What Aleria is, in sixty seconds

Aleria is a place where you bring your work — documents, data, questions, half-formed ideas — and the system helps you finish it. You chat with it the way you'd chat with a colleague who happens to know your files, your databases, and how to build things.

It's one interface with many capabilities underneath. You don't pick a mode. You describe what you want. Aleria figures out whether that means answering a question, writing a report, running a query, building an app, or kicking off work that takes an hour. You stay in the conversation. It handles the rest.

The 10-minute tour

Follow this once, end to end, and you'll understand how every feature in this guide connects. You don't need to remember the names — we'll cover them later. Just do the steps.

1. 1

   Open your workspace

   When you sign in, you land in a workspace. Think of a workspace as a shared folder for one team or one project. Everything you do — chats, files, data — lives inside it. If you're on a team, you likely share one. If you're solo, you have your own.
2. 2

   Say hi in chat

   The big text box on your screen is the way in. Type "summarize the most recent file I uploaded" or "what's on my calendar this week" — whatever is on your mind. The reply comes back in seconds.
3. 3

   Drop a file in

   Drag any document, spreadsheet, image, or PDF into the chat. Ask a question about it. Aleria reads the whole thing, not just the first page.
4. 4

   Ask for a report or a dashboard

   Try: "build me a one-page dashboard of our pipeline, grouped by owner, with a filter by quarter." Watch what happens — the chat starts working through steps, shows its progress, and hands you back a live page you can share as a link.
5. 5

   Ask it to build an actual app

   Now go bigger: "build me a small internal tool where our ops team can log incidents, tag them by severity, and see the weekly trend." Aleria builds the app — forms, data, charts — and deploys it to a temporary Aleria domain like yourteam.aleria.app/incidents. Copy the URL, send it to someone, open it on your phone. No setup, no hosting account, no deploy step.
6. 6

   Let it run in the background

   When a task is long (reviewing 200 contracts, ingesting a database, monitoring a feed), the chat promotes itself into a background agent. You can close the tab. It'll notify you when it's done.

Core concepts

A handful of ideas show up everywhere in Aleria, and they build on each other. The best way to understand them is as one story:

You bring your sources. You give your team skills and package them into reusable agent templates. You chat with those agents or drop them as steps inside a Flow. You watch the results on a Board. All of it lives in a Workspace.

Workspace

Your team's home. Holds sources, files, chats, skills, templates, Flows, Boards, and shared memory. People you invite have access to everything inside. People you don't, don't. Most companies run a workspace per department or per project.

Sources

Where your workspace's data lives. Three shapes:

• A database you spin up inside Aleria — create a new database in seconds and push data into it straight from chat. No external setup, no DBA ticket.
• An existing system you connect — your own databases, SaaS tools, cloud storage, event streams
• Files you upload — indexed automatically and available workspace-wide

Anything Aleria generates — a cleaned table, a report, the output of a Flow — can be saved back as a source so the next agent picks up where the last one left off. Sources are shared across the workspace, so once a teammate connects or creates something, it's available to everyone. When a source updates, every Board, agent, and Flow built on it sees the update too.

Chat

The main surface. You talk, Aleria responds. It can read your sources, query data, generate files, build apps, and run long tasks in the background. Most of what you do every day happens here.

Skills

Specific capabilities — reading a database, calling your pricing engine, rendering a chart, generating a deck. Some are built in; the best ones are the ones you describe in plain language and Aleria generates for your team. You don't invoke skills by name — the chat picks the right ones for the task.

Agent templates

Pre-configured starting points built from skills and sources. Instead of opening a blank chat, you open "Legal Reviewer" or "Ops Analyst" — an agent that already knows which sources to read, which skills to use, and how to behave. Like skills, you can describe a role in plain language and Aleria drafts the template for you.

Agents

A running instance of a template (or a chat that's been promoted to run on its own). An agent does work on your behalf — reviewing, drafting, monitoring — and keeps a visible log of every step.

Flows

When a task is repeatable, you chain agents together as steps in a Flow. A Flow is a deterministic pipeline with versioning, approvals, and a full audit log. The agents inside it are the same ones you'd chat with.

Boards

Live dashboards wired directly to your sources and files. Numbers update as the underlying data updates. You can drill in with a click or ask questions in plain English — and get a new panel back.

Where to go next

Pick the page that matches what you're trying to do. If you're not sure, start with Chat — it's the shortest path to getting something real done.

What a workspace actually is

Think of a workspace the way you think of a shared drive. It has a name, members, files, and content that grows over time. The difference: a workspace also has memory. The chats inside it, the sources connected to it, and the knowledge teammates have built up all become context the AI can draw on.

Most organizations run one workspace per department (Legal, Finance, HR) or per project (Acme acquisition, Q3 audit). You can belong to several at once and switch between them freely.

Creating a workspace

Go to the workspace switcher in the top-left and choose New workspace. Give it a name, pick an icon, and you're in. You can change everything later.

A new workspace starts empty: no files, no sources, no memory. Your first job is to bring in the material the team will actually work with.

Inviting people

Open Members in the workspace settings. Add teammates by email. They get an invite and show up in the workspace the moment they accept.

