01 / INTRODUCTIONFive perspectives. One shared horizon.
PULSE Horizon Intelligence is a prototype horizon-scanning platform designed to help explore emerging developments relevant to Uganda through structured public-source monitoring and evidence-led review.
Its interactive Intelligence Floor presents five specialist perspectives and a coordinating Floor Manager. The animated room makes the research workflow easier to follow; it is an interface to the underlying collection and reporting system, not a simulation of independent human researchers.
The purpose is to support exploration, questioning and strategic foresight, not to provide definitive predictions or official policy conclusions.
02 / SPECIALIST PERSPECTIVESThe five Ps
Each specialist works with assigned sources and thematic rules. Together, the five perspectives help organise evidence across connected areas of development.
PeopleSocial conditions, wellbeing and human development
ProsperityEconomy, livelihoods and opportunity
PlanetEnvironment, climate and natural systems
PeaceInstitutions, safety and social stability
PartnershipsCooperation, institutions and development relationships
The Floor Manager coordinates the output into a consolidated briefing for review. A cross-sector connection is a research hypothesis, not proof that one development caused another.
03 / WORKFLOWHow the Intelligence Floor works
- Launch a mission. An authorised user selects a time window, scanning options and participating perspectives.
- Assign and check sources. The platform works through configured public-source assignments and records whether each source could be accessed.
- Collect candidate evidence. Available feed, API and page records are parsed, filtered by scope and stored with source references where available.
- Analyse patterns. Deterministic thematic and lexical rules organise evidence into candidate signals, possible connections and research questions.
- Review the results. Users can open specialist workstations, inspect evidence and signals, and examine source coverage and gaps.
- Prepare a briefing. The Floor Manager presents a consolidated report, which requires human scrutiny before any publication or decision.
Agents visually travel to their desks, show source activity on their monitors and return to the lounge after their tasks. Those animations represent system states and do not imply human-like independent reasoning.
04 / DATA & TRACEABILITYWhere information comes from
PULSE uses configured public websites, feeds, structured datasets and optional approved imports. Example source families in this prototype include World Bank indicators, ReliefWeb, GDELT and other configured public information sources. Availability and coverage vary by source and mission.
Reddit and Facebook public-page gateways may be listed in the source configuration, but collection depends on official API credentials, access permissions and implementation status. A configured source is not evidence that it has been successfully scanned.
The Source Monitor and individual evidence records are the right places to inspect connection outcomes, timestamps, exclusions and available references. Collection respects access restrictions and robots policies where applicable.
05 / OUTPUTSWhat a report contains
Depending on available evidence, a briefing may include an executive summary, five-perspective findings, candidate signals, cross-sector hypotheses, source coverage, blind spots, contradictions and suggested follow-up checks.
Reports distinguish draft material from human-approved material. Signal cards and automated interpretations are research aids, not verified findings merely because they appear in a report.
06 / READ CAREFULLYLimitations and uncertainties
Important: PULSE is a bounded research prototype. Its outputs should never be treated as comprehensive surveillance, validated forecasts, confirmed facts or official United Nations positions.
1. Incomplete and uneven coverage
Only configured, reachable and permitted sources can contribute. Feeds have depth limits, records may be missing or delayed, and coverage is not representative of every Ugandan district, language, community or sector.
2. Source and access failures
Sites may change formats, restrict access, require credentials, impose rate limits or return no matching data. An empty result can mean no collected evidence, not that nothing important happened.
3. Time and historical limitations
Publication dates may differ from event dates. Bounded feed windows are not complete historical archives. Comparisons require suitable baselines; this prototype does not assume a validated historical baseline is available.
4. Rules-based interpretation
The current synthesis uses deterministic lexicon and theme matching, not a configured large language model. Keyword matches can miss context, irony, local terminology and nuanced developments; they can also create false positives.
5. No validated causality or forecasting
Co-occurrence, changes in record counts and cross-sector links do not demonstrate growth, diffusion, probability, causation or real-world impact. Hypotheses require independent investigation.
6. Evidence quality and independence
Public information can be inaccurate, duplicated, biased, outdated or based on the same original announcement. Source references do not guarantee independent corroboration or implementation of announced plans.
7. Social listening and personal data
The prototype is not a population-representative social-listening system and does not perform individual profiling. Public social platform access, if configured, remains subject to official permissions and technical limitations.
8. Operational and security limits
Data freshness, availability and scan completion depend on local infrastructure, network conditions, source health and authorised configuration. This prototype should be reviewed and hardened before production or high-stakes deployment.
07 / HUMAN OVERSIGHTHow to use PULSE responsibly
- Inspect the original source and publication date before repeating a claim.
- Check source health, missing perspectives and excluded assignments.
- Seek independent corroboration, especially for consequential findings.
- Separate observed evidence from interpretations, hypotheses and recommendations.
- Use qualified subject-matter review and approval before external publication or policy decisions.
Human judgement remains essential. PULSE helps organise questions and evidence; it does not replace research expertise, consultation with affected communities or institutional decision-making.