Case · Cleaning DM BLYSK · Odesa and region

DM BLYSK: a website, CRM and AI assistant for cleaning services

A customer writes in web chat, Telegram or WhatsApp. The assistant asks about the job, shares a calculator estimate and records the enquiry. For DM BLYSK, we connected that conversation to the team's daily work: the manager receives a record with the address, scope, price and customer preferences, ready to arrange the cleaning visit.

The DM BLYSK website, assistant conversation, enquiry and proposal
01

Challenge

DM BLYSK cleans homes and commercial properties, washes windows and provides upholstery cleaning in Odesa and the surrounding region. Customers usually start with a simple question: how much would cleaning my property cost? The answer depends on the job. An apartment needs questions about floor area and cleaning type. A residential building also needs details about floors, lifts and visit frequency. Each service has its own questions and pricing rules.

We needed to connect the website and messenger conversations with the work that follows. Customer answers should reach the enquiry record, with the estimate beside the agreed details, so the manager can move on to arranging the job. One conversation can cover several services. A customer may compare options, add work or accept only part of a proposal. The system needs to keep those choices together as the enquiry progresses.

The owner also needed to manage everyday changes after launch: update a rate, rewrite a service description, add a frequent question or teach the assistant a new company rule. We built a management panel with separate Ukrainian and Russian content fields. The website and working tools can then follow changes in the company's services.

39Pages in two languages
18Services on the website
3Assistant channels
78Addresses across two languages
02

What we built

The customer journey from choosing a service to the cleaning visit.

Customer

Helped customers find the right service

We built 39 pages in two languages, including 18 service pages covering different properties and cleaning needs. A quiz with 10 categories asks questions relevant to the customer's job. Visitors can move from a service page to an estimate or start a conversation to ask for advice.

System

Put pricing rules into calculators

We built nine calculators for regular services, plus dedicated calculations for residential building associations and pharmacies. Each uses the conditions relevant to that service, including area or volume, selected work, property type, extras and minimum charges. Customer answers become an estimate based on the company's rules, ready to use when handling the enquiry.

Bot

Connected the assistant to three channels

Web chat, Telegram and WhatsApp follow the same conversation rules. The assistant asks about the service, scope, address, preferred date and contact details, obtains a calculator estimate and offers to record the enquiry. Customers stay in their chosen channel while their answers prepare the request for the team.

System

Brought the conversation into the CRM

With the customer's consent, the assistant creates an enquiry with their contact details, address, channel and preferences. Several services from one conversation remain parts of the same enquiry, with calculations attached to the relevant work. The manager sees what the customer has already explained and the price they discussed.

Manager

Gave managers estimates and proposals

The enquiry record brings together questionnaires, estimate options, the current stage and the next task. A manager builds a client proposal from saved calculations and records the customer's decision on the full scope or selected parts. The discussion, figures and agreed work stay together throughout the enquiry.

Manager

Connected agreed work to the visit schedule

Accepted work becomes orders and cleaning visits, with recurring services organised as a series of future visits. Residential building maintenance has its own planning calendar. Managers arrange dates in the same system that holds the customer, property and agreed services, so the team can see what needs doing.

Assignee

Prepared a workspace for the cleaning team

Team members open their assigned visits, work through checklists and add completion photos. Their access is limited to their own assignments, while managers and the owner have separate permissions. The people doing the work continue the process that began with a question on the website or a messenger conversation.

Owner

Put everyday updates in the owner's hands

The owner edits prices, service descriptions, FAQs, quiz questions and the assistant's knowledge base in the panel. Saving updates the related website content, with separate fields for Ukrainian and Russian text. Important actions are recorded in a change history so the team can check who edited what.

03

Sonia gathers the details and records the enquiry

A conversation starts with an ordinary message. The customer may give the area and address straight away or ask about a service first. Sonia, the DM BLYSK assistant, picks up the request and asks for what is missing. Required booking details are configurable, so the questions follow the particular service.

The calculator works out the price from the customer's answers. Sonia shares the result in the conversation, and a changed amount needs confirmation when the details change. Before creating an enquiry, the assistant asks for consent. The manager later sees the same estimate alongside the job details.

Complex jobs, including windows, upholstery cleaning, communal areas and residential building maintenance, outdoor areas and pharmacies, go to a manager for an estimate. Photos help assess the property's condition; the assistant does not determine area or price from a photo.

