Product
Your team chats as usual. dotosi does the rest.
Before a message reaches any AI, dotosi covers what is sensitive, applies your rules and keeps a record. Your employee sees exactly what will be covered, and sends it with one click.
3 details covered before sending
How it works
Follow one message through dotosi
Your employee
Draft a contract for Ahmed Al-Falasi, Emirates ID 784-1990-1234567-8
Here is the contract for Ahmed Al-Falasi (ID 784-1990-1234567-8), ready to sign.
On the way out, dotosi
- Masks the name and Emirates ID
- Applies your policies and skills
- Checks usage limits
- Picks the right model
- Logs the message
Logged, nothing exposed
On the way back, dotosi
- Restores the real values, for this employee
- Records what it cost
- Logs the answer for management
The AI model
Draft a contract for PERSON_1, Emirates ID NATIONAL_ID_1
Here is the contract for PERSON_1 (ID NATIONAL_ID_1), ready to sign.
Watch one message go through dotosi
About 45 seconds
What you get
Safe AI for everyone, under your control
What each part does for you, and the specifics behind it.
Data masking
Keep sensitive data away from the AI
Sensitive details become placeholders before a message leaves, so the AI never sees them. Your employee still sees the real values.
- Names, ID numbers, bank and card details and more, in Arabic and English
- Files too: PDF, Word, Excel and CSV
- Add your own words to always cover, or always allow
- Employees check what was covered before anything is sent
Logs and monitoring
See how AI is used, without reading private data
A clear record of who used which AI and when, and an alert if anyone sends a sensitive value anyway.
- A masked record of every conversation
- Every time someone views a real value, it is logged
- Admins are alerted if a value is sent uncovered
- Six ready-made reports you can export
Budgets and limits
No surprise AI bills
See what AI costs as it happens, get warned before a budget runs out, and set hard limits where you need them.
- Monthly budgets for the company or each department
- Warnings at 80% and 100% of a budget
- Hard limits per employee, and the cost of every message
- Choose which AI each team or person may use
Every AI model
The right AI for every task, in one place
Every major AI in one familiar chat, so staff stop juggling separate accounts and subscriptions.
- OpenAI, Anthropic, Google and 13 more providers
- Add a new AI model in seconds
- Switch AI mid-conversation without starting over
- UAE-hosted models, or a private one you run yourself
People and roles
Every team gets the right access
Set things up once per department, and decide exactly who can see and change what.
- Employees grouped into departments
- Each department gets its own AI, budgets and rules
- Custom roles with exactly the permissions you choose
- AI accounts and protection settings stay with admins
Policies and skills
AI that follows your company's rules
Shape every answer to how your company works, from tone of voice to what is off limits.
- A ready-made library of skills to start from
- Write your rules in plain words
- Apply them to one department or the whole company
- Employees switch skills on right in the chat
Smart routing
Don't pay premium prices for simple questions
Use your premium AI for the work that needs it, and a lower-cost one for the rest.
- Everyday questions go to a lower-cost AI
- Complex work goes to your best AI
- Repeated content costs less on AI models that support it
- A report shows what routing saved, and what it cost
Decisions
Clear a backlog of tickets, invoices or leads
Sort a whole list in one run instead of item by item, with a person checking anything the AI is unsure about.
- One question across up to 200 items, with 2 to 10 possible answers
- Every item gets an answer, and how sure the AI is
- Unsure items go to a person to check
- The same protection and records as chat
Files: PDF, Word, Excel, CSV and plain text. Images and scans are not accepted, because dotosi covers text and a picture would reach the AI uncovered. Very long documents are shortened, and both you and the AI are told where.
Reports
Answers for management, not dashboards to decode
Six ready-made reports, by department and by employee. Export any time.
Usage
Who uses AI, which models, and how much?
Spending
What does AI cost us, who drives it, and are we on budget?
Protection
What sensitive data was caught, and how often was it sent anyway?
