# gohel-knowledge-context-procedural-graph-2026-10-25

## Veille

LinkedIn post by **Rakesh Gohel** (builder at **JUTEQ**, governance and proof infrastructure for agents) published around **October 25, 2026**, ~250 words, accompanied by an **infographic** credited to *@rakeshgohel01*. He distinguishes three graphs an agent needs to decide, using the example *"Rebook this passenger"*.

## Titre Article

Knowledge Graph. Context Graph. Procedural Graph. — What's true → What matters now → What do I do next

## Date

2026-10-25

## URL

https://www.linkedin.com/posts/rakeshgohel01_aiagents-agenticai-generativeai-activity-7520100714279239681-QgHj

## Keywords

knowledge graph, knowledge graph, context graph, context graph, procedural graph, procedural graph, ontology, Semantic Spine, AI agents, agent memory, decision-scoped subgraph, context scoping, procedure as a graph, procedure-relation-procedure triples, decision traces, audit trail, human escalation, learning from traces, context engineering, agent governance

## Authors

Rakesh Gohel — builder chez JUTEQ (gouvernance et infrastructure de preuve pour agents IA) ; LinkedIn, newsletter rakeshgohel.substack.com.

## Ton

Educational post in LinkedIn format: an opening question, three arrows unfolding a single example (rebooking an airline passenger), a one-sentence takeaway, then an engagement question to the reader. The vocabulary is assertive and cites no source; this is a conceptual framework, not a quantified field report. The infographic, dense and polished, carries part of the argument absent from the text (the ontology, the functions column, the *Trace & Learn* storage) and ends with a call to follow the author and his newsletter. The register is that of a practitioner selling an agent-governance infrastructure, as indicated by his displayed title.

## Pense-betes

- **Three questions per decision**: true (knowledge graph), relevant now (context graph), next (procedural graph).
- **Knowledge graph**: stable facts, *"enterprise-wide memory"*; nodes in the example: Passenger, Aircraft, Fare, Flight, Route, Airport, Weather, Policy.
- **Context graph**: *"Don't retrieve everything. Scope the graph to what this decision needs."* Live state (Flight 34/185, Gate C8, 4 seats left, 9 min to close), organizational context, decision history.
- **Procedural graph**: Check Availability → (if required) Check Visa; → Compare Alternatives; *No Options* → Escalate to human; *Options Found* → Apply Fare Rules → Resolved. *"Guides the next step. Doesn't dictate it."*
- **Agent**: reasons over the scoped graph, follows the next edge, deviates *"when the graph is wrong"*.
- **Trace & Learn loop** (infographic): Action → storage of success and failure traces → feedback to the agent; traces update knowledge, context, and procedure.
- **Semantic Spine** (infographic, absent from the text): Ontology (*what things are*) → Knowledge Graph (*what's true*) → Context Graph (*what matters now*) → Procedural Graph (*what to do next*) → AI Agent (*acting, and writing it back*).
- **Author's lesson**: scope the graph per step, store the procedure *"as a graph instead of a prompt"*, log every decision as an audit trail.
- Related to [[kb-context-engineering]].

## RésuméDe400mots

Rakesh Gohel, who is building a governance infrastructure for AI agents at JUTEQ, argues that every decision an agent makes rests on three questions: what is true, what matters now, what to do next. A knowledge graph answers the first well; each of the other two, in his view, deserves its own graph.

He takes the example of a simple request made to an airline agent: rebook a passenger. The knowledge graph holds what is true at the company-wide scale: passengers, aircraft, flights, fares, routes, airports, weather, policies, kept stable and governed. The context graph extracts from it a small slice limited to the decision at hand: Gate C8, Flight 34/185, four seats left, nine minutes before closing, and how similar cases have been handled. The infographic sums up this principle: don't retrieve everything, scope the graph to what the decision needs.

The procedural graph stores procedures as procedure-relation-procedure triples. The agent moves from checking availability to comparing alternatives then to resolution, checks the visa if needed, applies fare rules when options exist, and escalates to a human when there are none. The graph guides the next step without dictating it: the agent reasons over the scoped context, follows the next edge, and remains free to deviate from it if the graph is wrong.

Every trace gets written back. Precedents refine the context, failed runs correct the procedure, and recurring patterns become governed facts. The infographic embodies this loop through a "Trace & Learn" store that collects success and failure traces and updates knowledge, context, and procedure.

It adds an element absent from the text, the "Semantic Spine," a five-link chain: the ontology says what things are, the knowledge graph what's true, the context graph what matters now, the procedural graph what to do next, and the agent acts and writes the result back.

The author's lesson comes down to three instructions: scope the graph at each step, store the procedure as a graph rather than as a prompt, and log every decision, because that log becomes the audit trail. The post closes by asking the reader which of the three graphs their agent is missing.

## GrapheDeConnaissance

- Rakesh Gohel —travaille_chez→ JUTEQ (ORGANISATION, 0.9)
- Rakesh Gohel —a_créé→ Semantic Spine (CONCEPT, 0.85)
- Rakesh Gohel —affirme_que→ chaque décision d'agent dépend de trois questions : ce qui est vrai, ce qui compte maintenant, ce qu'il faut faire ensuite (AFFIRMATION, 0.95)
- Semantic Spine —utilise→ ontologie (CONCEPT, 0.9)
- Semantic Spine —utilise→ knowledge graph (TECHNOLOGIE, 0.92)
- Semantic Spine —utilise→ context graph (CONCEPT, 0.92)
- Semantic Spine —utilise→ procedural graph (CONCEPT, 0.92)
- context graph —est_basé_sur→ knowledge graph (TECHNOLOGIE, 0.9)
- context graph —est_basé_sur→ traces de décision (CONCEPT, 0.88)
- procedural graph —remplace→ procédure en prompt (CONCEPT, 0.85)
- procedural graph —permet→ escalade humaine (CONCEPT, 0.85)
- agents IA —utilise→ procedural graph (CONCEPT, 0.88)
- traces de décision —améliore→ procedural graph (CONCEPT, 0.88)
- traces de décision —permet→ piste d'audit (CONCEPT, 0.9)
- Rakesh Gohel —recommande→ scoper le graphe à chaque étape plutôt que tout récupérer (AFFIRMATION, 0.94)
- Rakesh Gohel —recommande→ stocker la procédure comme un graphe plutôt que comme un prompt et journaliser chaque décision (AFFIRMATION, 0.94)
- Rakesh Gohel —affirme_que→ le graphe procédural guide l'étape suivante sans la dicter ; l'agent s'en écarte quand le graphe est faux (AFFIRMATION, 0.9)

---
Canonical: https://www.thekb.eu/en/fiches/gohel-knowledge-context-procedural-graph-2026-10-25/