X.24 agentic AI platform logo

Enterprise Agentic AI Platform

The intelligence that acts

Conversational automation answers questions. X.24 completes the request. It establishes intent, reasons against approved enterprise content and the customer record, then executes the action in your systems, such as booking, amending, verifying or paying. It operates on the phone and on every digital channel, in Arabic and English.

See an agent handle a real journeyRequest a demo

X.24 dashboard, shown with sample data.

X.24Talk to X-Shift
Customer appointment booking journey Illustrative
X X.24 Agent ● Online · typically replies instantly Today I want to move my appointment to next week 09:41 Of course. Your appointment is on Sunday at 10:00. I can see Tuesday 09:30 or Wednesday 14:00 are free. 09:41 ✓✓ Tuesday please 09:42 Done. Tuesday 09:30, confirmed. I have sent you the details and a reminder the day before. 09:42 ✓✓ Appointment updated · REF 48192 Reply… AGENT REASONING 2.4s Intent understood Reschedule an existing appointment Customer identified Record matched on the calling number Knowledge applied Rescheduling window and notice policy System action Availability read, slot written back Outcome completed Confirmation and reminder issued CONNECTED SYSTEM Appointment updated · REF 48192 Booking engine · synced 2.4s ago ESCALATION RULE Clinical questions go to a person Routed to the on-duty nurse within 30s Sample data. Handled end to end - no queue, no callback.

A representative view of the X.24 interface, built from the product's own model.

Start here

Where teams usually begin

This page describes the platform in detail, which is appropriate if you are running an evaluation. If you came here because something is not working, the platform is the second question; in a good number of these engagements, X.24 proves to be only a small part of the answer.

Customers still call after using the bot. Containment looks healthy and call volume has not moved. AI-Powered Customer Service The contact centre has been re-platformed twice and the numbers have not changed. The constraint is the operating model, not the technology. Contact Center Transformation The backlog is in finance, procurement or shared services. Judgement-intensive work that rules could not automate. Agentic AI & Enterprise Automation
Staying here? You are evaluating agentic platforms and need the specification.

What it does

What X.24 does

A request arrives, and something must then happen to it. X.24 is the layer that decides what should happen and then carries it out, on the phone and on every digital channel, in Arabic and English.

THE DIFFERENCE IS STEP FOUR

Most conversational AI stops at step three. Answering is not completing.

01

A request arrives

Spoken or written, on whichever channel the customer already uses. No menu tree or fixed phrasing is required; the customer states what they want.

OutputAn intent, not a menu selection
02

It establishes what is being asked

Intent, entities and the context around them, read against who the customer is and what has happened before.

OutputAn understood request with its context
03

It reasons against what it is allowed to know

Your approved content and your customer record, not a general model’s opinion. Where the answer is not available, it says so rather than inventing one.

OutputA grounded answer with its source
04

It acts in your systems

Books, amends, verifies, pays, raises, updates. The action is performed within the permissions granted and is recorded.

OutputA completed action, not a ticket
05

Or it hands over, with everything attached

Where policy, confidence or sensitivity says a person should take it, the handover carries the full context rather than starting again.

OutputA resolution, or a warm handover

The loop

Four steps, not one answer

Conversational automation stops at the answer. An agent carries the request through to a result, and reports whether it reached one.

01

Understand

What the person is trying to achieve, in Arabic or English, however they phrase it.

02

Reason

Against approved enterprise content, the customer record and what the policy permits.

03

Act

In the connected system: book, amend, raise, verify or pay, rather than handing over to a form.

04

Complete

Confirm the outcome, record it, and escalate with full context when it cannot be finished.

The X.24 execution loop

The challenge

Most bots stop where the work starts

Organisations have deployed conversational automation for years. Deflection improved; resolution did not. Answering a question and completing a request are different problems.

HOW IT WORKS TODAY What traditional bots do Answer questions from a scripted tree Route the user to a queue or a form Automate the first turn of a conversation Lose context the moment the channel changes Hand over to an agent with no summary of what happened Stops at step one. Resolution never happens. WHAT IT BECOMES What complex journeys need Context held across turns, channels and sessions Decisions made against policy, not keywords Live enterprise data, retrieved at the moment of need Real actions in real systems, not a promise Workflow execution, exception handling and a recorded outcome Carries the request all the way to a result.
HOW IT WORKS TODAY

What traditional bots do

  • Answer questions from a scripted tree
  • Route the user to a queue or a form
  • Automate the first turn of a conversation
  • Lose context the moment the channel changes
  • Hand over to an agent with no summary of what happened

Stops at step one. Resolution never happens.