There are three roles:

• Owner — full control, including deletion and billing. Usually one or two people per workspace.
• Member — can do anything inside the workspace, but can't delete it or change billing.
• Viewer — read-only. Useful for executives or auditors who need visibility but shouldn't modify.

Files and data rooms

Drag files into the workspace to make them available to everyone in it. Aleria reads them once and remembers. You can attach a file to a specific chat for a focused conversation, or leave it in the workspace library to be picked up automatically when relevant.

Supported out of the box:

• Documents — PDF, Word, TXT, Markdown, RTF
• Spreadsheets — Excel, CSV, TSV, Google Sheets exports
• Slides — PowerPoint, Keynote exports
• Images — PNG, JPG, HEIC, SVG
• Audio & video — MP3, WAV, MP4, MOV (transcribed on upload)
• Archives — ZIP (extracted automatically)
• Code & logs — any text-based source file

Sources

Beyond one-off file uploads, the workspace holds sources — the live places your team's data lives or gets stored. There are three kinds, and most workspaces end up with a mix of all three:

Databases you create in Aleria

Open Sources → New database, give it a name, and you have a ready-to-use database in seconds. Push data into it straight from chat ("import this CSV", "add a row for the Acme deal"), and query it from any agent, Flow, or Board. No external setup, no hosting account, no DBA ticket.

Existing systems you connect

Bring in the data you already have — your company databases, warehouses, SaaS tools, cloud storage, event streams. Connection is a one-time admin step. After that, everyone in the workspace can ask questions that span every connected source.

Generated artifacts, saved back

Anything Aleria produces — a cleaned table, a merged dataset, a report, a Flow output — can be saved back as a source. That's how a one-off analysis becomes a reusable dataset the next chat, agent, or Flow can pick up.

When a source updates, every Board, agent, and Flow that reads from it sees the update. No refresh button, no schedule, no "which snapshot is this?" conversation.

Shared memory

As your team uses Aleria, it builds up a workspace memory — conventions ("we always refer to Q1 as the fiscal Feb-Apr window"), people ("Priya leads procurement"), and past decisions ("we rejected the Acme proposal in February because of exposure limits"). New teammates inherit this context the moment they join the workspace.

You can inspect and edit the memory directly in workspace settings. Nothing is hidden.

Switching between workspaces

Use the workspace switcher in the top-left, or press Cmd + K and type the workspace name. Your chat history is scoped to each workspace, so switching away doesn't mix conversations.

Just ask

You type what you want. Aleria answers. That's the default experience, and it covers most of what people do every day. Questions, summaries, drafts, calculations, translations, code snippets, quick data lookups — they all come back in seconds.

No prompt-engineering required. Talk to it like you'd talk to a smart, patient colleague.

Voice input

You don't have to type. Press the microphone next to the input, start talking, and Aleria transcribes as you go. Useful when you're on the move, thinking out loud, or describing something too long to type (an incident, a meeting takeaway, a request with ten conditions).

Voice input supports English, Arabic, and a long list of other languages out of the box. You can switch between typing and speaking mid-message — the mic icon stays available the whole time.

The enhance button

Next to the send button is an enhance button. If you're not sure your message is clear — vague wording, missing details, too short — press it. Aleria rewrites your draft into a sharper version: more specific, better structured, more likely to produce the result you want. Review the suggested wording, then accept it, edit it, or keep your own.

It's especially useful when:

• You're about to ask for a big deliverable (a deck, a report, an app) and want to make sure the request is complete
• You're new to Aleria and still figuring out how much to say
• You want to turn a scribbled note into a well-formed task in one click

Upload anything

Drag files straight into the chat or paste them from your clipboard. You can attach as many as you want in a single message — they're read together, as one context.

What you can upload

• Documents — PDFs (even scanned), Word, Markdown, plain text, RTF, HTML, ebooks
• Spreadsheets — Excel, CSV, TSV. The whole grid, including formulas and multiple tabs
• Slides — PowerPoint and Keynote exports; Aleria reads text and image content
• Images — photos, screenshots, charts, diagrams, scans. Describe what's in them, extract text, compare to a baseline
• Audio — meetings, voice notes, calls. Transcribed on upload, diarized if possible
• Video — transcripts plus keyframe analysis for what's on screen
• Archives — ZIP files get unpacked; contents become attachable
• Code & logs — any text source file, any log format. Ask for explanations, refactors, error diagnoses

Generate any kind of file

Ask for a deliverable, and that's what shows up — not a description of one. A PDF report with your logo. A CSV of reconciled rows. A formatted Word contract. A deck with speaker notes. An image. Source code. A zip of many files.

Common outputs

• Reports — polished PDFs or Word documents with headings, charts, citations
• Slide decks — real .pptx or Google Slides files, branded, with speaker notes
• Spreadsheets — Excel or CSV, multi-tab, formulaic, formatted
• Images — diagrams, renders, logos, mockups, edits of uploads
• Documents — contracts, policies, memos, briefs — drafted from your templates and precedent
• Code — scripts, migrations, config, components — in any language your team uses

Everything generated is downloadable, editable, and saved to the chat thread. Ask for revisions in plain language ("make the margins tighter", "use our brand colors", "drop the footnotes") and a new version replaces the old one.

