A connected digital system for Dacha Plus and Karkas Plus. Enquiries, sales, home design and documents come together in one customer journey.
Selling a home means helping someone understand dimensions, layouts, insulation, materials and price. A lead card alone cannot do that.
We connected the customer experience with the company’s internal work. Essential information follows the deal from first contact to contract preparation. Select a stage.
A message or website form enters the shared process. The system retains its source, conversation history and the customer’s requirements.
We built CRM around the timber-home sales process. The pipeline, tasks, inbox, calls and project information come together in one interface.
Stages follow the actual sales process: introduction, contact, qualification, proposal, discussion and contract. Each deal has an owner and next action.
Work lists and daily planning surface tasks, due dates, outstanding replies and enquiries requiring attention.
Conversations, customer requirements, the selected home, proposals and documents connect to the deal instead of being assembled again.
The integration layer brings conversations and lead intake into a shared workflow. Source and message history support the entire project, beyond the first reply.
Channels use connected accounts and provider APIs. An implemented integration is distinct from uninterrupted provider availability; October diagnostics, for example, recorded Avito API restrictions.
We configured the AI assistant around the company’s domain: home use, seasonality, dimensions, land, packages and next steps. Conversation context and a knowledge base support more relevant responses.
The workflows account for pauses, follow-ups and human handover. Conversation history and collected requirements stay in the system. The knowledge base, prompts and rules can be developed further.
The conversation is an illustrative example of the workflow. It is not a customer quote or evidence that the AI is indistinguishable from a person.
Telephony connects to CRM: calls, recordings, transcripts and contact outcomes sit alongside the deal. AI workflows support requirement discovery and handover to a designer.
Schedules and handling rules are configured separately, including workflows for enquiries outside team hours. Recordings and statuses make call outcomes and technical failures reviewable. Availability depends on provider, routing and schedule; this case does not claim uninterrupted handling of every call.
We developed a custom home planning tool. The floor plan and spatial model share one screen, making design discussions tangible.
Exterior walls, partitions, windows, doors, rooms, floors and dimensions. Review the full plan or focus on a selected object.
Roofs, foundations, floors, extensions and terraces. Furniture and surroundings help communicate scale and possible use.
Move from design to package and estimate, prepare a proposal and use the project in contract preparation. Revisions keep changes traceable.
We connected house geometry to the calculator and product package. The selected option defines the specification, and the calculation supports the proposal. Managers can explain what is included and how packages differ.
Each catalog item can link to separate designs and package calculations. Customer projects are separate from templates, so catalog edits do not silently rewrite an existing proposal.
This example illustrates linked parameters. The exact estimate depends on the design, specification and agreed terms.
The product includes a 3D home view, camera angles, floors and roof controls. This case includes an interactive explainer: rotate the house and separate its structural layers.
Review the shape, roof, windows and entrance.
Understand the layout and room proportions.
Explain the structure and assembly sequence.
Open the customer presentation and explore.
Instead of scattered attachments, the customer receives a personal page. Visuals, a floor plan, specification, price and stages form one coherent presentation.
Visuals and the floor plan connect to the selected home. The interactive view supports a more informed discussion.
The package, estimate, validity period and payment stages appear alongside the design, ready for review before contract preparation.
Editing, preview, history and publishing separate internal preparation from the customer version. Proposal views are tracked separately.
People often choose a home on their phone. We focused on structure, generous visuals and a clear next step, from the first impression to the plan, package and discussion.
Customers do not need access to the internal CRM. The manager sends a prepared proposal that they can review when convenient.
The phone frame contains an actual mobile proposal screenshot with personal data removed. The English image is a localized presentation adaptation.
The next step is seeing the home in the context of a real space. The 3D module includes camera and AR workflows for compatible devices.
The home can be explored on a plot or as a tabletop model. Available modes depend on the browser, operating system and AR support. Server and browser checks have been performed; universal native AR operation on physical phones has not been established.
We built a contract module connected to projects and proposals. It uses house, estimate and package information, helps check required fields and generates the document set.
AI does not alter legal clauses. An uploaded receipt is supporting material, not automatic evidence that funds were received.
The catalog connects item codes, photos, packages, descriptions and projects. It provides a shared foundation for repeatable sales work and calculations.
Home variants and their specifications share one place. Standardizing descriptions preserves technical characteristics.
Photos, layouts and linked designs sit beside the item. Calculation readiness and discrepancies remain visible.
Shared product rules help the team discuss the same specification instead of different price-list copies.
The new Karkas Plus website presents the catalog, completed homes, layouts, packages and building process. Forms send enquiries into the shared workflow.
