Describe it. Generate it. Own it.
Describe how a part of your business works. Nx·Studio's AI business analyst designs digital workers for it — AI agents, business rules and human approval steps, laid out as a flowchart you can read. Test them on realistic data, then download the complete source code and run it yourself.
Your description
“Invoices from new vendors or above $50,000 need the controller's approval. Flag duplicates. Everything else goes through.”
Illustration of a worker's journey through the studio.
What is a digital worker?
Software that handles one piece of work from start to finish — like checking supplier invoices. Inside, it's a flowchart of steps: AI agents that read and decide, fixed business rules, calls to your systems, and points where a person approves.
Because the rules and approvals are real steps in the flowchart — not instructions hidden in a prompt — anyone can read what the worker will do before it runs.
Invoice exception handler
example- 01 AI step Read the invoice and pull out the vendor, amount and purchase-order number
- 02 Rule Over $50,000, or a vendor we've never paid? Send it for approval
- 03 Data Check past invoices for a duplicate
- 04 Approval The controller approves or rejects — the run waits until they do
- 05 Outcome Approved, rejected, or sent back with a reason
From a description to code you own, in six steps.
Describe your business
Answer a ten-step guided form, paste a description into the Fast Lane box, or start from one of eight starter packs. The AI pulls out your rules, roles, approval limits and goals — and scores how complete the picture is.
Get up to five designs
In a few minutes, an AI business analyst proposes up to five designs, from cautious to ambitious. It checks each one and recommends the best fit.
Review the flowchart
Every design is a flowchart you can read: AI steps, rules, approvals and system calls. Edit it on the canvas, or describe a rule in plain English and let the AI write it — with live checks as you go.
Test it on realistic data
Create realistic sample data for the worker, or load your own spreadsheets. Run it in the studio and see every step, every approval and the cost of each AI step.
Compare and improve
Save test cases, compare versions side by side, try alternatives, and keep every version so you can always go back.
Download the code
Download all of the worker's code, ready to run on your own systems with your own AI account. It comes with a test, a security risk assessment and a signed list of what's inside.
Watch Nx·Studio design a worker.
One description, through the studio's own tabs: what it understood, the five designs it weighed, the flowchart it chose, a test run that waits for an approval — and the code you take home. Click any tab to look closer.
You describe the work in your own words. The studio picks out the rules, roles and data — and scores how complete the picture is.
Fast Lane — describe the work
“Invoices from new vendors or above $50,000 need the controller's approval. Flag duplicates. Everything else goes through.”
What Nx·Studio understood
Illustration based on Nx·Studio's own tabs, stages and labels. The worker, its data and the scores are examples.
The Nexgile difference.
AI models can be swapped out. What's built around them is what counts. Four principles are built into every Nexgile digital worker — each one something you can see, test or run yourself.
Proven on realistic data before go-live.
Every worker gets its own realistic sample data — or you can load your own spreadsheets. Run it in the studio and see every step, every approval and the cost of each AI step. Save test cases, compare versions side by side, and try alternatives before anything touches a live system.
- Sample data for every worker
- A step-by-step record of every run
- Versions compared before you ship
Your policy isn't a suggestion to an AI.
Write a rule in plain English, such as “refunds over $10,000 need the finance director's approval”. It becomes a real step in the worker — not a line hidden in an AI prompt. Rule steps don't use AI at all, so the same facts always get the same answer. AI handles only the steps that need judgment, and the worker stops wherever a person must decide.
- Rules that don't use AI
- Approvals built in
- Risky steps without approval get flagged
Your auditors read it before it runs.
Every digital worker is a flowchart you can read — not a tangle of AI instructions. It shows the AI steps, the rules, the connections to other systems and where people approve. A business view shows who does what. Every change is saved as a new version you can compare, and a health check runs before anything is downloaded.
- A readable flowchart
- Every version kept
- Health check before download
Take the source code home.
When a worker is ready, you download all of its code — including the engine it runs on. Your team runs it on your own systems with your own AI account. There's no license key, nothing to rent and no connection back to us. Every download also includes a security risk assessment and a signed list of every AI model, software library and dataset inside.
- All the code
- Runs on your own systems
- A list of what's inside
Approval steps are built into the worker. When it reaches one, it stops and waits for a person to decide. And if a worker could change or delete data without an approval first, the studio warns you before you download it.
Recently added.
The biggest additions to Nx·Studio — each marked with where it stands today.
Test data made for each worker
Each worker gets its own realistic sample data — or load spreadsheets and data files of up to 200,000 rows. The worker uses it while you test.
Workers that take action
39 kinds of steps: calling your other systems, looking things up in a database, reading web pages and documents, creating images, and using connected tools.
Teams of agents
A lead agent that hands work to specialist agents, agents working side by side, steps that repeat for every item, and retries with a limit.
Approvals that pause and resume
In the studio, a worker stops at an approval step and waits — even through a restart. It carries on as soon as a person approves or rejects, from the run page or the Actions inbox.
Knowledge for each worker
Give each worker its own documents to search, plus facts and memories to draw on — while you test, and in the code you download.
Rules from your own words
Write rules in plain English, like “refunds over $10,000 need the finance director's approval”. They become real steps and approval chains — not lines hidden in an AI prompt.