Smart routing
Where did messages go, and what did routing save or cost?
Decisions
How much sorting work did AI take off people?
Weekly digest
How did this week compare with last week, in plain words?
Where it runs
Run it where your data needs to be
Let us run it, run it in your own cloud account, or keep it in your own building. Same product, same protection.
dotosi Cloud
We run it for you. Best for getting started fast.
- Up and running in minutes, nothing to install
- We handle updates and maintenance
- Pick AI providers by region, including UAE-hosted models
- Encrypted, and never used to train AI
Your cloud
In your own cloud account. Best when IT already runs on a cloud.
- Runs on one server in your account, on any major cloud
- Pick a region in your country for everything dotosi stores
- Your encryption keys are made there and never leave
- Nothing reports back to us
On-premise
In your own building. Best when nothing may leave your walls.
- Runs on one server in your own data centre
- With your own AI model, it runs with no internet at all
- Your encryption keys are made there and never leave
- Nothing reports back to us
In your cloud or on-premise, this is what stays inside
Your servers
dotosi app
Chat, admin and reports
Detection engine
Finds sensitive data, locally
Database
Real values stay locked
Encryption keys are created here, and never leave.
The AI provider
On your own account. The only thing that goes out.
or your own AI model
Run one yourself, and nothing leaves at all.
Your engineers get the full setup guide on request: requirements, install, upgrades and backups.
Questions, answered
The things buyers ask us first
Do our employees need training?
No. They chat the way they already do with any AI tool. Before a message is sent they see what dotosi covered, and the answer comes back with the real names and numbers in place.
Do we have to install anything?
No. dotosi runs in any modern browser. If your teams prefer an app, there are desktop apps for Mac, Windows and Linux, with notifications and one-click access to a new chat or decision.
Which AI models can we use?
Models from 16 providers, including OpenAI, Anthropic and Google, plus UAE-hosted models and any private model you run yourself. You choose which ones each team may use.
Does dotosi know what is sensitive in our industry?
Every field has its own sensitive data and rules, whether finance, healthcare, legal, government, retail or yours. We adapt what dotosi looks for, your policies and your list of terms to what your field cares about, and you can add your own words to always cover at any time.
If dotosi covers our data, doesn't dotosi see it?
Only for the moment it takes to cover it. The original text is held in memory while the message is handled and is never saved. What is kept is exactly what went to the AI, which is masked text unless your employee chose to send a value uncovered (and that is logged), plus the real values, locked with strong encryption. Nothing is used to train AI. If your policy rules out even that, run dotosi in your own cloud or on-premise, so the original text never leaves your network.
What if something sensitive is missed?
dotosi is set to cover too much rather than too little. It combines fixed rules for things like ID and card numbers with an AI that understands context, in Arabic and English. You can add your own words to always cover, employees check what was caught before sending, and if protection is ever down, nothing is sent at all.
Does covering data make the AI's answers worse?
No. Each detail becomes a consistent placeholder, so the AI keeps the full context: "draft a contract for [PERSON_1] due Thursday" reads and reasons exactly like the real sentence. The real values are put back into the answer for your employee automatically.
Can our data stay in our country?
Yes, when dotosi runs in your own cloud or on-premise. Choose a cloud region in your country or your own data centre, and pair it with an in-country AI provider such as UAE-hosted models. Then the whole flow stays in your country. Tell us your requirements and we will set dotosi up to meet them.
How quickly can we start?
dotosi Cloud is ready in minutes, with nothing to install. In your own cloud or on-premise, dotosi needs just one server, and we send your engineers the full setup guide.
How do we get our data out, or delete it?
Your company's data can be exported, and deleted for good, on request. Once the locked values are erased, the real details cannot be recovered, while the masked record stays in place for your own compliance.
Let your whole company use AI, safely.
See dotosi on your own kind of data in a 20-minute demo. No commitment.
Request a demo and our sales team will contact you to schedule it.