Build and deploy apps

This is the part that surprises people. You can ask Aleria to build something — a dashboard, a calculator, a custom viewer, an internal tool — and it doesn't stop at showing you code. It builds, tests, and deploys the app on a temporary Aleria domain, like yourteam.aleria.app/pipeline. You get a live URL in chat. Share it, open it on your phone, send it to a stakeholder before the meeting.

What counts as an "app"

• A one-page dashboard reading from your data ("pipeline by region, filter by owner")
• A branded calculator or quote tool
• An interactive document viewer or data explorer
• A lightweight internal tool (approvals, form intake, tagging)
• A prototype to show stakeholders before committing to a real build

When a chat becomes an agent

Some things don't finish in a few seconds. Reviewing every lease in a 400-file data room. Ingesting a database and building a schema map. Watching a folder for changes and summarizing what's new each morning.

When you ask for work like that, the chat promotes itself into a background agent. You'll see a status panel open with the steps it's running, the files it's touching, and the progress bar. You can close the tab — it keeps going.

When the agent finishes (or hits something it needs you on), you get a notification. Open the chat again and you'll see a complete log of every decision it made and every action it took.

What the chat has access to

Out of the box, the chat can reach:

• Your attached files and everything in the workspace library
• The workspace's sources — the databases you've created inside Aleria, the external systems you've connected, and the files your team has uploaded (see Sources)
• Skills — both built-in and the ones your team generated from your own know-how (see Skills)
• The web, when the answer isn't internal
• Other Aleria features — it can trigger a Flow, build a Board, or open a Fusion query when relevant

And in the other direction: anything the chat produces — a cleaned table, a generated report, a query result — can be saved back as a source for your team to reuse.

Sharing and continuity

Chats are scoped to your workspace, so teammates can pick up where you left off. You can share a thread with a direct link, fork a chat to take it in a new direction, or turn a recurring request ("every Monday, prep last week's pipeline summary") into a Flow in one click.

Shortcuts

Inside the composer, three single-character triggers bring up inline pickers so you can attach context without leaving the keyboard:

• @ — attach a file from your workspace
• / — invoke a skill
• $ — attach an agent template

And the usual keyboard shortcuts:

• Cmd/Ctrl + K — jump to any chat, file, or setting
• Cmd/Ctrl + Enter — send
• Shift + Enter — newline
• Cmd/Ctrl + / — toggle the keyboard-shortcut cheat sheet

What a skill is

A skill is a small, named capability: "reconcile a bank statement," "score a credit application the way we score them," "rewrite this in our brand voice," "pull the latest price from our rate sheet." Each one knows one job well.

You don't call skills by name. The chat picks the ones that fit the request and uses them while it works. You just see them in the conversation — "applied pricing rules", "generated PDF" — so nothing happens behind your back.

The fastest way to build one

Most skills you'll ever need are specific to how your team works: the rules, the judgments, the one-off conventions that live in a senior person's head or a long Notion doc. Aleria makes those repeatable.

1. 1

   Describe the skill in plain language

   Open Skills → New skill and type what you want it to do. Be as detailed as you'd be when explaining the task to a new hire. Paste relevant docs, screenshots, example inputs, and example outputs. Don't worry about "the right way" — the more context, the better the result.
2. 2

   Aleria drafts the skill

   In seconds you get a draft: a clear name, a description, the inputs it accepts, the output it returns, and the logic it follows. If your description included examples, those become the test cases.
3. 3

   Review and tune

   Read the draft. Adjust anything that's off — tighten the description, add an edge case, rename an input. You can chat with Aleria to refine it ("add a branch for tier-3 customers", "treat missing revenue as zero"). Test it with sample inputs right in the editor.
4. 4

   Save to the workspace

   Publish the skill. Now it lives in your workspace skill library. Every chat, every agent, every Flow in the workspace can use it — automatically, no one has to invoke it by name.

What kinds of skills people create

Almost anything in "how our team actually does it" is a candidate:

• Pricing rules — "quote a rate for product X in region Y for term Z, using our current rate card and volume discounts"
• Credit scoring — "score this application the way our credit team does: income, debt, tenure, risk flags, our adjustment for the sector"
• Brand voice — "rewrite customer replies to match how we sound: warm, concise, never promise a refund before checking eligibility"
• Contract review checklist — "read a SaaS contract and flag: unlimited liability, auto-renew with long notice, cross-border data transfer, anything unusual in termination"
• Tax classification — "given a transaction, tell me which VAT category it falls into, based on our internal taxonomy"
• Ticket triage — "classify incoming support tickets by severity and route to the right team, following our SLA rules"

Built-in skills

Every workspace starts with a standard kit that covers the generic 80%:

• File reading & generation — PDFs, docs, spreadsheets, slides, images, audio, video; produce any of the same formats
• Web search & fetch — search the web, read specific URLs
• SQL & data queries — against any connected source
• Charting — line, bar, pie, funnel, heatmap, rendered natively in chat
• Email, messaging, calendar — via your configured providers
• Code execution — Python, SQL, shell in a sandbox
• App building & deployment — deploy apps to an *.aleria.app subdomain

Authoring a technical skill (for admins)

If the skill needs to call an existing internal system (a pricing microservice, a custom database, a legacy API), admins can wrap that system as a skill directly. Three things to provide:

1. A name and description — clear enough that the chat understands when to use it
2. Inputs — the parameters the skill accepts, with examples (Aleria uses them to pick the right arguments)
3. An implementation — an HTTP endpoint, a SQL query, a script, or a function

These technical skills can be mixed freely with the ones you generated from description. The chat treats them all the same.