The customer website uses the company’s real photos and video. Illustrations are identified separately.
The catalog, filters, favourites and comparisons help turn general interest into a specific home.
An enquiry is saved before confirmation. Retries avoid accidental duplicates, and CRM delivery is tracked.
We created two display modes: Sales and Team Pulse. Leaders configure participants, periods, targets and display rules.
Reporting periods, team results and individual or shared targets. Closed-deal values remain distinct from cash receipts.
Configurable activity indicators give the team a shared basis for discussing work.
An error retains the last confirmed snapshot and shows a warning. Customer contacts and messages are not sent to office displays.
We connected tasks, due-date visibility and shared results. The team gets clear rules; leaders can see where progress has stalled.
An illustrative animation of the display concept. This is not a real employee ranking or a measured productivity uplift.
We analysed CRM exports, stage history, conversations, advertising spend and web analytics. The diagnostic work identified where a new process was needed, from first contact to project discussion.
Historical funnel by reached stage. Snapshot: 17 September 2026; enquiries from November 2025. This is diagnostic context, not the implementation outcome.
In the historical snapshot, 24% of enquiries reached the proposal stage and 1.86% reached a held meeting. We therefore connected communication, next actions, design and proposals, making progression observable and manageable.
The client estimates that proposal preparation fell from two hours to 5–10 minutes. Connecting the design, package, calculation and presentation reduced repeated data entry.
faster preparation of one proposal
Basis: client estimate supplied on 24 October 2026, rather than instrumented timing. Speed ratio: 120 ÷ 10 and 120 ÷ 5. Estimated time saved per proposal: 110–115 minutes.
of estimated working time saved
Scenario calculation based on the client estimate: proposals × 110–115 minutes ÷ 60. It is not measured payroll savings or a sales forecast.
This case shows the implemented system’s scope and design. Recorded data, client estimates and demonstration material are identified separately.
| MEASURE | WHAT IS SUPPORTED | BASIS |
|---|---|---|
| Proposal preparation | 2 hours → 5–10 minutes, implying a 12–24× speed ratio. | Client estimate |
| Conversations | 6,408 conversations; 7,567 AI-authored messages in August–17 September. | Archived records |
| Product catalog | 69 active items. 3D and calculation readiness is checked separately. | Live module |
| Funnel | 15,026 records in the historical snapshot. Analysis uses stage history, beyond the current status. | Archived diagnostics |
| Functionality | CRM, planner, proposals, contracts, catalog and office displays are evidenced by interfaces and source code. | Module verification |
The aggregate analysis uses the 17 September 2026 amoCRM export, stage history, module conversation table and monthly message-author log. Active catalog items were rechecked in the live module on 24 October 2026. Raw exports contain personal information and are not published.
CRM, planner, proposal, catalog and website images are screenshots of actual interfaces. Names, contact details and sensitive text are anonymised. English screenshots localize the case presentation and do not imply that the internal product has launched in English. The cover, AR scene and office setting are illustrations. The conversation and 3D explainer are demonstrations.
The implementation includes language-model integration, conversation context, a knowledge base, prompts and business rules. AI message counts are log entries, not autonomously closed sales. This case does not claim proprietary model-weight training or human indistinguishability.
A won CRM status and its budget do not prove payment or profit. Different traffic sources and periods are not treated as evidence of automation impact. Measuring uplift requires consistent event definitions, comparable channels and spend, and sufficient cohort maturity.
External APIs may restrict messaging and calls. Telephony behavior depends on configuration and provider. AR validation across physical devices remains incomplete. A catalog item does not imply that every package already has a linked, ready 3D model and estimate.
Behind the visuals are processes, data and integrations. We designed the system so that interfaces support the work and outcomes remain reviewable.
Website, forms, messaging and personal proposals. Customers see their project and next step.
CRM, tasks, inbox, planner, calculation, catalog and documents share connected data.
Roles, revisions, logs, integration diagnostics, analytics and office displays.
Internal operations are separate from the customer view.
Project and document changes remain traceable.
Persistence and retries support recovery from network failures.
Channel and background-process issues appear in diagnostics.
From analysing the original process to connected software modules. The project combines development, interface design, integrations, domain logic and data verification.
The funnel, sources, conversations, sales work, materials and home requirements.
A connected customer journey across enquiries, projects, proposals and contracts.
Operational modules, interfaces, integrations and the customer’s project presentation.
Data, roles, calculation links, user workflows and diagnostic limitations.
Tell us where your team loses time and context. We’ll examine the customer journey and design around your product, people and working rules.