Three things a standard communication app doesn't give you: an assistant that does the editor's work for you, developing the app yourself within an agreed scope and a private assistant for employees running on your server - without sending any data outside.
See the assistant at work
This is not a chatbot that tells you where to click. The assistant performs the editor's work itself: it creates content, publishes it, directs it to the right people and decides on requests - in the admin panel and from a phone, in the same conversation.
A screen from the AllAbout app: the assistant renames the draft to the chosen suggestion and confirms that the draft is not published. The company and the content are demonstration data.
Tell it what the post is about - the assistant creates a draft, writes the title and body itself, suggests a category and offers three titles to pick with one tap. You add a photo, narrow the audience, schedule the publishing time. Publishing itself and the push notification you confirm with a fingerprint or PIN on your phone.
A quiz with questions and their correct answers, a time limit, random question order, shown only to the group you choose, and then the results and the ranking. Questions about your own procedure it writes from the source text you give it - it does not invent facts about your company, it asks for the source.
It creates an onboarding checklist template, adds tasks with deadlines counted from the hire date, names the supervisor and the buddy, and then starts the onboarding for one person - the checklist shows up on their phone. Later it tells you who is at which stage and what is overdue.
It tells you how many requests are waiting for a decision and which days they cover, and on your instruction approves or rejects a particular one. The employee gets a notification. It does the same with plant pass requests.
Screens from the AllAbout app. The company and the content are demonstration data.
Publishing, sending a notification, deleting content - every such action first shows a card with what will happen and how many people it reaches. You confirm it with a fingerprint or a PIN.
Procedures, rules, names, dates and numbers it does not make up. When the topic needs them, it asks for the source text instead of writing something plausible.
After saving it reads the result back from the database and builds the receipt from that reading. If something else was created than intended, the action is rolled back and you get an error, not a “done”.
Permissions are checked exactly as in the panel, on the account that is talking. The assistant will not show or do anything your account could not do by itself.
The assistant also works outside the app: as part of the rollout we connect it to your website, your ATS recruitment system and the other tools you use. It then changes a photo on the website, prepares a job offer in your ATS and updates content - in the same conversation in which it runs the app. The scope of the connected systems is agreed during the rollout.
The reader in the phone covers two completely different things: entering the app and confirming an action that cannot be undone. Below is exactly how each of them works.
An employee can turn on unlocking by fingerprint (on an iPhone also by face) in the app settings. The session token then sits in the phone's system store - Keystore on Android, Keychain on iPhone - with a "require biometrics" condition. The system enforces it, not the app: without a successful fingerprint check the store simply does not release the value. What is stored is a token, not a password - a token can be revoked and expires on its own.
When the conversation reaches an instruction that cannot be undone - publishing a post to everyone, deleting content, scheduling a publication, publishing a sports challenge or changing its date - the assistant shows a card: exactly what will happen and how many people it will reach. The phone then opens the system reader window by itself. A phone without a reader asks for the PIN; the PIN also stays as the fallback when the reader is on. If you back out, nothing at all goes to the server.
The fingerprint confirms that the action is requested by the owner of the phone. It is an extra layer alongside the permissions on the account - exactly the same ones the publish button in the admin panel requires - and alongside the assistant's activity log, which records every change together with the confirmation method.
No waiting weeks for a developer or queuing for changes. You tell the assistant in plain language what should change - and within the scope agreed with you it does it right away: it adds a field to a form, a new category, a tile in the “More” tab, sets up a quiz or a survey, corrects content and the order of sections. All under your brand and your control.
Small changes you'll make yourself, in minutes. For bigger ones our team is on hand to take over. You decide how much you do on your own.
A screen from the AllAbout app. The company and the content are demonstration data.
The employee asks in natural language — “how much leave do I have?”, “what to do if a machine breaks?”, “where do I return equipment?” — and gets an answer right away. No searching through binders, no waiting for HR. And it all stays fully private.
A large language model (LLM) can be deployed on the company's infrastructure. Data and documents do not go to OpenAI or any other external company.
The assistant answers solely on the basis of documents uploaded by the administrator — regulations, procedures, safety, benefits. It doesn't use general internet knowledge and doesn't make things up.
The administrator decides what the assistant knows and can say. Change a document — the answer changes. No surprises.
Employee data stays in the company. Privacy and security from day one.
You write a message once — in Polish — and the AI instantly translates it into each recipient's language: Ukrainian, English, German and more. No one is left out just because they don't understand the announcement.
And in international companies: one deployment for many sites — an administrator manages their country's content, while a super-administrator manages everything.
The AI in AllAbout not only builds features and answers employees — it also works “behind the scenes”, guarding the security of the server dedicated to your company.
An AI agent continuously monitors the server, detects intrusion attempts and unusual events, and keeps us informed of threats in real time.
Automatic updates run every day — tools and libraries don't stay exposed to known vulnerabilities for long.
After an update the AI independently walks through the admin panel, assembles a test build of the mobile app and checks that everything works correctly.
Each day's result comes to us as a report — so we react before any problem affects users.
This is real, daily use of AI to increase the security of the tool — not just a marketing slogan.
Book a free demo - we'll show the assistant and how you develop the app yourself, using your company as an example.
Book a Demo
AllAbout