Permissions and audit

Skills respect your organization's access rules. A data skill only sees rows the user is allowed to see. A skill that sends emails sends as the user who triggered it. Admins can restrict sensitive skills to specific roles.

Every skill invocation is logged: which skill, which inputs, which user, what it returned. Admins can review the audit log from workspace settings at any time.

How skills fit with everything else

This is the story that connects the whole product:

• You save a skill — your team's pricing rule, say.
• You bundle a few skills into an agent template — "Deal Desk Analyst" — that uses the pricing skill plus your sources (CRM, product catalog).
• Teammates open that template as an agent and chat with it. Or the agent runs in the background.
• When the workflow becomes repeatable, you drop that same agent as a step inside a Flow — triggered on schedule, audited, approved.

One skill, many uses. Build once, leverage everywhere — in chat, inside templates, inside Flows, even as panels in a Board.

What a template actually contains

A template is a saved recipe with four ingredients:

• Instructions — what the agent is for, how it should behave, what tone to use, what "done" looks like
• Skills — the skills it reaches for first (your pricing rule, your credit scoring, your brand voice rewrite)
• Sources — the sources it knows about by default (CRM, contract library, policy folder)
• Examples — sample prompts and sample outputs, shown to new users so they know what to ask

Open a template and you get a chat that already knows what it's doing. You just tell it the specific thing you need today.

Create a template by describing the role

The fastest path — the one you'll use for anything specific to your team — is to describe the specialist in plain language. Aleria drafts the whole template for you. You review, you save, you use.

1. 1

   Describe the role and what it should do

   Open Templates → New template and write the brief: "A compliance reviewer for our bank's merchant onboarding. It reads a merchant pack, checks it against our policy, flags any risks, and drafts a summary for the approver. Tone: cautious, precise, cites the policy clause by number."Attach the policy docs, the sample outputs you want, and any specific sources it should always check.
2. 2

   Aleria drafts the template

   You get back a ready template: name, instructions, recommended skills (compliance classifier, PDF extractor, your policy lookup), recommended sources (policy library, sanctions list), and example prompts. Everything is editable.
3. 3

   Review and tune

   Read the instructions out loud — if anything feels wrong, fix it in place or ask Aleria to revise ("be more direct", "always cite the clause number", "don't approve anything marked tier-1 without escalation"). Open the skills panel to add or remove skills. Swap sources if needed.
4. 4

   Save to the workspace

   Publish. The template appears in your workspace library next to the built-in ones. Anyone can open it, chat with it, or drop it into a Flow as a step.

What's in the library

Alongside anything your team builds, every workspace has access to a starter set:

General business

• Legal Reviewer — contract review, clause extraction, risk flags, redlining
• HR Partner — policy Q&A, onboarding playbooks, job-description drafting
• Finance Analyst — statement analysis, reconciliation, variance reporting
• Procurement Analyst — vendor comparison, RFP drafting, spend analysis
• Compliance Officer — policy mapping, audit preparation, evidence collection
• Sales Partner — account research, proposal drafting, pipeline review

Industry-specific

• Real Estate — property valuations, due diligence, lease abstraction
• Financial Services — credit memos, KYC packets, portfolio analysis
• Telecom — spectrum filings, churn analysis, tariff comparisons
• Government — RFP response drafting, constituent Q&A, policy comparison

Using a template

Open the gallery from the workspace home, preview what the agent can do (example prompts, sample outputs), and click Open to spin one up. It behaves like any chat — everything in the Chat guide still applies. Upload files, generate outputs, deploy apps, promote to a background agent.

Adapting a template someone else built

Every template is editable, including the built-in ones. Clone "Legal Reviewer" and turn it into a specialist that reads your jurisdiction's case law and flags clauses using your house style. Changes are scoped to your workspace — you don't affect anyone else's copy.

You can change:

• Name, icon, and description
• Instructions — the brief the agent follows
• Skills — add or remove, including the custom skills your team has generated
• Sources — which sources it starts with
• Example prompts — what new users see when they open the template

Sharing templates across teams

Once you've tuned an agent to how your team actually works, save it as a workspace template. Anyone in the workspace can open it. New teammates get the same specialist on day one — no weeks-long ramp-up into tribal conventions.

If your organization runs multiple workspaces, admins can promote a workspace template to the organization level so every workspace inherits it.

The story so far

Before Flows, everything you built in Aleria was already composable. Flows are where it all clicks together:

• You saved your team's know-how as skills — pricing rules, scoring methods, brand voice.
• You packaged skills and sources into agent templates — Deal Desk Analyst, Compliance Reviewer, Ops Triager.
• Teammates chat with those templates as agents.
• Flows wire those same agents together as steps, with triggers, branches, approvals, and outputs. The agents you already trust become the components of your automation.

In other words: a Flow isn't a new system to learn — it's the machinery that runs your agents on a schedule, in sequence, with the guardrails a production process needs.

