AI and agents
Running large language models and AI agents on governed data: self-hosted and managed models, Cohere integration, retrieval, coding assistants, prompt-injection defence and why agents fail in production.
- November 1, 2026, by Ashis Parajuli
A semantic layer for AI agents: governed metrics over REST and MCP
A semantic layer defines each metric once, so dashboards, applications and AI agents compute it the same way. In DataNature, cubes are YAML files over Gold Iceberg tables that compile to warehouse SQL, can be materialized as warehouse views, carry row-level security per role, and are served over a REST endpoint and over MCP, which Voir agents use to answer numeric questions without guessing joins.
- November 1, 2026, by Praful Nepal
Private RAG on governed lakehouse data: how DocuContext, Milvus, cubes and Voir fit together
Private retrieval-augmented generation keeps every step inside your boundary: documents are parsed and embedded by DocuContext into a Milvus vector store, structured questions are answered from governed cubes over Gold tables, and Voir agents combine both under scoped access. The model is called through an AI gateway pointed at your approved endpoint, and nothing is sent to an external embedding or LLM API.
- November 1, 2026, by AI Team
MLOps and LLMOps on a sovereign platform: experiment tracking, model registry, tracing, evaluation and prompt versioning
DataNature AI covers the full model lifecycle inside your jurisdiction. Experiments log parameters, metrics and artifacts; a model registry versions every model with aliases such as champion and challenger and links it to its training data snapshot; tracing records every step of an LLM or agent call; evaluation scores answers with automated judges and human feedback; and prompts are versioned like code.
- November 1, 2026, by Sudip Shrestha
Knowledge graph and ontology for agentic AI: how DataNature gives multi-agent systems semantic context
An ontology defines the entities your business runs on (customers, accounts, devices, policies, suppliers) and how they relate. DataNature builds a governed knowledge graph from that ontology over Gold tables and documents, so agents can follow relationships instead of guessing joins. Voir agents use it for multi-step work such as deep research, investigations and relationship briefings, with every hop governed and cited.
- November 1, 2026, by David Lowry
Hyper-personalization with real-time signals: from next-best offer to next-best conversation
Hyper-personalization means choosing the reason, timing, channel, language and content of each interaction for one customer, from their profile and live signals. On DataNature, AI360 event SDKs and StreamHouse supply website, app and transaction signals within seconds, AI360 Studio scores intent and picks the next-best action, and the action is delivered as a voice call, message, app content or a staff briefing, with consent checked by DataNature Atlas.
- November 1, 2026, by Ashis Parajuli
How to build predictive models on DataNature: from lakehouse data to production predictions
Yes, DataNature includes everything needed to build prediction models. Data lands in Iceberg tables, dbt builds clean Silver and Gold data, the feature store holds reusable features, Spark trains models with experiment tracking, the model registry versions and approves them, and model serving delivers batch or real-time predictions. Predictions flow into AI360 Studio for decisions and into Voir agents as skills, and DataNature Atlas records lineage, access and every prediction served, all inside your jurisdiction.
- November 1, 2026, by James Smith
Build a churn prediction model on DataNature, step by step
To build a churn model on DataNature, define churn and the prediction window, build a point-in-time training table from Gold customer and behaviour tables, register the features in the feature store, train and compare models with experiment tracking, register and approve the best one, score customers weekly with a scheduled job, and send the highest-risk customers to retention teams through AI360 Studio, reverse ETL and Voir. Real-time signals from StreamHouse can trigger save actions within minutes.
- November 1, 2026, by Sudip Shrestha
Hyper-personalized AI voice calls from your customer data platform: how AI360 connects to voice agents
AI360 turns the customer profile into personalized outbound voice calls. It decides who to call, why, when and in which language, checks consent, and sends each customer's approved details as variables to a voice agent platform such as ElevenLabs Agents or the calling platform you already use. Call outcomes, answers and opt-outs flow back into the profile, so every next call is better informed.
- November 1, 2026, by Sudip Shrestha
AI in insurance claims: from first notice of loss to settlement with document AI, fraud scoring and claims agents
AI can speed up every step of an insurance claim. On DataNature, claims arrive through StreamHouse and Connect, DocuContext turns forms, invoices, medical reports and photos into structured data, AI360 Studio scores complexity, fraud and leakage, the knowledge graph links related claims, and Voir agents give claims handlers summaries, next steps and draft letters with citations. Straightforward claims are fast-tracked, complex ones get expert attention, and every decision is traceable in DataNature Atlas.
