with you. The GDPR is the framework we work within, not a certification.
Organizations first
The work focuses on organizations. Individuals appear only as people acting in relevant public or professional roles, where the engagement supports it. No private-person profiling or intrusive personal surveillance.
No breach data
Access to breach data is not promised. We do not buy leaked credentials or access stolen databases.
Security context, not a SOC
Security exposure work supports your existing security team. It is not a SOC, SIEM, EDR or penetration-testing service.
Scoped monitoring
Periodic monitoring is an output only where scoped, with a defined topic, entities, cadence and sources. It is not a real-time monitoring platform.
Working environment
The working environment is agreed according to the material, requirements and engagement. Sensitive analysis can be carried out within architectures designed around your security, residency and infrastructure requirements.
Outputs
Agreed per engagement.
Outputs are agreed per engagement and reviewed with your team as the work progresses.
Intelligence brief
A focused answer to the question, with key findings and confidence.
Sourced research dossier
The underlying material, organized and referenced.
Due diligence summary
Relevant findings on an organization, with open points to clarify.
Entity & relationship map
Organizations, people in public roles and their documented connections.
Monitoring summary
Periodic updates on a defined topic or set of entities.
Limits
What intelligence can and cannot tell you.
Intelligence reduces uncertainty. It does not eliminate it.
- Public information may be incomplete.
- Licensed information may be outdated.
- Sources can be wrong or intentionally misleading.
- Absence of evidence is not evidence of absence.
- Correlation does not establish causation.
- AI-generated interpretations require verification.
- Confidence is not certainty.
- Source availability can change.
- Findings apply to the scope and time window investigated.
Start with the question
What do you need evidence for?
Tell us the decision you are facing, the entities or topic involved and what you currently know. We can define the question, the permissible source categories and the evidence required to investigate it.