When to build a Flow

Reach for a Flow when you catch yourself doing something the same way twice a week. The signs:

• The task has more than one step, and the steps are fixed
• You need the output to look the same every time (same columns, same format, same audit trail)
• Someone needs to review what happened later — an auditor, a compliance officer, a new manager
• You want it to run on a schedule, or when a file lands, or when a customer submits a form

For one-off work, stay in Chat. For repeatable work, Flows pay back the setup cost fast.

The anatomy of a Flow

Every Flow has three parts: a trigger, a sequence of steps, and one or more outputs.

Trigger — what starts the Flow

• Manual — someone presses Run
• Schedule — daily, weekly, cron-style
• File drop — a new file lands in a watched folder or source
• Webhook — an external system calls Aleria
• Chat — a user types a phrase that matches the Flow's intent, and the chat offers to run it

Steps — the work itself

Each step is one of:

• Agent step — drop in any agent template from your workspace. The Flow passes inputs to it, the agent does its job using its skills and sources, and the output becomes available to the next step. This is the main kind of step.
• Data step — pull from a source, join, filter, aggregate. Uses Fusion under the hood.
• Code step — run a bit of Python or SQL you wrote, when you need precise control.
• Branch — take a different path based on a value ("if total > $1M, route to compliance").
• Approval — pause and wait for a human to sign off before continuing.

Steps pass data to each other automatically. You build the shape of the Flow on a canvas; Aleria handles the wiring.

Output — what the Flow delivers

• A file (report, spreadsheet, deck, PDF) saved to the workspace
• A write-back to a connected database or SaaS tool
• A notification (email, Slack, Teams) to the right person
• A Board update
• A new chat thread with the Flow's result attached

Building your first Flow

Open Flows in the sidebar and choose New Flow.

1. Pick a trigger. For your first one, choose Manual — easier to iterate.
2. Add your first step. Start with an AI step. Describe what you want it to do in plain language. Attach the files or sources it should read.
3. Add the next step. Chain steps together by dragging from one to the next. Each step can refer to outputs from any earlier step.
4. Define outputs. Tell the Flow where the result goes — a file in the workspace, an email to a team, a row in a database.
5. Test run. Hit Run. Watch the steps execute live and inspect the output at each stage.
6. Publish. Once it's working, publish the Flow. Now anyone in the workspace can run it (and schedules activate, if configured).

Versioning and audit

Every change to a Flow creates a new version. Older versions stay accessible — you can diff, restore, or fork. Every run of a Flow captures:

• Which version ran
• Who triggered it (or what triggered it)
• The inputs
• The output of every step
• The final output and where it was delivered
• How long it took and any errors along the way

Auditors can export the run history with one click. This is why Flows get used for regulated work — the answer to "what did the system do?" is always "here's the full log."

Examples people actually run

• Due diligence binder — a data room lands, the Flow pulls every document, extracts key terms, runs a risk screen, and produces a 30-page binder the deal team reads
• Monthly board prep — pulls KPIs from six systems, checks for anomalies, generates commentary, and drops a branded PDF in the board's shared space
• Vendor intake — new vendor form submission → compliance screen → contract draft → approval → e-signature routing → finance onboarding
• Daily anomaly scan — every morning, scan transaction data; if anything is off, summarize and page the on-call analyst

Chat vs. Flow — picking one

Start in Chat. If you end up running the same prompt more than twice, save it as a Flow. If the result needs to be auditable or scheduled, it's a Flow from the start.

You can always have both: a Flow for the daily scan, and chat for the "tell me more about this one case" follow-up.

The simplest way to picture a Board

A Board is a live dashboard. Three things make it different from the BI tool you're used to:

• It reads from everywhere, directly. Databases (Postgres, Snowflake, BigQuery), SaaS (Salesforce, HubSpot, Stripe), files you dropped into the workspace (CSV, Excel, PDFs), even event streams. Whatever you've connected as a source is fair game.
• It's online. No schedule, no nightly batch. When a row lands in the database, the panel reflects it. You're never looking at Friday's numbers on Tuesday.
• You can ask it things. At the top of every Board there's an ask bar. Type (or speak) a question and the answer comes back as a new panel. Don't like it? Dismiss it. Like it? Pin it.

What Boards read from

Boards are built on top of sources. The common starting points:

• A database you created in Aleria — the fastest path when you don't have an existing warehouse. Spin one up in seconds, push data in, and build the Board directly on it.
• An existing database or warehouse you connected as a source, for curated metrics
• Your SaaS systems — pipeline, tickets, finance, HR — through the connectors already in your workspace
• Workspace files — drop a spreadsheet in and build a Board on it in a minute, no ETL required
• A Fusion view when you need to join across several sources (see Big Data Fusion)
• Whatever agents and Flows have saved — a Board can sit on top of any source, including the tables your own Flows produce

Creating a Board

Open Boards in the sidebar, click New Board, and describe what the Board is for: "Quarterly pipeline by region and owner" or "Branch operations health for the Emirates region."

Aleria reads your connected sources, proposes a starting set of panels (charts, tables, KPIs), and lays them out. You keep the ones you like, swap the ones you don't, and move on.

