An AI copilot that actually runs the work.
Most AI features only answer questions. Xentral's copilot understands the screen and record you're on, then takes real action — creating contacts, deals, tasks, meetings, quotes and invoices on your behalf, scoped to your workspace and your role.
Bring your own AI keys · Permission-aware & audited · OpenAI · Claude · Gemini & more
Xentral AI
Your key · ClaudeAI that operates the company, not just a chatbot.
An assistant that only summarises text leaves the actual work to you. Xentral's AI is agentic: it reads your live CRM, communications and billing data, decides what to do, and performs the action — then confirms what it did with a clickable result.
Every action runs through the same permissions as the rest of the platform. A Support user can open a ticket but not draft an invoice; a Sales user can create deals but not change admin settings. Sensitive actions require explicit confirmation, and everything is logged.
The AI is available everywhere — a floating assistant on every screen, a copilot on every record, and module agents tuned for sales, support, finance, marketplace and operations.
Tell it what to do, and it does it.
Ask the copilot to add a contact, schedule a call for tomorrow at 3pm, move a deal to negotiation, draft a follow-up, open a support ticket or draft an invoice — and it completes the task with the correct data, in your workspace's timezone and currency.
It works from natural language in English or German, resolves the right records, and returns a result you can click straight through to.
- Create contacts, companies, tasks, deals and meetings
- Draft invoices and quotes with automatic numbering and VAT
- Open tickets, add notes, update records and move deals
- Timezone- and currency-correct, in plain language
Xentral AI
Your key · ClaudeYour AI, your provider, your terms.
Connect your own OpenAI, Claude, Gemini, Perplexity, DeepSeek, Grok or OpenRouter key. Usage is billed by your provider, runs on your account, and is never blocked by platform credits when you use your own key.
Keys are encrypted at rest, never logged or exposed, and resolved per user, per workspace — so each team controls exactly which model powers their AI.
- OpenAI, Claude, Gemini, Perplexity, DeepSeek, Grok, OpenRouter and more
- Per-user and per-workspace key resolution
- Encrypted at rest; never logged or shared
- Usage tracked per provider, model and feature
Sara Khan
Procurement Lead · Acme Health LLC
Open deals
3
Lifetime
AED 48k
Health
Good
Activity
Powerful, but never reckless.
The AI can suggest and draft freely, but it executes within the boundaries you set. Non-destructive actions run after a permission check; sensitive actions — sending, buying, deleting, changing credits or permissions — require explicit approval.
Every action is recorded in a usage log you can review in the admin and billing dashboards, so you always know what the AI did and on whose behalf.
- Role-based permission checks on every action
- Explicit confirmation for sensitive actions
- Full per-event usage and action log
- Workspace isolation end to end
Revenue
AED 412k
+18%
Pipeline
AED 1.2M
+9%
Win rate
38%
+4%
Pipeline development
Last 12 weeksAI woven through the whole platform.
Takes action
Creates, updates, schedules and drafts — not just answers.
Your keys
Bring your own provider; usage runs on your account and your terms.
Permission-aware
Every action respects the same roles as the rest of Xentral.
Everywhere
A floating assistant, a record copilot and module-specific agents.
Tracked
Per-event usage by provider, model, feature and action.
Plain language
Works from natural prompts in English or German, in your timezone.
AI — frequently asked questions
Whose AI keys does Xentral use?
Your own. Connect your OpenAI, Claude, Gemini or other provider key. Usage is billed by your provider, and the AI is never blocked by platform credits when you use your own key.
Can the AI really create records?
Yes. It can create contacts, companies, tasks, deals, meetings, quotes, invoices and tickets, update records and move deals — all scoped to your workspace and role.
Is it safe to let AI take action?
Every action runs through role-based permission checks. Non-destructive actions execute directly; sensitive ones require explicit confirmation, and everything is logged.
Which providers are supported?
OpenAI, Claude, Gemini, Perplexity, DeepSeek, Grok, OpenRouter and custom OpenAI-compatible endpoints, with both OpenAI- and Anthropic-style tool calling.
Where can I see what the AI did?
Every AI call is recorded as a usage event — visible in the workspace billing dashboard and the platform admin portal, with provider, model, feature and actions.
Put an AI copilot on every screen.
Connect your own AI key and let Xentral read your data and do the work — safely, on your terms. Start free.