Automation is not “a bot instead of a person”. It is a request that does not get lost between tabs, an owner who hears about it at once, and a customer who gets an answer even when everyone is busy.
Six steps a person usually does by hand — and at each one a share of requests goes missing.
flow.vmtech.rs
Scenario “new request”
LIVE
written to CRM
A message on Instagram
source: Instagram
The request is recognised
booking
The customer is identified
customer since 2024
The task reaches its owner
assigned · 12 sec
The customer gets a reply
confirmed
The outcome in a report
written to CRM
THE REQUEST'S PATH
01A message on InstagramThe enquiry is pulled from DMs, comments, WhatsApp, a form or a call into a single list
02The request is recognisedThe system works out the subject — booking, price, order status, complaint — and picks a scenario
03The customer is identifiedThe CRM record opens: past requests, purchases, the account manager
04The task reaches its ownerThe request is assigned by rule: area, workload, working hours, customer priority
05The customer gets a replyThe confirmation goes back in the same channel the customer used, not “we'll call you back”
06The outcome in a reportYou can see how many enquiries arrived, how many closed without a human and where the process stalled
A FAMILIAR PICTURE
Where routine eats the team's time
Before the processes exist
Requests arrive in five places — DMs, WhatsApp, email, forms, calls — and nobody sees them together
The reply depends on who noticed the message first
Reminders about appointments and repeat visits are sent by hand, when someone remembers
The same customer retells their story to every new employee
Nobody can say how many enquiries came last week and how many went unanswered
Once the processes are described
Every channel flows into one list with conversation history and status
Routine questions close immediately; complex ones reach a person with the context ready
Reminders, confirmations and review requests trigger on events, not on memory
The customer record opens with the request — nothing is asked twice
A report shows the load by day, channel and employee
SCOPE OF WORK
What an automation project includes
Mapping the current process
We follow a request as it works today: who sees it, who answers, where it waits and why it gets lost.
One inbox for every channel
Instagram, WhatsApp, Viber, email, site forms and calls land in one shared queue.
Assignment rules
Who gets a request: by area, language, workload, working hours and customer priority.
Reply scenarios
Confirmations, answers to common questions and follow-ups — with a clear handover to a person.
Reminders and follow-ups
Appointment reminders, review requests after a purchase, nudges for abandoned baskets or invoices.
CRM and back-office links
A request becomes a deal or a task, and the outcome returns to the customer record.
Error handling
If an outside system fails to answer, the scenario does not break: it retries and alerts the owner.
Reporting and control
How many enquiries, from which channels, how many closed automatically and where a scenario stalls.
COMPARISON
A chatbot, a ready-made service and a described process
A button-based chatbot
An off-the-shelf messaging tool
A process by VMTech
Enquiry channels
One messenger
Email and SMS
Every channel in one queue
Knowledge of the customer
None
Just an address and a name
History, purchases, account owner
Complex enquiries
A dead end in the menu
Not covered
Handed to a person with context
Changing the logic
Within the builder's limits
Through the vendor's support
Around your process, without limits
What stays with you
Conversations inside the bot
Contact lists
The database, the scenarios and the access
TECHNOLOGY LANDSCAPE
What it runs on
Messenger APIsInstagram, WhatsApp and Viber connect through official interfaces, not emulationScenario engineConditions, branches, waits and retries when an outside system failsWebhooks and queuesEvents arrive instantly and heavy steps run in a queue without lossesCRM and databaseThe customer record, the request history and the outcome of every scenarioExecution logEvery step is recorded: what worked, what failed and exactly whereMonitoringIf a channel drops or a scenario stalls, we hear about it before the customer does
Messenger accounts, numbers and third-party subscriptions stay yours. VMTech designs the scenarios, builds the connections, wires them up and maintains them.
QUESTIONS
What people usually ask before we start
Where does automation start?
With a look at what already happens. We follow one request from the first message to closure and mark where it waits for a person, where data is retyped and where it disappears. Automating a mess is pointless — the process has to be described first, and that step alone often shows half the problems are solved by rules rather than software.
Will customers notice they are not talking to a person?
We do not pass automated replies off as a live employee. A booking confirmation, an order status or a reminder is a service the customer expects, and it reads as normal when it arrives fast and to the point. Anything outside the scenario goes to a person along with the conversation history, so the customer never repeats themselves.
We have no CRM. Is that a problem?
No. You can start with a simple queue of requests with history and owners — enough to stop losing enquiries. If a CRM arrives later, the scenarios connect to it: they write outcomes wherever you point them and are not tied to one system.
What if an outside service stops responding?
The scenario does not fail silently. The step goes into a queue and retries, and if the retry also fails the owner gets a notification with the error text and a link to the request. The request stays in the queue with a status rather than vanishing.
How long does it take to deploy?
The first scenario — usually the most painful process — goes live within a few weeks so the effect is visible at once. The rest follow one at a time: easier to verify each one works, and no big-bang launch that stops the business.
Won't the staff worry about their jobs?
Usually the opposite happens: the dullest part is taken off people — retyping data, digging for history, sending identical messages. The work that needs a human — the conversation, the disputed case, the sale — stays with people, and there is more time for it.