Panel types

• Metric card — a single big number with delta vs. prior period (revenue, pipeline, open tickets, anomalies)
• Chart — line, bar, stacked bar, pie, funnel, heatmap, map
• Table — rows from your data with sorting, filtering, in-line editing if writable
• Narrative — a plain-English summary that updates as the data updates ("Pipeline is +12% vs. last quarter driven by three large deals in the Gulf region")
• Alert — a panel that's invisible until a threshold breaks, then appears at the top ("DSO over 60 days in Finance")

Multiple layouts on one Board

A single Board can hold several saved layouts — the same underlying data, arranged differently for different audiences or moments. Common patterns:

• Exec view — three big KPIs and a narrative panel, nothing else
• Ops view — the same Board with 15 operational panels, filters set to "today"
• Meeting view — just the panels you're walking through on the call, sized for a projector
• Personal view — your own filters and favourites, not shared with anyone

Each layout can have its own filter presets, time range, and pinned panels. Switch between them with one click, or share a direct link to a specific layout so the recipient lands exactly where you meant.

Querying — plain English, or SQL

Every Board has an ask bar at the top. You have two ways to use it:

Plain English

The default. Type the question the way you'd say it:

• "Break revenue down by product and region for Q2"
• "Which accounts slipped from commit to pipeline this week?"
• "Same chart, but exclude the renewal business"
• "Compare this month to the same month last year"

The answer comes back as a new panel on the Board (or a drill-down into an existing panel). If you like it, pin it to the current layout. If not, dismiss and it's gone.

Drop to SQL when you need precision

Switch the ask bar into SQL mode and write the query directly against any source in the workspace. Useful when the English phrasing would be clumsier than the query ("join these two tables on the fuzzy customer match"), or when you want exact control. Results show up as a new panel the same way.

Your query history is saved, so you can go back to any query you ran, tweak it, and re-run it. Promote a query to a full panel to make it part of the Board.

Drilling down

Click any panel to open the full view. From there you can:

• Slice by any dimension (region, segment, owner, date range)
• Jump to the rows behind a number
• Ask a follow-up question anchored to the specific panel
• Export the current view as CSV, PDF, or an image

Auto-drafted minutes and reports

When a Board is used for a meeting, Aleria can auto-draft minutes from what was discussed. Open the Board, run the meeting, ask your questions — at the end, click Generate minutes and Aleria produces a clean document: what was shown, what was asked, the answers, and the decisions recorded. It's a real file you can share or attach to the corporate record.

You can also schedule weekly or monthly reports built from a Board — a PDF shows up in people's inboxes without anyone building it.

Sharing and access

Boards are scoped to their workspace by default. You can share a Board with:

• Specific teammates — as viewer or editor
• A link — any workspace member with the link can view
• External recipients — read-only, with optional password and expiry, subject to your workspace's external-sharing policy

When your data needs reshaping first

If a panel needs to join five systems, clean messy records, or pre-compute something expensive, do the heavy lifting in Big Data Fusion and point the Board at the Fusion view. Boards sit on top of whatever Fusion produces — no change in how the Board feels to use.

What Fusion does

Three things:

1. Connects to your data wherever it lives — cloud warehouses, transactional databases, SaaS applications, file stores, event streams
2. Transforms that data — joining, cleaning, enriching, deduping, splitting — in pipelines you can edit and version
3. Serves the result — as tables other tools can query, as files in the workspace, or as inputs to Flows and Boards

Most teams who use Aleria heavily end up relying on Fusion for the messy parts. Chat answers are only as good as the underlying data — Fusion is how you make the data good.

Connectors

Fusion ships with 400+ connectors, covering the categories enterprise teams actually need:

• Warehouses & databases — both the major cloud warehouses and the transactional databases most teams run
• SaaS tools — CRM, billing, accounting, support, HR, project management
• Cloud storage — the major object stores and file platforms
• Streams & events — message buses, webhooks, change-data-capture feeds
• Aleria's own databases — the ones you spin up inside Aleria in seconds; Fusion reads and writes them just like any external system

If a specific system isn't in the catalog, add one via the generic REST or JDBC connectors, or have your admin build a custom one in-house. The full up-to-date list lives in the Sources → Add source catalog in your workspace.

Building a pipeline

A Fusion pipeline is a sequence of operations on data. You build them on a canvas:

1. Sources — pick one or more connected sources as input
2. Transforms — filter, project, join, aggregate, union, pivot, explode, enrich, dedupe, clean
3. Sinks — write the result to a table, a file, an existing warehouse, or keep it as a virtual view for other Aleria features to query

Every transform is previewable: you can see the rows flowing in and out at each stage, which makes debugging obvious.

Scale

Fusion is designed for enterprise data volumes. Pipelines process billions of rows in seconds, not hours. You don't need to pick "sample mode" or "full mode" — it runs at full scale by default.

Behind the scenes, Fusion uses a parallel execution engine that pushes work down to the source where possible (so a join against a warehouse runs in the warehouse), and otherwise runs in Aleria's compute fabric.

Governance and lineage

Every column in every pipeline has a lineage: which source it came from, which transforms it went through, who changed the pipeline and when. You can click any column in any output and see the full path back to its origin.

This is what makes Fusion usable for regulated work. When compliance asks "where does this number come from?", the answer is a clickable path, not a Slack thread.

