Alex
AI employees / Alex · Customer service

Alex

“This one needs Karla in billing. I've attached the four messages and told the customer she's coming.”

< 4s
median first reply
80%
resolved without a human
100%
of first contact answered
+35%
appointment attendance
The process he owns

Every customer who writes should know within seconds that somebody has read it.

He takes it from the moment a message arrives to the moment the question is answered or somebody named is holding it. Throughout he works from your resolution protocols rather than a generic script, and he is measured on what he hands over as much as on what he closes. Anything touching money or an exception stays human. What changes is that nobody is left waiting to find out whether they were heard.

The artifact

The conversation handoff, exactly as Alex filed it.

Every line carries the check that produced it, and anything held stays visible with its reason attached. Structure and logic from a live deployment; values changed.

ConsoleTue 24 Jun 15:38
Alex — conversation #8821 · whatsapp
AlexCustomer serviceHuman handoff
Conversation #8821 · WhatsApp11:07filed at
10:58 · customerThird charge for the same order, asks for a personReceived
10:58 · AlexMatched to order #3982 — three charges confirmed realClear
11:01 · AlexRefund not attempted — outside his protocolHeld
11:02 · AlexCustomer told a person is coming, and by whenClear
11:07 · Karla M. · billingThread and charge history receivedEscalated
Customer message · WhatsApp · 10:58

«Ya van tres veces que me cobran el mismo pedido y nadie me contesta. Quiero hablar con una persona.»

Alex’s decision

He confirms the three charges are real, tells the customer a person is coming and by when, and then stops. Issuing the refund is outside his protocol, so the thread goes to Karla with the charge history attached — not a promise he can’t keep.

conversation #8821 · resolved 11:34 by K. Mendoza, Billingrefund issued by a human · 36 minutes from first messageFull audit trail →

What Alex does.

Every capability trained on your knowledge base, your resolution protocols and your brand voice — not a default template.

01

Answers on every channel at once

WhatsApp, web chat, Instagram and Facebook, with the same answer and the same tone on all of them.

02

Books, moves and confirms appointments

Straight into your calendar, with the reminder scheduled and the reschedule handled without a phone call.

Appointment confirmed · dental cleaning · 25 Apr 10:30 · reminder set for 24h before
03

Resolves what your protocol covers

Returns, order status and the common complaints, using the resolutions your team already approved.

04

Answers the questions before the sale

Products, pricing, hours and policies, from your knowledge base rather than a guess.

05

Knows when to bring in a person

Recognises the conversation that needs judgement and hands it over with full context, rather than looping.

▲ Escalated to Karla (billing) — customer #8821, four-message history attached
06

Reports what customers keep asking

Captures satisfaction and surfaces the recurring question that should have been a policy change.

Sound like what your support team needs?
How he gets specialized

He doesn’t arrive knowing your operation. He is taught it, on your records, before he touches anything live.

Four stages. You see his calls next to your team’s before anything is delegated.

Weeks 1–2

Diagnostic

Someone from our forward-deployment team works your inbox by hand for two weeks — the real volume, the real tone, the requests your protocols don’t cover.

Weeks 3–4

Your protocols, ingested

The knowledge base, the refund and rescheduling rules, and the voice your brand actually uses rather than the one in the guidelines. Permissions scoped per system in the same pass.

Weeks 5–8

Shadow mode

He reads a quarter of closed conversations and drafts replies nobody sends. You compare them to what your agents said, and the differences tune the protocols.

Month 3

Live

He answers and books. Refunds, credits, price changes and anything outside protocol stay with your team.

First usable results in weeks · full operation ≈ 3 months
Reach

No new platform. He works inside what your team already opens every morning.

Customers write where they already are — WhatsApp, Instagram, the web chat. Alex is on all of them with the same answer, in the language they used.

Systems

Connected on the systems you already run

WhatsApp BusinessWeb chatInstagramFacebookCRMGoogle Calendar+ any system with an API
What he handles

Everything that arrives as a message

MessagesVoice notesWeb formsEmailTranscribed calls+ anything else that arrives
In production in

Where Alex is already deployed

HealthcareRetailAutomotiveHospitalityServices+ more
Languages
Spanish · English
Same rules, either language, in one run
Regions
Mexico · US · LatAm
Data stays in the region you nominate
Availability
24 / 7 / 365
Work picked up as it lands, not in a batch
Time to launch
≈ 3 months
First usable results inside the first weeks
Scope, escalation and error

What he can reach, what he can’t, and what happens when he is wrong.

Every permission is set by you, per system, and visible on the trust page. Alex talks to your customers and cannot move their money.

Permissions, as deployed
SystemReadWrite
CRM · customer recordsyesnotes only
Calendar · bookingsyesbook and reschedule
WhatsApp Businessyesreplies only
Order systemyesno
Refunds and creditsyesnever
Issue a refundn/anever
When a conversation needs a person
01 · Hold

He stops rather than improvising, and tells the customer so in plain language — including when a person will reach them.

02 · Escalate to a named person

Within 60 seconds to the agent who owns that queue, with the full thread and the order history attached.

03 · When he’s wrong

He misreads about 3% of conversations — usually an unusual request phrased like a common one. The customer is handed to a person, the misread is logged, and the protocol is updated.

04 · The line he doesn’t cross

Alex cannot issue a refund, change a price or commit to anything outside your protocols. Those stay with your team.

Results

The workload, redistributed.

What a six-person call centre was carrying alone, and what it looks like after Alex has been through shadow mode and gone live.

Before
Your team, alone.
Agent hours / week on first contact168
Error / miss rate22%
Process coverage35%
With Alex
The same team, on the hard conversations.
Agent hours / week on first contact20
Error / miss rate3%
Process coverage100%
One dental clinic chain that previously ran a six-person call centre · six months either side of go-live · the team moved onto treatment coordination.
100%
Healthcare

First contact without a call centre

Every message answered within seconds, with 85% of rescheduling and routine questions closed without an agent touching them.

Dental clinic chain · MX · NDA
85%
Retail

Orders and returns, handled overnight

Order status, returns and product questions resolved on WhatsApp and web chat at the hours customers actually write.

Omnichannel retailer · MX · NDA
40%
Automotive

Support and reconciliation in tandem

Alex handles customers while Memo matches the bank movements: two fewer administrative hires and a team that isn’t underwater.

Automotive group · MX · NDA

Put Alex on your first contact.

Two weeks of diagnostic on your real conversations, with our forward-deployment team on site.

Talk to us on WhatsApp
No commitment · 2-week diagnostic · clear proposal