People on site and in the workshop
Reporting things as they come — a photo, a voice note — without learning any management software.
operations + ai ecosystem
From a message on site to a published post
An ecosystem for an exhibition-design studio — field requests become tickets and tasks; one editorial flow becomes content for website and social channels, with AI and a person's final approval.
Context
A studio that designs and produces exhibitions — shows, spaces, events — lives between the workshop and the construction site. Operational requests are born there, in front of a wall to finish — a photo, a thirty-second voice note, a technical document photographed on the fly.
Somewhere on the way to the office, something always got lost. Meanwhile communication — website, social channels, portfolio — remained the postponed task, because it meant redoing the same work six times, once per channel.
The project was born to join the two worlds in a single ecosystem — the public website, the internal digital studio and a direct channel from the field.
The problem
- Operational requests scattered across personal chats, voice notes and photos, with no owner and no deadline.
- Complex quotations — bills of quantities with items, measures and amounts — managed with generic tools.
- Documents, suppliers, purchases and job orders without a single archive linked to the work.
- Publishing a project on website and social channels meant six versions of the same content — so it often did not happen.
Users and goals
Office and coordination
Seeing every request as a ticket with materials, deadlines and assignments; governing tasks, quotations and documents.
Whoever curates communication
Going from finished project to published content quickly, keeping the studio's voice.
- No field request lost — everything becomes a tracked ticket.
- Structured quotations and bills of quantities, with numbering, measures and professional print options.
- One editorial flow for six destinations, with realistic previews before publishing.
- AI accelerates; people decide — always.
The process
- Field analysis
Observing how requests are actually born — on site, by voice, with both hands busy.
- Ecosystem architecture
Three connected surfaces — public website, internal studio, Telegram bot — on a single base of data and access control.
- Iterative development
Modules released in sequence — board and tasks, then tools (CRM, quotations, documents, purchases, sales, invoices), then the multi-channel editorial flow.
- Operation and evolution
The platform grows with the studio — new needs become modules, without breaking what exists.
Architecture and modules
Three surfaces, one system — the field channel (Telegram), the Studio with its board and tools, and the multi-channel editorial flow feeding website and social channels.
Board and tickets
Field requests arrive as tickets with photos, videos, voice notes and documents, to be turned into tasks.
Tasks and calendar
Assignments with deadlines, priorities and areas — what is late surfaces on its own.
Quotations and bills of quantities
Offers structured by items and chapters, with measures, amounts, automatic numbering and PDF options.
CRM, documents, purchases, sales, invoices
The administrative tools linked to projects, in a single archive.
Document signing
Sending documents with a personal link and a dedicated signing page for the recipient.
Multi-channel editorial flow
One piece of content, six destinations — website journal and portfolio, LinkedIn, Facebook, Instagram and Stories.
The key flow
From the field to the ticket
The most important flow is the shortest one — whoever is on site should do nothing different from what they would do in a chat.
- Message
Photo, video, voice note, document or dictated text — sent to the studio's Telegram bot.
- Transcription
Voice notes become text; materials stay attached to the request.
- Ticket
The request appears on the board — "Arrived from the site" — with all materials, and a confirmation goes back to the sender.
- Task
The ticket becomes an assigned task, with deadline and priority.
The bot replies with a readable confirmation ("The office will find…") — whoever writes from the site knows immediately that the request has landed.
Write once, publish everywhere
The editorial flow is a four-step path, with AI preparing per-channel versions and a person approving.
- Write
Title, story, tone of voice (elegant, professional, warm, direct) and images.
- Destinations
Website journal and portfolio, LinkedIn, Facebook, Instagram, Stories — with toggles and a summary.
- Previews
AI adapts text and format to each channel; the preview shows the real result — phone included.
- Publish
Final approval and coordinated publishing to the selected channels.
Main features
Tickets with everything inside
Photos, videos, voice notes with a player, documents and transcriptions — the request arrives complete, not to be reconstructed.
Professional bills of quantities
Items and chapters with measures, units, amounts and automatic numbering; options for prices in the PDF, safety charges and public clients.
Tone of voice
The studio's voice is an explicit choice, not an accident — four defined tones for the editorial flow.
Per-channel previews
A Story is judged on a phone, a post in its own format — before publishing, not after.
Per-channel photography
For each destination you choose framing, crop and focal point of the images.
Built-in guides
Every section has its own "How it works" — the tool explains itself.
UX decisions
The field uses no management software
For people on site the interface is Telegram — zero training, zero forgotten passwords, zero friction.
The language of the physical world
Requests are "tickets" pinned to a board, with confirmations written the way a person would say them — the software adopts the studio's metaphors, not the other way round.
The preview is the truth
No blind publishing — each channel shows exactly how the content will appear, phone included.
AI proposes, people dispose
Every generated text remains editable and nothing goes out without explicit approval — the last button is always human.
Security and privacy
- Access control on every private route of the Studio, with centralised verification.
- The public signing page exposes no data — security lies in the personal link and dedicated APIs.
- Client and supplier data confined to the internal tools.
- No automatic publishing — final approval always belongs to a person.
- The images on this page are anonymised (names, references and data redacted).
Challenges and solutions
Challenge. Voice notes dictated on site arrive with noise, typos and broken sentences.
Solution. Transcription is a declared starting point — the text stays editable and the office confirms before it becomes a task.
Challenge. Six channels with different formats and tones risked fragmenting the studio's identity.
Solution. Explicit tone of voice, per-channel rules and real previews — consistency is built into the flow, not left to memory.
Challenge. An ecosystem growing for years must not break with every new module.
Solution. Independent modules on a shared base of data and access control, released and tested one at a time.
Outcomes
- Field requests land in one place, complete with materials, and become tracked tasks.
- Quotations and bills of quantities are produced on the platform, with structured PDF documents.
- Website and social publishing starts from a single editorial flow with real previews and human approval.
- The ecosystem is in production and keeps growing module by module.
The images come from the real system, anonymised — names, references and recognisable data are redacted.
Possible evolutions
- Direct link between field tickets and supplier orders.
- Channel metrics in the editorial flow (what works, where).
- Extension of the signing flow to more document types.
Does this process sound familiar?
Tell me about it: the first orientation call is free of obligations.