Access control

• Row-level security carried through from the source (Fusion never exposes rows the user isn't allowed to see upstream)
• Column-level masking — mark a column as PII, and it's hashed or redacted everywhere it flows
• Full audit log of who read what, who changed what, when

How Fusion fits with the rest of Aleria

Fusion is the data plane. Everything else sits on top:

• Chat queries Fusion tables when a question is about your data
• Boards use Fusion views as the backing store for panels
• Flows pull from Fusion for inputs and push results back into it
• Custom skills can read from or write to Fusion tables

What teams do with it

• Customer 360 — join Salesforce, Stripe, Zendesk, and the product database into one clean customer table
• Regulatory reporting — assemble exactly the view a regulator wants, with full lineage for the audit trail
• M&A data rooms — pull financials, contracts, and HR data out of the target, clean them, and stage them for diligence
• Ops warehousing — feed a daily-refreshed warehouse that the whole org queries for operational decisions

What Video AI does

Video AI watches your streams the way a (very attentive, never tired) security analyst would. It detects people, vehicles, objects, and behaviors; it tracks things across frames; it raises alerts when something matches a rule you defined; and it keeps a searchable log of everything it saw.

You can query the log in plain English — "show me every vehicle that entered the east gate after hours last week" — and you get a scrollable list of clips with timestamps.

Connecting streams

Three ways in:

• RTSP / ONVIF — most enterprise and municipal cameras
• Cloud VMS — Milestone, Genetec, Avigilon, Eagle Eye, and similar
• Upload — for after-the-fact analysis of recorded footage

Each stream gets a name, a location, and optional metadata (what's in the frame, which area it covers). That metadata is what makes the system useful later — when you ask "anything unusual at the loading bay?", Aleria knows which streams cover the loading bay.

Detectors and rules

A detector is a rule that turns a pattern in the video into an event. The built-in library covers:

• People — count, cluster, presence, loitering, forbidden-zone entry
• Vehicles — count, type, direction, license-plate capture
• Objects — left-behind items, specific categories (bags, hazards, equipment)
• Behaviors — running, falling, queue length, crowd density, altercations
• Environmental — smoke, fire, water, spills
• Custom — train a detector on your own examples if the off-the-shelf ones don't cover what you need

Building a rule

A rule has three parts:

1. Where — which streams, which regions inside the frame
2. What — which detector, which parameters (e.g., "more than 5 people in the queue for over 2 minutes")
3. Then — what happens when the rule fires: log an event, send an alert, page on-call, trigger a Flow

Dashboards

Every stream can feed a dashboard — a Board tuned for video data. Panels you'd typically put on one:

• Live thumbnails of each stream
• Event counts by type over time
• Heatmaps of where activity clusters in the frame
• Dwell-time or queue-length curves
• Alert log — most recent at the top

Like any Board, you can ask it questions in plain English and pin the answers as new panels.

Alerts

Alerts fire through the channel you configure: email, SMS, push, Slack, Teams, webhook. Each alert carries a clip, a timestamp, the detector that fired, and a link back to the live Board. Recipients can acknowledge, escalate, or dismiss directly from the notification.

If the same rule fires a lot, Aleria rolls alerts up into one summary so the channel doesn't get flooded.

Searching footage

Ask the chat directly. For example:

• "Show me every vehicle entering the east gate after 10pm last week"
• "Was anyone in the server room between Friday 6pm and Monday 8am?"
• "Summarize what happened at camera 7 between 14:00 and 14:30"
• "Any incidents involving forklifts in the last 30 days?"

Results come back as a list of clips with timestamps and detector tags. Click any clip to watch it inline.

Privacy and deployment

Video is sensitive. Video AI is designed so none of it has to leave your environment. Standard deployment options:

• On-prem — all processing happens in your data center, no frames leave the building
• Edge — detection at the camera or gateway, only events (not frames) forwarded
• Sovereign cloud — runs in an Aleria region you pick, with data residency guarantees
• Air-gapped — fully disconnected install for the highest-sensitivity environments

See Deployment & security for the full picture.

The four deployment modes

Pick the one that matches your constraints:

Aleria Cloud

Managed, multi-tenant. You get an Aleria workspace, we run the infrastructure. Fastest to start. Fits most commercial teams. Data is isolated per tenant; every tenant gets its own encryption keys.

Sovereign Cloud

Aleria runs inside a cloud region you pick (UAE, EU, KSA, UK, and others), with data residency guaranteed. Useful when compliance says "data must not leave this country." Same managed experience as Aleria Cloud, with the locality constraint.

On-premises

Aleria installs into your own data center. Your team (or ours, managed) operates it. Data never leaves your infrastructure. Fits regulated industries, defense, and organizations with existing compute they want to reuse.

Air-gapped

No internet connectivity at all. Everything — models, connectors, updates — is delivered through a one-way physical channel. Fits the highest-sensitivity environments: classified networks, nuclear operators, certain defense programs.

Which one fits you

Answer these three questions in order:

1. Can your data leave your country? If no, start with Sovereign Cloud or on-prem.
2. Can it leave your data center? If no, on-prem or air-gapped.
3. Can the environment have internet access at all? If no, air-gapped.

If the answer to all three is "yes," Aleria Cloud is the simplest path and you can migrate later if constraints tighten.