WHAT IT BECOMES

What complex journeys need

  • Context held across turns, channels and sessions
  • Decisions made against policy, not keywords
  • Live enterprise data, retrieved at the moment of need
  • Real actions in real systems, not a promise
  • Workflow execution, exception handling and a recorded outcome

Carries the request all the way to a result.

Getting started

Your agent, live in minutes

Deploying an agent is a matter of configuration, not a build. You connect the channels you already run, brief the agent as you would brief a new team member, point it at the content you have approved, and publish. The first conversation does not depend on an integration programme.

NO MODEL TRAINING REQUIRED

Configuration, not a data science project.

01

Configure your channels

Create voice, web chat, WhatsApp, email and portal channels, connect provider keys, numbers and email identities, and set how conversations are routed by service, language or journey type.

OutputA live channel with a default agent assigned
02

Design your agent and its persona

Name it, give it a voice, a language and a dialect, then set instructions, tone, guided dialogue flows, handover rules and the boundaries it must not cross. This is the same brief you would give a new human agent.

OutputAn agent with a defined persona and remit
03

Add your knowledge in one click

Point it at documents, URLs or free text. X.24 indexes the content and shows you what is ready, pending, drafted or synced, so answers are grounded in sources you approved rather than in a general model.

OutputAnswers grounded in your approved content
04

Test, publish, then improve

Validate the agent against real journeys, publish it to the channel, then work from sessions, transcripts, tool-call logs, outcomes and cost in the dashboard, and adjust prompts, knowledge and tools based on what you see.

OutputA running agent with measurable outcomes

Platform

Everything an enterprise AI agent needs to work

X.24 is a platform, not a bot builder. You can use each part on its own. You govern all of them in one place.

01

AI Agent Studio

Design, configure, test and version AI agents. Define goals, boundaries, tone, escalation rules and the actions each agent is permitted to take.

and 3 more below

X.24 Illustrative
FEATURE AI Agent Studio Live AGENT PREVIEW X Booking Agent ● Testing · v1.4 I need to change my flight to Riyadh Sure. Your flight SV-1142 to Riyadh is on 12 Oct. I can move it to 14 Oct. Shall I confirm? Yes, please Type a message… Configuration Behaviour Tools PERSONA Polite · Concise · Bilingual LANGUAGES Arabic English BOUNDARIES No refunds · Escalate to human ESCALATION On low confidence Studio illustrative

AI Agent Studio

Design, configure, test and version AI agents. Define goals, boundaries, tone, escalation rules and the actions each agent is permitted to take.

Omnichannel AI

One agent definition serving voice and digital channels, with shared context so a customer never repeats themselves after switching.

Knowledge & Grounding

Answers are generated against approved enterprise content, with the source available. Unapproved content is not a source of truth.

Integration & Actions

APIs, enterprise systems, tools and workflow execution. The agent reads records, writes updates and triggers processes under permission.

Outcome Management

Define what a completed outcome is for each journey, then measure against it. Completion is the metric, not conversation volume.

Monitoring & Optimisation

Sessions, containment, performance, interaction quality, model cost and outcome rates in one operational view.

Governance

Permissions, traceability, data privacy, approval workflows and controlled release between environments.

Channels

Voice and digital, on the same agent

Voice is not an add-on here, and digital is not an afterthought. You define the agent once. One knowledge base. One set of permitted actions. It then works on every channel your customers already use, with context carried between them.

ONE AGENT · EVERY CHANNEL

Context follows the customer, not the channel.

THE CUSTOMER ARRIVES Voice Inbound & outbound calls WhatsApp Business API · AR / EN Web chat Embedded & portal Email Threaded · attachments Portal Self-service & forms Mobile app Native · push Messaging SMS & social Any future channel Same agent, same context One X.24 agent One knowledge base · one set of permitted actions · one governance One customer record, one case, one set of service standards Context follows the customer instead of restarting at every door Never repeats themselves Journey continues, not restarts One standard, every channel

Arabic

Arabic-first, Saudi by design

This is not an English product with Arabic added to it. The agent is built to hold a conversation in Saudi dialect, on the phone and in writing, and to act on what it understands.

