How a Czech Smart-Metering Operation Automated Handover Protocols with Smartsheet + Pipedream
A Czech smart-metering operation used Smartsheet, Pipedream and Adobe document generation to auto-build 100–200 photo-rich apartment handover PDFs a month.
TL;DR: A boutique construction-operations consultancy in the Czech Republic, running Smartsheet for clients that install around 10,000 utility meters a year, replaced hand-assembled apartment handover documents with a Pipedream workflow that builds a photo-rich PDF the moment an apartment is marked complete, attaches it to the right Smartsheet row, and sends it to the owner. Its own estimate of the volume is 100–200 protocols a month.
Smartsheet handover protocol automation for meter installations. Photo by Godspower Abdulahi on Unsplash
The situation
The consultancy is small: a founder who designs operational systems in Smartsheet for construction and building clients, with plans to hire. Two of those clients work in advanced metering. Their technicians install water meters, heat meters, gas meters, electricity meters and heat cost allocators in residential and industrial buildings. Across both clients, that adds up to roughly 10,000 meters a year.
Everything already ran in Smartsheet. Each building project had its own installation sheet, with one row per meter, the apartment it belongs to, its serial number, its reading, and a photo taken by the technician on site. Sheets carry column names in both English and Czech. Location matters for this story for two reasons: the bilingual setup, and the European legal rules that came into play once signatures were involved.
The problem
When a technician finishes an apartment, the owner gets a handover protocol: a formal record of which meters were installed, their serial numbers and readings, and photos of each one. That document is proof of the work and the starting point for future billing.
Building it by hand meant pulling the right rows out of the right sheet, matching every meter to the right apartment, placing photos and readings into a template, saving a PDF, filing it against the apartment, and emailing it out. One apartment's protocol can carry a dozen photos. At 100–200 protocols a month, it was a steady stream of slow, error-prone document work landing on people who should have been scheduling and installing.
The founder was clear about what he wanted from the start of the build:
"Ideal way would be getting rid of all the manual steps that the technician needs to do in order to make this working." (the consultancy's founder, on a working session)
What they'd already tried
Two things had stalled before Connex got involved.
Smartsheet's own automations could react to changes inside one sheet, but this operation had many sheets, a new one for every building, and no clean way to run one shared document process across all of them. Copying automations into each new sheet by hand was the kind of setup step that gets missed.
The second wall was signatures. The original plan routed each protocol through Adobe Acrobat Sign. After talking to Adobe, the founder found it was the wrong tool. A legally verified e-signature platform expects identity checks, recipient accounts, and documents sent by email or SMS, which didn't fit an owner standing next to a technician in a hallway. He also priced it at roughly $6,000 a year for their signature volume. In his words, what the client actually needed was "basically like a physical signature on the touch screen of a device, saved to the PDF file. That's all."
What we built with Smartsheet and Pipedream
A Connex consultant worked with the founder in recurring one-hour ZoomFlow sessions, starting in 2024 with an AI meter-reading workflow and moving to the handover protocol system in 2025. Pipedream sits in the middle; Smartsheet stays the system everyone works in.
1. Every installation sheet joins the workflow automatically. Instead of hand-wiring each new building, a central register sheet collects sheet IDs. When a new installation sheet is added, the workflow creates the Smartsheet webhook for it and ticks a confirmation checkbox, so the office can see at a glance which sheets are live.
2. Marking an apartment complete triggers the document. The webhook fires on the completion column. The workflow gathers every meter row for that apartment (serial numbers, readings, photos) and fills a branded template using Adobe's document generation API. A typical protocol includes up to 12 photos, laid out in tables with the readings next to them.
3. The PDF lands where people already look. The finished file is attached directly to the apartment's row in Smartsheet as proof of generation and delivery, and the owner receives it by email.
4. Batches don't collide. When technicians close out several apartments at once, the workflow splits the events and processes each apartment on its own, so every owner gets their own document and nothing gets merged or dropped.
5. AI reads the meters, people stay in charge. In the first phase, technician photos are sent to an OpenAI vision model that extracts the serial number and reading and writes them back to separate columns on the same row, where they can be checked against the photo. Heat cost allocators were the hard case: the model kept reading the wrong part of the display. The consultant built a labeled set of 50 real images with correct answers and fine-tuned the model on it, and the founder took on training the field suppliers to capture photos more consistently.
Guardrails throughout. The signature step went through several versions (Adobe Sign, then a Google Drive and mobile PDF editor flow with a save button embedded in the document) before the end client changed its requirements and dropped on-site signing. Because the workflow was built in steps, removing signing and the Drive hop was a short change, not a rebuild. Later, deprecated Smartsheet calls were replaced with direct API requests after a connection broke. When the founder handed the system over to his client, all accounts and access tokens (Smartsheet, Google, OpenAI, Adobe, Pipedream) were moved to the client's ownership, and the workflow was paused during testing so no owner got a test email.
The results
100–200 handover protocols a month, generated automatically. Marking an apartment complete in Smartsheet now produces the finished handover PDF, files it on the row, and sends it. No one assembles documents by hand.
From ~2 minutes to ~2 seconds. The workflow went from about a 2-minute lag between completion and document to about 2 seconds after the delays were taken out.
Photo-rich documents without copy-paste. Readings, serial numbers and up to a dozen photos land in the right place in every document, pulled straight from the rows the technician filled in.
Every new project is covered. New building sheets join the shared workflow through the register sheet instead of a manual setup step, which matters for an operation installing around 10,000 meters a year.
~$6,000/yr in signature licensing not spent. Dropping the verified e-signature platform avoided what the founder priced at roughly $6,000 a year in licensing for their volume.
The $6,000 figure and the 100–200/month volume are the client's own estimates from working sessions; the 2-minute to 2-second change was measured during testing. No labor-hours estimate is given: the sessions never captured a reliable time per protocol before automation, and we'd rather show real scale than invent a number.
Client voice
On the session where the final version was tested end to end, the founder reviewed the generated PDF and called it "wonderful" and "perfect", then had the workflow unpaused for live testing. He has kept booking sessions as his client's needs changed, most recently to simplify the workflow after on-site signing was dropped.
FAQ
Will this work if every project has its own Smartsheet?
Yes. That was the core design problem here. A register sheet tracks which sheets are connected, and the workflow creates the webhook for each new sheet, so one process serves every building.
Do we need a paid e-signature platform for handover documents?
Not always. A legally verified signature platform solves a specific problem: proving identity remotely. If what you need is a record of completed work, possibly with a simple on-device signature, a generated PDF stored against the record may be enough. Check your own legal requirements first; this team did, and it changed the design.
How reliable is AI at reading meter photos?
Good on standard meters, weaker on unusual displays like heat cost allocators. A small labeled dataset (50 images here) plus consistent photo capture made the difference, and readings are written next to the photo so a person can check them.
What access did it need?
A Smartsheet account with access to the installation sheets, a Pipedream account, an Adobe developer key for PDF generation, and an OpenAI key for the meter-reading step. All of it ended up owned by the client.
Still building documents by hand from your Smartsheet data?
If your team finishes jobs in Smartsheet and then copies the same data into reports, certificates or handover documents, that step can run itself. Book a call and we'll look at your workflow together.
A hospitality staffing firm used Bullhorn and Zapier to turn booking, extension, and signed-agreement emails into self-updating tracking for ~100 consultants.
How a branded-merchandise company used Make and Pipedrive to automate self-serve store onboarding: next-step emails, 24-hour reminders and a 48-hour drip.
How a 3-tech equipment service business turned AI-answered calls into a technician dispatch board in Airtable that sorts calls, routes them, and closes jobs.