Security posture

Encryption

• At rest — AES-256, per-tenant keys you can rotate or bring your own (BYOK / HSM)
• In transit — TLS 1.3 end to end, mutual TLS between services
• In use — confidential computing available on supported hardware

Identity

• SSO via SAML 2.0 and OIDC — Okta, Azure AD, Google Workspace, Ping, and others
• SCIM provisioning for automatic onboarding/offboarding
• MFA enforced at the organization level
• Session controls with configurable idle timeout and forced re-auth

Access control

• Role-based access at organization, workspace, and resource level
• Attribute-based access for sensitive skills and sources
• Row- and column-level policies carried through from connected sources

Audit

• Every action — reads, writes, skill invocations, flow runs — is logged
• Logs streamed to your SIEM (Splunk, Datadog, Sentinel, Chronicle)
• Immutable audit store with retention you control

Data handling

• No training on your data. Ever. Your data is used to serve your queries; it is never mixed into shared model training.
• Configurable retention. Set how long chats, files, and generated artifacts are kept, per workspace.
• Right to delete. Data removal is honored within contracted windows, with cryptographic verification of erasure.
• Regional models. For sovereign deployments, Aleria runs open-weight models locally so prompts never leave your boundary.

Compliance posture

Aleria is designed to fit regulated environments. Active certifications and alignments include ISO 27001, SOC 2 Type II, GDPR, and local regulatory frameworks in the markets we serve. For your current compliance package, talk to your account team — it updates as programs renew.

Operational considerations

• High availability — multi-zone by default on managed deployments; active-active on request for on-prem
• Disaster recovery — RPO and RTO targets configurable by tier
• Updates — managed deployments update weekly under a change-window policy you can configure
• Observability — metrics and traces exported via OpenTelemetry; logs streamable to your stack

For deep architecture details — the AI Factory, reference models, sovereign stack — see the AI Factory overview on the main site. For anything specific to your organization, your account team can walk through the right configuration.

Frequently asked questions

Does Aleria train on our data?

No. Your data is used to answer your questions; it's never mixed into shared model training. This is a hard product rule, not a setting you have to find.

Which language models does Aleria use?

Aleria selects the best model for each task from a curated catalog — including leading frontier models and open-weight alternatives. For sovereign and air-gapped deployments, Aleria runs open-weight models locally so prompts never leave your boundary. Admins can set a default model per workspace.

Can we bring our own model?

Yes, on on-prem and sovereign deployments. BYO-model fits teams with existing relationships or compliance constraints that require specific weights. Your account team will help with the integration.

How is billing calculated?

By user seat plus metered usage for heavy operations (long-running agents, large data volumes, video streams). Your plan details are in Pricing.

Can non-technical people really use this?

Yes — it's the whole point. The chat is designed to work like a colleague who happens to be fluent in your data. If someone on your team can write an email, they can use Aleria. Technical depth is there when you need it (custom skills, Flow authoring, Fusion pipelines), hidden when you don't.

What happens if the AI gets something wrong?

It shows its work. Every answer can be traced to the source files, the queries it ran, and the steps it took. If something is wrong, correct it in the conversation — the chat updates its understanding and keeps going. For regulated work, approvals in Flows keep a human in the loop.

Do deployed apps stay up forever?

Temporary deployments (*.aleria.app) are kept for a configurable window and then recycled. For apps worth keeping, admins can promote them to a permanent internal URL or export the code.

Can I export everything?

Yes. Workspaces, chats, files, Flows, and Boards can all be exported as structured archives. Aleria is not a lock-in product — we'd rather earn the renewal than trap the data.

Glossary

Agent

A chat that has been given a background task — it runs on its own, with a visible step-by-step log, and notifies you when done.

Board

A live dashboard of panels backed by your data, queryable in natural language.

Chat

The main interface. One box where you ask, upload, and get results. Covered in detail in the Chat guide.

Connector

A prebuilt integration to an external system (database, SaaS, storage) that feeds data into Aleria.

Flow

A deterministic, multi-step pipeline. Runs on demand or on a schedule; every run is logged.

Fusion

Aleria's data-engineering layer. Connects sources, transforms data, serves it to other features.

Pipeline

A Fusion construct: a sequence of data transforms from sources to outputs.

Skill

A specific capability the chat can use (run SQL, send an email, render a chart). Some built-in, some custom.

Source

A live place data lives in the workspace — a database you spun up in Aleria, an external system you connected, or a collection of uploaded files. Anything Aleria generates can be saved back as a source.

Sovereign

A deployment mode where Aleria runs in a cloud region you pick, with data residency guaranteed.

Template

A pre-configured starting point for a chat — role, skills, and sources bundled as a ready-to-use agent.

Workspace

A shared space for one team or project. Holds files, chats, sources, and shared memory.

Keyboard shortcuts

On Windows and Linux, substitute Ctrl for ⌘.

Release notes

New features, improvements, and fixes are announced weekly. The most recent releases are posted in the Newsroom, with full release notes available from inside your workspace under Help → What's new.

Getting help

• In-product — press ? or click the help icon in the bottom-right
• Your account team — the fastest path for enterprise customers
• Email — support@aleria.com
• Status page — current system health and incident history