01

Saudi dialect understanding

Dialect conversation rather than formal Modern Standard Arabic alone, so that customers can speak naturally and still be understood, verified and acted on.

and 2 more below

X.24 Illustrative
FEATURE Saudi dialect understanding Voice AGENT CONFIGURATION Arabic Agent Saudi dialect · bilingual DIALECT Saudi · Najdi LANGUAGE AR + EN LIVE TRANSCRIPT · AR أبغى أعدّل موعدي للأسبوع الجاي تمام. موعدك الحالي الأحد ١٠:٠٠ متاح الثلاثاء ٩:٣٠ والأربعاء UNDERSTOOD Reschedule Najdi Date · Time Next: confirm slot with customer → write back

Arabic knowledge grounding

Arabic source content indexed and retrieved alongside English, so an Arabic answer cites the Arabic policy rather than a machine translation of an English one.

Local and sovereign deployment

On-premise, private cloud, public cloud or a sovereign region, with the option to configure any LLM or bring your own, where residency or policy requires it.

Operating model

AI resolves, AI assists, humans handle

The question is not whether AI or a person should handle requests. It is who should handle this request, right now. X.24 routes on that basis.

ACCOUNTABILITY AT EVERY STAGE

A named owner and a defined output, start to finish.

Mode 01

AI resolves

The agent completes the request end to end. The customer gets an outcome, not a ticket number.

Mode 02

AI assists

A person handles the interaction with the agent supplying context, knowledge, suggested responses and the next best action.

Mode 03

Human handles

The agent prepares the case, summarises what has happened and hands it over with full context rather than a transcript.

Routing is determined by complexity, policy, model confidence and customer need, and every routing decision is logged.

Architecture

How a request becomes an outcome

A request arrives on any channel. The agent establishes what is wanted and who is asking. It retrieves what it needs from your approved content and your systems. It then acts inside the rules you set and records the result against the defined outcome.

FIVE LAYERS · ONE JOURNEY

From channel to outcome, everything logged.

01 · CHANNELS Where the request arrives Any channel, any language, same agent Voice WhatsApp Web chat Email Portal Mobile 02 · AI AGENTS Understand, reason, decide The agent holds intent, context and policy together Intent Reasoning Policy Handoff Confidence 03 · KNOWLEDGE & CONTEXT What the agent reasons against Approved sources only - never a general model Approved content Customer record Entitlements History 04 · ENTERPRISE ACTIONS Where the work is completed The agent writes back - not a promise to follow up APIs Systems of record Workflow Payments ERP 05 · MONITORING & GOVERNANCE Control and traceability Every decision recorded, every action auditable Permissions Audit trail Outcomes Model & cost tracking

X.24 reference architecture - illustrative

The dashboard

See what every agent is doing, as it happens

Every conversation an X.24 agent handles is measured in one dashboard: what is live right now, how this period compares with the last, which channels and hours carry the highest volume, and what each agent has resolved. Operations teams can identify problems before customers report them.

01 · LIVE MONITORING

Know what is happening now, not at month end

The dashboard opens on the conversations in progress and anything that needs a person's attention. It refreshes every minute, so supervisors watch live sessions instead of waiting for a report.

  • Live sessions across voice, chat and email, with one click into any conversation
  • Automatic alerts when evaluation accuracy, volume or duration moves sharply
  • Filter by agent, channel and period, in your local time zone
X.24 DashboardLive now · 3

Three conversations are in progress. One caller is booking an appointment in Arabic, right now.

Helpdesk Voice · ArabicLive session · 01:42
Needs attentionReview ›

Evaluation accuracy on Helpdesk Voice · Arabic dropped since Tuesday. I've flagged the affected conversations for review.

Accuracy fell from 92% to 78%
02 · PERFORMANCE AT A GLANCE

The numbers that matter, compared with last period

Every key metric sits next to its previous-period value and the percentage change, so you can tell at once whether things are improving or slipping, without exporting anything.

  • Conversations, average duration and messages per conversation
  • Prompt and completion tokens, so AI usage and cost stay visible
  • A trend line on every card shows the direction, not just the total
Conversations
0
↑ 18%
Previous: 1,088
Average duration
0m 0s
↓ 12%
Previous: 2m 32s
Messages per conversation
0
→ steady
Previous: 6.3
Token usageThis period
Prompt tokens
0
Completion tokens
0
PromptCompletion
03 · CONVERSATION VOLUME

See demand move, day by day

Daily volume for this period is drawn over the previous one, so spikes, quiet days and campaign effects stand out straight away. It helps you plan staffing, time outbound campaigns and spot unusual demand early.

  • This period against the previous period on one chart
  • Grouped by day in your time zone
  • Narrow it to one agent or channel to compare journeys
Conversation volumeLast 7 days
By day, in Asia/Riyadh
240180120600
This periodPrevious period
This period0
Peak dayTue · 236
Change↑ 18%
04 · CHANNELS & BUSIEST HOURS

Where demand comes from, and when it peaks

One agent can serve many channels. The dashboard breaks traffic down by channel, including enabled channels that carried nothing, and maps every conversation onto a day-by-hour grid to show your real peak hours.