This short exchange comes from a test run in Ukrainian. The address and phone number are masked; the wording and amount are preserved. The customer asks for a deep clean of a 55 m² apartment the next day. Sonia gives an indicative price of UAH 5,500, asks for contact details and records the enquiry, with the manager confirming the visit.

The assistant replies in Ukrainian and understands requests in Russian. A preference such as ‘please write here’ stays in the manager's summary alongside the other arrangements. Details sent after submission can be added to an existing enquiry. If photo processing or the conversation fails, the request can pass to a person with the context collected so far.

A test conversation in Ukrainian. Address and phone number are masked.

Клієнт

Хочу генеральне прибирання квартири 55 м², Прикладна 1, завтра

Соня

Вітаю, я Соня з DM BLYSK. Генеральне прибирання квартири площею 55 м² коштує орієнтовно 5500 грн. Адресу Прикладна, 1 в Одесі зафіксувала… Як до вас звертатися і який номер телефону залишити для зв’язку?

Клієнт

+380 67 000 12 34

Соня

Заявку прийнято, номер DM-212F9. Записала генеральне прибирання квартири площею 55 м²… Менеджер передзвонить протягом 15 хвилин; виїзд підтверджує менеджер.

The CRM now holds an enquiry with the address, floor area, estimate and source channel.
04

Customers can agree to the work they actually need

One cleaning enquiry often includes several different jobs. The CRM keeps them together while letting the manager work on each part: complete a questionnaire, clarify the scope and prepare an estimate. It shows which services were discussed, what has been calculated and what to ask next. After a pause, the manager can pick up the conversation from a specific point.

An enquiry can hold several estimate options. The manager selects saved calculations to build a proposal listing the work and total for the customer. Internal cost calculations stay in the working part of the CRM. A separate print view lets the manager save the proposal as a PDF through the browser.

A saved estimate keeps the terms used when it was calculated. The owner can update the price list later while the earlier proposal retains its figures. The manager can return to it and check exactly what amount was discussed and what work it covered, even when time passes between the estimate and the customer's decision.

Customers can accept part of a proposal. The selected work connects to orders or recurring visits, giving the team a clear scope to carry out. The system prevents an already fulfilled saved package from being sold again. From the first calculation to the visit schedule, the manager can trace the customer's decisions and the work agreed.

05

Stack and reliability

Technology

Next.js, PostgreSQL and Docker. The website, CRM, calculators and Telegram and WhatsApp connections work as parts of one system.

Updates and backups

During routine updates, the proxy holds requests while the application restarts. A daily database backup is configured.

Team permissions

The owner, managers and assigned team members have separate permissions enforced on the server. Important operations are recorded in a change history.

Forms and consent

Incoming requests have phone validation, rate limits and spam protection. Forms ask for consent to process personal data, and the cookie banner saves the visitor's choice.

06

Website and workspace

Select an image to view the details.

The live site: first screen and the services strip
07

Frequently asked questions

We start with a free review of your services, pricing rules, team and enquiry channels. That gives us the scope for an individual project estimate. For DM BLYSK, we connected the website, calculators, chats and cleaning visits. Your workflow determines what your system needs.

Yes. The DM BLYSK assistant asks about the job, collects contact details, the address and preferred date, then creates a CRM enquiry with the customer’s consent. The calculator estimate stays with the enquiry. A manager can open the record, agree the work and arrange the visit.

We connect incoming messages to the assistant conversation and the CRM enquiry record. DM BLYSK uses shared rules across WhatsApp, Telegram and web chat to collect details, create enquiries and hand conversations to a manager. For a new project, we first check the company’s access and agree how the team will continue each conversation.

Yes. The owner can edit prices, page text, FAQs, quiz questions and the assistant’s knowledge base in the management panel. Saving updates the related content. Existing estimates and proposals retain their original pricing terms, so a manager can always refer back to the figures discussed with a customer.

The enquiry goes to a manager with the details already collected. This handover also covers failures when processing a photo or continuing a conversation. The manager can review the context, ask any remaining questions and help the customer. Contact preferences are kept in the enquiry summary too.

Built for your business

How do customer enquiries reach your team?

Tell us where your customers contact you, how you calculate prices and what your manager copies by hand. In a free review, we will identify the steps and connections your business needs. We estimate development individually for your workflow.

CRM for business
— Also

Other services

— Journal · insights

From our blog

All articles →
Call+38 095 108 62 12 +38 097 599 66 89 Telegram Viber WhatsApp