Just Launched: Senzing Is Now a Kiro Power

Helping Builders Build

Think of the Senzing power as a forward-deployed engineer in a box: top technical talent from Senzing with decades of entity resolution expertise, available to your Kiro agents the moment a task calls for it.

Kiro agents now help builders deploy and activate identity intelligence in the infrastructure — identity graphs ready to service all kinds of workflows/pipelines: Customer 360 that prevents onboarding duplicates; fraud-detection that catches synthetic identities before approval; KYC that continuously screens applicants and customers against sanctions lists in milliseconds; patient matching that unifies records across healthcare systems; or investigations that trace beneficial ownership through corporate structures with third party data. Anywhere that identity has to be accurate, the Senzing power gives Kiro agents the expertise to build it.

What Is Entity Resolution?

Entity resolution (ER) is the process of determining when different data records refer to the same real-world entity – a person, an organization, a location – and when they don’t. Records at most organizations are typically fragmented across systems, inconsistent, and full of duplicates. ER fixes that, answering two deceptively simple questions: who is who, and who is related to whom?

Every downstream analytic, customer count, risk score, fraud model, or AI agent decision depends on getting those answers right. ER has to do two things at once: catch matches that look different but are the same (“Sue Jones” and “Susan Jones” who recently moved) and tell apart records that look the same but are different (a father and son sharing a name, address, but have one letter off ; “Jr” versus “Sr”).

The discipline of entity resolution goes by many names; record linkage, fuzzy matching, data matching, deduplication, identity resolution. All of these terms relate to the same problem.

What Is Agentic Entity Resolution?

Historically, it’s been incredibly time-consuming to combine data. Adding a data source to a system can take experts many months to deliver. Such time and effort are unacceptable in this agentic era.

Agentic entity resolution (Agentic ER) is when an AI agent can autonomously prepare data, load and resolve entities in real time or batch, and let users conversationally explore the results without experts to configure, train, or fine-tune the system for each new data source.

Agentic ER reshapes how identity intelligence infrastructure will serve the enterprise:

• Agentic Data Preparation : agents profile, map, and validate any source
• Agentic DevOps : agents design, code, and deploy pipelines
• User Exploration: analysts conversationally interrogate resolved data, and
• Agent Conversations: autonomous workflows share identity context across multi-agent systems.

Senzing Was Built For This Moment

The Senzing power for Kiro has two components: 1) The Senzing SDK (software development kit), the actual entity resolution engine, running inside your perimeter (air gapped, if required), invoked by the DevOps code the Kiro agent generates; 2) The Senzing MCP Server, a knowledge service that teaches Kiro agents to work with Senzing, deploying the engine, coding data-source mappers, writing ingestion pipelines, querying the identity graph, troubleshooting, never touching transactional data.

This split is privacy by design in action, something Senzing has been delivering since day one.

When it comes to Agentic ER, the architectural principle is simple: the Kiro agent should build what runs your entity resolution, not run it. The MCP Server teaches the agent how to construct the pipeline; the SDK runs in the pipeline once it’s built. Write the integration once, run it a million times without the LLM/MCP being involved. Better tokenomics. Lower latency. Explainable.

Senzing was built for industrial-grade agentic workflows from the ground up. It’s principle based entity resolution, not rules-based; preconfigured for people and organizations, so it deploys without pre-training, tuning, or experts, no matter how many data sources you add.

It uses an Entity Centric Learning™ approach, comparing new records against everything known about each resolved entity rather than record-to-record matching, delivering real-time accuracy that gets sharper as evidence accumulates, and catching matches that rules-based systems miss. Sequence neutrality ensures the same data resolves to the same entities regardless of arrival order, a critical property for any production system where data streams in from multiple sources simultaneously.

And it runs at scale: production deployments in the billions of records. No LLM/MCP in the transactional path, no black box in your compliance story.

What The Senzing Power Gives Kiro Agents

The power covers the full entity resolution lifecycle:

• Data mapping – profile any source (Parquet, Kafka, JSON, CSV, databases) and map it into Senzing record format, without manual schema work
• Deployment – stand up Senzing as working code in your team’s language of choice (Python, Java, C#, Rust, or TypeScript/Node.js), designed to integrate with the AWS services you’re already using; Amazon S3, AWS Fargate, Amazon SQS, AWS Lambda, Amazon Aurora PostgreSQL, Amazon Neptune, and others
• Ingestion and resolution – load records in batch or real time, resolve them into an identity graph continuously maintained as new data arrives
• Querying – ask natural-language questions against the resolved graph and get explainable answers
• Explainability – Senzing explains why, why not, and how for every record-matching decision. Every match and every non-match comes with a full evidence chain.

Kiro Powers, In Ten Seconds

A power bundles a Model Context Protocol (MCP) server, steering files, and hooks into a single install. Unlike a raw MCP server, a power activates only when a developer’s task calls for it so your context window stays clean and your agent focuses on the right tools at the right time.

One click from the IDE, and your agents know how to work with that technology. That’s what we’re delivering for entity resolution.

How It Activates

The Senzing power activates automatically when a Kiro conversation touches entity resolution work. Trigger keywords include:

• Core ER vocabulary – entity resolution, identity resolution, record linkage, fuzzy matching, deduplication, identity graph, identity matching
• Use case categories – MDM, master data management, customer 360, headless360, fraud, KYC, AML, sanctions screening, watchlist screening, continuous vetting, and more
• Vertical specifics – patient matching, master patient index (MPI), vendor risk, third-party risk, and more

Mention any of these in a Kiro prompt, and the power’s tools and guidance load into context. When the conversation shifts elsewhere, the power unloads, and your context window clears.

Install It In A Few Seconds

From the Kiro IDE:

• Open the powers panel
• Search for Senzing
• Click Install

From the web: Click “Add to Kiro” on the Senzing power page at kiro.dev/launch/powers/senzing.
Source and documentation: github.com/Senzing/senzing-kiro-powers.

On first activation the power will walk through any setup prompts your project needs, API credentials and target deployment configuration.

Try It

Activate the Senzing power for Kiro, grounding every conversation in Senzing domain expertise, not the more general knowledge of the LLM alone.

If you’re building anything that depends on knowing who’s who, install the power and let us know what you ship.

For a deeper read on the architectural pattern behind this, Jeff Jonas’s recent paper “What Is Identity Intelligence? How It Works & Why It Matters” lays out the principles behind agentic entity resolution and the identity intelligence layer that supports it.

SENZING® and ENTITY CENTRIC LEARNING® are registered trademarks of Senzing, Inc. Kiro and Kiro Power are trademarks of Amazon.com, Inc. or its affiliates. All other trademarks are the property of their respective owners.