  • Voice in Arabic and English, WhatsApp, web chat and email, side by side
  • A heatmap of conversations by day of week and hour
  • The busiest hour called out, ready for rota and campaign planning
By channel6 channels
Busiest hoursDay × hour
Busiest: Sunday 10:00 · 41 conversations None More
05 · AGENTS & OUTCOMES

Hold every agent to the result it delivered

Each agent has its own row: volume, average duration, messages per conversation and token use. The outcomes you define, such as booked, resolved or handed over, are recorded for every conversation, so you measure results, not just activity.

  • Volume and token use per agent for this period
  • Outcome definitions with versioned, recorded values
  • The latest conversations, newest first, one click from the transcript
AgentsView all ›
AgentConv.Avg dur.Msgs/convTokens
HDHelpdesk · Arabic6121m 58s5.81.12M
APAppointments · English4182m 21s7.10.86M
COCollections · Voice2542m 44s6.60.62M
Outcomes
Helpdesk Form · v2
68%
Resolved68%
Booked22%
Handed over10%
Recent conversations
Newest first
9665•••• 482114:02 · 6 msgs
9665•••• 193713:58 · 9 msgs
9665•••• 774013:51 · 4 msgs
Live, every minuteAuto-refresh keeps the view current without reloading the page.
Compared, alwaysEach figure sits next to the previous period.
Per agent, per channelFilter down to the journey you are responsible for.
Outcomes, not activityResults you define are recorded on every conversation.

Use cases

Where organisations start

A first deployment follows a consistent pattern: high volume, an unambiguous definition of completion, and a system of record the agent can reach. The channel mix varies; the pattern does not.

01

Beneficiary status enquiries with X.24

Situation

A ministry support line fills with beneficiary calls asking where an application has reached, and each answer requires an advisor to open three separate systems.

What happens
  1. Understand the request
  2. Retrieve and check
  3. Act in your system
  4. Confirm the outcome
Outcome

An X.24 agent retrieves the live status, explains the next step in Arabic or English, and records the contact.

Commonly applied in Government and across other sectors

02

Card disputes raised out of hours with X.24

Situation

Card disputes arrive through the night and the overnight team can only log a reference number, so investigation starts the following morning at the earliest.

What happens
  1. Understand the request
  2. Retrieve and check
  3. Act in your system
  4. Confirm the outcome
Outcome

Intake, classification and case creation happen on the call, with the block applied and the customer told what happens next.

Commonly applied in Banking and across other sectors

03

Outbound collections at cycle end with X.24

Situation

A telco runs payment reminders across its whole consumer base each cycle, and customers who answer want to agree an arrangement the campaign team cannot handle live.

What happens
  1. Understand the request
  2. Retrieve and check
  3. Act in your system
  4. Confirm the outcome
Outcome

The outbound agent confirms the balance, agrees an instalment plan within policy and updates the billing record during the same call.

Commonly applied in Telecommunications and across other sectors

More ways to apply it

Appointment booking and rescheduling with X.24

Availability is checked, the slot is reserved or moved, confirmation and a reminder are issued, and the clinic list stays accurate.

Healthcare

Outage surge on the fault line with X.24

The restoration answer is given on the automated channel, and the queue is held for customers reporting new damage.

Utilities

Order status after a promotion with X.24

Live status is retrieved and explained on whichever channel the customer used, with returns and replacements started in the same conversation.

Retail

Complaints into a multi-entity ministry with X.24

The complaint is classified, rated for severity, raised as a case against the right entity and acknowledged before the call ends.

Government

Seasonal volume the roster cannot cover with X.24

Repetitive contacts are handled by agents that scale with demand, and the trained team remains focused on work that requires judgement.

BPO

Back-office exception clearing with X.24

Enterprise agents read the records, apply the rule, complete the update and pass only genuine exceptions to a person.

Banking
Explore by industry

These are illustrative applications rather than accounts of delivered projects. What any one of them involves depends on the channels, systems and policies already in place.

Beyond service

Agents are an enterprise capability, not a service channel

The same platform that handles a customer request can handle an internal one. X.24 agents operate across finance, HR, procurement, operations and shared services, reading systems, applying rules, executing workflow steps and escalating whatever requires human judgement.

Enterprise AI agents
Workflow actions
System updates
Back-office execution
Task completion
Exception handling
Human escalation

Deployment

Deployed where your data has to stay

You choose the deployment model, the hosting location and the model configuration. The appropriate choice depends on the sensitivity of your data and the requirements of your regulator.

