# AI Development and Automation in New York | Nadya Construction Studio

> AI assistants, agents and automation built on Claude: grounded in your data, permission aware, measured with evals, with cost controls from day one.

Canonical: https://nadiaconstructionstudio.com/ai

[Home](https://nadiaconstructionstudio.com/) / AI and automation

AI and automation

# AI that does *real work.*

Assistants, agents and automations that answer from your real data, take real actions, and prove they work with evals before they reach your team.

[Start a project](#start)[See our work](https://nadiaconstructionstudio.com/#work)

## What can AI do for my business?

Take over the repetitive reading, writing and lookup work that slows your team down: answering questions from your data, drafting replies, processing documents, routing tickets, writing reports, reviewing code, and taking actions in your systems with a person in the loop where it matters.

## Assistants, agents or automation?

Whichever is simplest for the job. A fixed workflow with one model call is often more reliable and cheaper than a free roaming agent. We use agents when the task genuinely needs planning and tools, and we tell you which one you are getting and why.

## How do you keep AI from making things up?

We treat the model as untrusted and check its work in plain code. Answers are grounded in sources it actually retrieved, numbers are recomputed, links are verified, and anything that changes data needs a dry run and a person's approval.

## Can it see data a user should not see?

No. We design AI to act as the person asking, with that person's permissions, so it can never show someone more than they could already see.

## How do you know it works?

Evals. Before launch we build a set of real tasks with known good answers and run the system against it on every change, measuring accuracy, latency and cost. Failures become tests, so they stay fixed.

## What does it cost to run?

Model spend is designed in, not discovered later. Every call is metered, prompt caching cuts repeated context, cheaper models handle the steps where evals show they are good enough, and a budget limit can switch it off before a bill surprises you.

## Does it need memory?

Assistants that talk to the same people over time usually do. We built MenteDB, an open source memory engine for AI, so an assistant can remember what it learned, notice contradictions and recall the right context later.

## Do you fine tune or distill models?

When it pays off. For a narrow, high volume task, a smaller model fine tuned or distilled from a larger one can match quality at a fraction of the cost and latency. We build the training set from your real data, compare against the large model with evals, and only switch when the numbers hold.

## Can AI search our documents and data?

Yes. Retrieval over your documents, tickets, wikis and databases with vector and keyword search combined, permissions respected per user, and answers that cite the exact source. We measure retrieval quality separately from answer quality, because most wrong answers start with the wrong documents.

## What about voice, images and scanned documents?

We build with them too: reading invoices, forms and scanned PDFs, describing and checking images, transcribing calls and meetings, and voice assistants. Each one gets the same evals and cost controls as text.

## Which models and tools?

Mostly Claude, on the Anthropic API or Amazon Bedrock, plus other models where they win your evals. Tool use over MCP, retrieval with vector and keyword search, LangGraph where orchestration helps, and infrastructure on AWS.

Start a project

## Tell us what you want *built.*

A few lines is enough. The person who reads it is the person who builds it.

- A reply within one business day
- A written scope with price and date before any work
- Code, infrastructure and accounts in your name

Prefer email? [nam@nadiaconstructionstudio.com](mailto:nam@nadiaconstructionstudio.com)