Self-checking designs
Every design is checked against 73 rules, fixed automatically where possible, and scored before you see it. An optional second pass adds detail to every step.
Views for the business
A view of who does what, a readiness score, a health check before download, a map of the data, and a cost estimate for each action.
Walled-off workspaces
Each customer's data is kept separate inside the database, with roles, folders and usage limits for every workspace.
A list of what's inside
Every download lists the AI models, software libraries and data inside it, signed to show where they came from — plus a security risk assessment.
More model choices
Claude by default, or ten other AI providers such as OpenAI, Google Gemini and Amazon Bedrock, or models on your own servers. You use your own account.
Connect to your other systems
Pick from a catalog that describes thousands of business systems, add your own, or paste a system's documentation and let the AI set up the connection.
Built in, and off until you switch them on for a workspace or a download:
39 kinds of steps to build with.
Every worker is assembled from these steps — the AI picks them when it designs a worker, and you can add, remove or rewire them on the canvas. Each one runs the same way in the studio and in the code you download.
Start and finish
2Where work comes in and where results go out.
Reasoning
5AI steps that read, decide and write — alone, side by side, or as a team with a lead.
Custom tools
2Small pieces of custom code, and other tools the AI can use.
Connections
6Connections to your other systems, websites and workers — or a hand-off to your existing automation bots.
Documents
2Pull details out of documents, then check them.
Knowledge
6Search your documents, look up facts, remember things — or ask a database, read-only.
Quality
2Check a result against rules or a scoring guide before moving on.
Governance
3A person's approval, fixed business rules, and a permission check before a tool runs.
Compliance
4Rule checks, hiding personal data, and safety guardrails.
Paths and repeats
6Choose a path, filter, combine, repeat for every item, or retry within a limit.
Reuse
1Use another worker as a single step.
A few steps need an optional add-on in the downloaded code — the export screen shows which. Custom steps can arrive as outlines for your developers to finish.
Watch it work before it touches anything real.
The studio runs your worker on the same engine you'll download — so what you test is what you get. Every run is recorded step by step, and your other systems can be replaced with simulated responses.
Sample data for every worker
Made to fit the worker's own data, or imported from your files.
Every step, in view
A live timeline of each step: what went in, what came out, who approved, and what each AI step cost.
Real AI, safe systems
Test runs use real AI, while your other systems can be replaced with simulated responses.
Test cases from your rules
Sample inputs for each expected outcome (approved, rejected or needs a person), created from your rules.
Compare versions
Run two or three versions on the same input, or let a statistical test pick the better one.
Replay any run
Step back through a finished run and see what each step saw and did.
What you take home.
Nx·Studio is where workers are designed and tested — it doesn't run your day-to-day work. When a worker is ready, you download it and run it yourself.
The whole program
Ready-to-run code, plus the full source of the engine it runs on.
The worker itself
Its flowchart, AI instructions, rules, approval steps, data and knowledge files.
Ready to run
Setup files to start it on your own servers — plus cloud setup files if you want them.
Proof of what's inside
A security risk assessment, and a signed list of every AI model, software library and dataset inside.
Independent of Nexgile
The downloaded code has no license key, no tracking and no connection back to Nexgile, and it needs no Nexgile service to run. Your team runs it on your own systems, with your own AI account and database. An “Export contents” screen shows exactly what's in the box before you download.
No meter on the work
Rule steps don't use AI, and AI is used only where a step needs judgment — at your own provider's rates, with a cost estimate for each action before you go live. There's no Nexgile meter on the decisions your workers make.
Built for everyone around the work.
From the people who know the process to the people who run the code.
Business leader
COO, VP of Operations
Find the processes worth automating and see them designed
Business analyst
Analyst, process owner
Describe how the work really happens, rules and all
Technical architect
Solutions architect, CTO
Review the design, the connections and the approvals
AI engineer
ML engineer, developer
Refine AI instructions, rules and tests until it's right
Platform team
DevOps, IT operations
Set up and run the downloaded code
Risk & compliance
Risk, audit, security
Read who approves what before anything goes live
Questions people ask first.
Do I need to be technical to use it?
Not to get started. Business users can describe the work, review the designs and run tests in plain language. Your engineers then finish the connections to your systems and put the downloaded code to work — some steps arrive as outlines for them to complete.
Can I test with our real data?
You can. Each worker gets generated sample data by default, and you can import your own spreadsheets when you're ready. What you upload is kept separate for your organization.
What exactly do we get when we download a worker?
All of the worker's code and the engine it runs on, ready to run — with a test, a security risk assessment and a signed list of what's inside. There's no license key and no connection back to us.
Which AI models does it use?
Claude by default, or ten other AI providers such as OpenAI, Google Gemini and Amazon Bedrock, or models on your own servers. Your downloaded workers use your own AI account.
Can a generated worker go live right away?
Not straight away. A generated worker is a checked first draft. Review it, test it on your own cases, and finish its connections to your systems before it goes live. The studio is built to make that loop fast.
From a business idea to software you own.
Bring one process — a paragraph is enough. We'll design a worker for it, test it with you, and show you the code.
You describe it. You test it. You own it. You run it.
Nx·Studio designs the digital workers that run your business. Nx IDE is there when you want to build or extend the software around them. The Cognitive Engine learns routine decisions and makes them without AI.