WHERE YOU NEED IT

Cloud, on-premise or a private environment.

01

Cloud

Managed deployment for organisations without a residency or isolation constraint.

X.24 Illustrative
DEPLOYMENT Cloud Managed CLOUD Channels Voice, chat, email, portal Inside X.24 runtime Agents, tools, orchestration Inside Model runtime LLM inference Inside Knowledge and data Indexed content, records Inside Fully managed by X.24 · illustrative

Cloud

Managed deployment for organisations without a residency or isolation constraint.

On-premise

Deployed inside your own data centre where policy requires the platform and the data to remain there.

Private environment

Dedicated, isolated environments where applicable to the organisation's requirements.

Flexible model configuration

Model selection configured per deployment to suit performance, language and governance requirements.

Deployment options should be confirmed against the current X.24 technical specification for any specific engagement.

Common questions

What buyers ask before an evaluation

Next

Where to go next

See it running

A working session against a journey close to one of yours, on the channel that matters most to you, including voice.

Recommended when you are comparing agentic platforms and need to see completion, not conversation.

Request a demonstration

Work out what should be automated first

Contact Center Transformation looks at the operating model, the demand mix and where automation actually pays before any platform is selected.

Recommended when you have automated before and the call volume has not changed.

Contact Center Transformation

Most organisations gain more from the second than from the first. We will advise which applies after one conversation.

Where this sits

Where X.24 is used

Deployed asAI-Powered Customer ServiceVoice and digital AI across the service operation.
Deployed asAgentic AI & Enterprise AutomationAgents that complete enterprise work, wherever it runs.
Deployed asOmnichannel Customer ServiceOne customer, one context, every channel.
Deployed asKnowledge & Self-ServiceAnswers grounded in approved content.

Also relevantContact Center TransformationX.Orbit

Industry applications

X.24 across sectors

The platform is the same in each of these. What changes is what it is configured to do and which systems it connects to.

Government & Public Sector

Beneficiary channels, case handling and compliance evidence, built as one programme.

How the portfolio applies here Telecommunications

Service cost per subscriber, handled through channel mix, AI and interaction analytics.

How the portfolio applies here Banking & Financial Services

Remote branch service, tracked cases and service journeys that stand up to audit.

How the portfolio applies here Healthcare

Patient access, case follow-up and remote consultation across a connected care network.

How the portfolio applies here Utilities

Billing, outage and connection demand handled across channels, field and back office.

How the portfolio applies here Retail & E-Commerce

What shoppers say online, what staff do in store, measured against the same standard.

How the portfolio applies here Automotive

Aftersales service, dealer networks and the retention that depends on both.

How the portfolio applies here Education

Applicant and student service that holds up through admissions peaks and audit.

How the portfolio applies here BPO & Shared Services

The platform, quality and managed operations layer behind outsourced service providers.

How the portfolio applies here Insurance

Claims, policy servicing and remote assessment, held against one customer record.

How the portfolio applies here Real Estate & Property Management

Tenant and owner service, maintenance cases and property condition, evidenced.

How the portfolio applies here Manufacturing & Industrial

Line safety, quality evidence and customer service across plants and distributors.

How the portfolio applies here Logistics & Supply Chain

Shipment enquiries, exception handling and condition verification across the chain.

How the portfolio applies here Aviation, Transport & Mobility

Passenger service at scale, terminal conditions and disruption handling.

How the portfolio applies here Hospitality & Tourism

Guest service across languages and channels, with requests that close.

How the portfolio applies here

These are the sectors in which this platform is most often applied. It is not limited to them; the requirement, not the industry, determines applicability.

Built on this

Solutions that deploy X.24

A solution is built from this platform, scoped to your organisation and delivered through to go-live.

Customer Service & Engagement

AI-Powered Customer Service

AI that understands the request, takes action and completes it, on voice and digital channels.

Process & Automation

Process Orchestration & Automation

Coordinate people, systems, rules and automation across one end-to-end process.

Process & Automation

Agentic AI & Enterprise Automation

AI that reasons towards a goal and completes the work in your systems.

Services that support the decision to deploy itContact Center TransformationAutomation & Orchestration

See X.24 against a scenario close to yours

We run demonstrations on a journey you actually handle rather than on a script written to show the product well.

Start a conversation
See X.24 on your own journeys

Your AI agent, live in minutes.

Tell us your top journeys and channels, and we will show X.24 handling them in Arabic and English with your team.

Book an X.24 demo