ContextOps
How to build the context layer that gets enterprise AI to production.
Most enterprise AI teams have cycled through the same fixes — better prompts, bigger context windows, newer models — and the agents keep failing the same way.
If none of that has moved the needle, that's your cue to look further upstream — at the knowledge the agent is actually running on. Who assembled it? Is it current? Does anyone own it?
The operators who have taken enterprise AI to real production, the kind with users behind it, are the ones who asked that question early. ContextOps is a podcast about how they did it.
A discipline the enterprise has been missing.
A fortnightly conversation between Vivek Khandelwal and Joshua Thomas, joined by practitioners and researchers doing real work on this problem. We're thinking through what it actually takes to make enterprise AI reliable at the knowledge level — the layer that sits upstream of model benchmarks and prompt tips.
Latest episode: A different kind of answer, not a better number→Episodes
Conversations with the operators and researchers building the context layer behind production enterprise AI. New episodes drop fortnightly through Season 1 — subscribe to get each as it lands.
A different kind of answer, not a better number
Karthik Soman invented KG-RAG — knowledge-graph-based retrieval augmented generation — while building biomedical knowledge graphs at UCSF, and now leads enterprise-scale agentic AI at SAP America. One question carries from a PhD in computational neuroscience to the enterprise: how do you build intelligent systems that actually work in the real world? His answer isn't a better accuracy number but a different kind of answer — one a human can trace, question, and learn from. He walks through the case that convinced him: enriching patient records with a 40-million-node biomedical graph surfaced an olfactory-receptor gene that flagged Parkinson's five years early, catching a prodromal case a clinician had missed — not because the model was more accurate, but because it pointed at a mechanism. Then he moves to the enterprise, where the curated ontologies of biomedicine don't exist. You lean on the topology already inside your documents. Graph-based reasoning turns out to be a sixty-year-old idea that LLMs merely made usable on the fly. And a graph earns its keep over vector RAG in specific places — multi-hop questions, smaller models, private data the model never saw — before the least glamorous advice in AI: data hygiene first, then AI hygiene.
A graph is one tool, not the destination
Melli Annamalai has spent 27 years at Oracle watching technology waves crest and break — multimedia retrieval, the semantic web, big data, property-graph analytics, and now knowledge graphs for AI. As the Distinguished Product Manager who leads graph technologies there, she makes an argument you rarely hear from a database vendor: a knowledge graph is one tool in the kit, not the destination. She traces why semantic-web tech stalled for two decades: a steep learning curve, a custom RDF/OWL/SPARQL ecosystem, and a year-and-a-half payback that senior management wouldn't fund. Then what AI finally changed, and where these projects still quietly fail — over-engineering everything into a graph, tuning and tooling gaps, and the security officer who shuts it all down. Along the way, a working definition of "ontology" for non-technical buyers, natural language as the new query language, and Oracle's converged-database bet to collapse the graph, vector, and agent layers into the place the data already lives.
Discover your ontological core
Tony Seale spent two decades stitching together siloed data inside investment banks — Lehman, Deutsche Bank, UBS — until he concluded the hard part was never the model, it was the data. Now widely followed as "the knowledge graph guy," he makes a precise argument: a large language model is rented intelligence you don't own, while the ontology and knowledge graph you build are the durable IP that stays inside your organization. He walks through the neuro-symbolic loop — why LLMs are continuous and knowledge graphs discrete, and why you cycle between them rather than pick one — coins Seale's Law on how marketing hollows out a word's meaning, and warns that companies that don't consolidate their "ontological core" will watch their advantage quietly leak out. Along the way: semantic data products and the DPROD standard, why true open standards (schema.org, W3C) beat vendor "open," and why the agentic web arriving next makes getting your data estate in order the only move that counts.
Own your meaning, or rent it
A CTO has a 2-million-token context window and thousands of recorded cancer-care calls. The plan: pipe them in, ask for provider performance, done. Elliott Risch — who runs R&D on semantic AI at Enterprise Knowledge and came to it through mathematical logic, not engineering — walks through where that holds and where it quietly stops flagging the contradictions that matter, until someone has to stand in court and say the computer made a mistake. He separates inductive systems (LLMs that guess the next token) from deductive ones (rules that must hold every time), and makes the case for owning the meaning of your company instead of renting it.
The Vocabulary Problem — what 'context' actually means
Alan Morrison has spent two decades watching technical vocabularies get coined, hyped by vendors, and hollowed out — knowledge graphs, grounding, and now "context." Drawing on 20 years running PwC's Technology Forecast think tank, he argues the vocabulary problem in enterprise AI isn't a communication failure but an architectural one: when words stop meaning anything, architecture follows marketing and teams hold the repair bill for years. He makes the case for data-centric over application-centric, why incumbents stay stuck, and the smallest credible first step toward a knowledge-graph layer.
Vivek Khandelwal
IIT Bombay grad. Previously built iZooto from scratch into a market-leading marketing automation platform. Spent two years in enterprise AI sales watching teams blame the model for failures that were actually upstream — in how their knowledge was assembled, governed, and served. Co-founded CogniSwitch to build the infrastructure layer that was missing.
Joshua Thomas
Leads the neuro-symbolic architecture behind the platform's reliability and auditability. Over a decade in AI and computational linguistics; earlier, Product Lead at Aikon, where he learned firsthand how AI systems behave in production. Authored CogniSwitch's CQI benchmark and the Trust-but-Verify HotpotQA evaluation — work grounded in the conviction that verification has to be deterministic and auditable, not a probabilistic second opinion from another model.
No one has written this playbook yet.
We're working through it one episode at a time. If you're figuring it out too, come build it with us.