- November 1, 2026, by AI Team
An AI gateway for the enterprise: controlling LLM cost, routing, quotas and guardrails in one place
An AI gateway is the single entry point between your applications and agents and every language model you use. DataNature's AI gateway routes each request to an approved model endpoint, applies budgets, quotas and rate limits per team or agent, caches repeated answers, enforces PII and safety guardrails, keeps provider credentials out of applications, and logs tokens, cost and latency for every call, all inside your jurisdiction.
- November 1, 2026, by Ashis Parajuli
Self-hosted, Cohere or Dlytica-managed: three ways to run LLMs on sovereign data
There are three practical ways to run large language models on sensitive data: self-host open-weight models on your own GPUs, use Cohere models through its API, Model Vault or private deployment, or let Dlytica run models for you inside your jurisdiction. Choose by data sensitivity, volume, in-house skills and how much capability you need.
- November 1, 2026, by Sudip Shrestha
How to use Cohere with DataNature: sovereign RAG and agents on your own data
DataNature connects to Cohere’s Command, Embed and Rerank models through Cohere’s API, its OpenAI-compatible Compatibility API, a dedicated Model Vault, or a private deployment inside your own VPC or data centre. Lakehouse tables and documents stay governed in DataNature; Cohere models generate, embed and rerank; Atlas logs every prompt and answer.
- October 31, 2026, by AI Team
Self-hosted LLM vs API: the real cost and when self-hosting pays off
Self-hosting an LLM pays off when volume is high and steady, GPUs stay busy, and data cannot go to a third party. Below that, APIs or a managed service are usually cheaper once you count engineering time, monitoring, upgrades and idle capacity. Research on small and mid-sized firms found self-hosting can reach cost parity within months at moderate, sustained volume.
- October 30, 2026, by AI Team
Why most AI agents never reach production, and how to build ones that do
Most enterprise AI agents stall in pilot because of systems problems, not model problems: hallucinated actions, prompt injection, runaway loops, missing permissions, poor data and no monitoring. Agents reach production when guardrails live in credentials and APIs rather than prompts, every action is logged and reversible, and agents work on governed data.
- October 30, 2026, by Praful Nepal
Private AI coding assistants for banks and regulated teams: self-hosted models that keep code in-house
Regulated teams can use AI coding assistants without sending source code to a third party by serving open-weight coding models, or Cohere’s North Mini Code, inside their own network. Self-hostable coders now score roughly 60–72% on SWE-bench Verified versus 80–95% for frontier cloud models, so pick per task and keep a human review step.
- October 29, 2026, by David Lowry
Agentic AI use cases that work today: 20 examples for banks, credit unions, telcos, insurers and government
Agentic AI works best today on well-defined, data-heavy workflows with a human approving the final step: drafting credit memos, triaging fraud and AML alerts, reconciling payments, answering members with cited policies, diagnosing network faults and processing claims. The power of agents comes from combining reasoning with governed data and narrowly scoped tools.
- October 29, 2026, by Sudip Shrestha
Prompt injection: how to protect enterprise AI agents that read documents, email and data
Prompt injection is the top risk in the OWASP Top 10 for LLM applications: attacker text hidden in a document, email, web page or record tells an agent to do something else. It cannot be solved with prompt wording alone. Defend in layers: treat all retrieved content as untrusted, give agents least privilege, require approval for risky actions, and log everything.
- October 23, 2026, by Ashis Parajuli
What is a multi-agent system, and when does the enterprise need one?
A multi-agent system coordinates several specialized AI agents, each with its own scope, tools and data access, to complete multi-step work such as onboarding, claims or reporting. Enterprises need one when a process crosses systems and approvals. On DataNature, Voir runs these agents with governance enforced in code.
- October 14, 2026, by Sudip Shrestha
AI agents for telecom network operations
AI agents help telecom network teams answer questions about KPIs, investigate incidents and draft tickets from governed network data. Built on DataNature with StreamHouse and Voir, they run in-country, use certified KPIs from the semantic layer, and only act within the scope each engineer is allowed.
See it on your own data.
Book a demoDemos and quotes are handled by Dlytica, the company behind DataNature.