Process automation
Repetitive tasks in sales, marketing, or operations—automated using n8n, Make, or custom Python. Say goodbye to copy-pasting between five different tools; generate quotes, reports, and data without a team clicking away all night long.
If you recognize at least two of these—then it makes sense to talk about automation.
01
A lead from a form entered manually into the Customer Relationship Management. A quote generated in Word, sent via email, and then entered back into the Customer Relationship Management as a value. A returning customer—you have to search for their name across three systems. This isn’t a problem with the tool; it’s the lack of an automation layer.
02
Newsletters, post-trade show follow-ups, payment reminders, segmentation by lead source. Every "send" requires clicking through three systems. Without automation, the Same message arrives late or not at all.
03
PDF invoices received via email entered into the accounting system, orders from Allegro transferred to the warehouse, monthly reports compiled from three Excel exports. Every repetitive task yeses hours each week and leaves room for human error.
04
The SaaS pricing is calculated per task/operation. When you’re dealing with tens or hundreds of thousands of operations per month, SaaS becomes more expensive than running n8n on your own self-hosted server (VPS starting at 5PLN 0/month). You can scale without paying a per-task fee.
A clickable prototype in a test environment before we write a single line of production code. Payment in four installments.
Week 1
A process map showing the current state of affairs (who does what, in which systems, and how long it yeses). A list of integrations (Customer Relationship Management, ERP, email, Allegro, online store). A decision on what to automate in the MVP and what to address later. Deposit: 30%.
Weeks 2–3
A diagram of each workflow: trigger, steps, error branches, retry policy, alerts. Choice of stack—n8n self-hosted, Make / Zapier, or custom Python. Architectural decisions laid out in two pages, not fifty.
End of Week 3
You approve the prototype. We won't start building the MVPithout your approval. Upon approval: another 30%.
Weeks 4–7
Workflow implementation with technical documentation. Each step is tested in a sandbox using your actual (anonymized) data. Weekly demos of the working pipeline. After the MVP release: another 30%.
Week 8
Migrate the workflow from the sandbox to production. Set up monitoring (health checks, Slack/email alerts if the pipeline fails). Document the process and hand it over to the client’s team—you can modify the workflow yourself. After implementation: last 10%.
Salespeople return to selling instead of administrative work. The back office no longer has to piece together reports from three different exports. It’s not about “productivity” in the abstract—it’s about being able to handle a higher volume of leads or orders with the Same staff.
The lead from the form won’t disappear. The price from the quote is sent to the Customer Relationship Management exactly as it was accepted. The PDF invoice goes to accounting without any typos in the tax ID. Everyone makes misyeses. Workflow doesn’t.
The workflow can handle a volume limited only by server capacity, not by the number of staff. As you expand your sales funnel, operating costs increase only minimally. Hiring another administrative assistant costs 6,000–8,00PLN 0 per month, while an additional workflow typically costs a fraction of that.
Every event is logged: when, who, what request, what response. GDPR-compliant, easy to debug, easy to generate management reports. "Why didn't the customer receive a quote?" — an answer in 30 seconds, not after three days of asking around the team.
Apollo Cold, Brevo welcome sequence, WordPress deployment, price scraping—it’s all part of our pipeline. We don’t sell anything we don’t use ourselves.
KG Electronics · Boiler Manufacturer
12 000
The controller sends data every 30 seconds, the backend aggregates, service alerts go to the administration panel. Classic IoT pipeline = automatic layer over hardware. 12,000 connected devices, operational for 18 months.
View the full case studyDigital quoting · related service
3 hours → 4 minutes
In Akpil's case, the offer required 3 hours of the salesperson + approval from the boss. After automation - 4 minutes from clicking in the PDF configurator in the customer's inbox. It's not just UI - it's the generation, price mapping and shipping workflow.
View digital biddingERP/Customer Relationship Management Integrations · Related Service
5 → 1
Most of our workflows are based on the integration of 2-5 systems (Customer Relationship Management + ERP + mail + accounting + e-commerce). If the main problem is data synchronization between them - start with ERP/Customer Relationship Management integration, automation will go a layer higher.
See integrationsTwo paths, two different company profiles.
Make/Zapier (no-code SaaS)
n8n self-hosted/custom Python
Make and Zapier are sufficient for simple pipelines. Custom solutions are needed when SaaS stops scaling, Polish APIs are non-standard, or data cannot leave the EU.
What happens after implementation, and who we are.
After implementation
Optional packages. We will discuss the details during a meeting after implementation.
See also
30 seconds on JSON Crew
A software company specializing in digital transformation for B2B sales. Product configurators, dashboards, and sales process automation. Three key case studies (Akpil, Forest, and the internal JSON Hub platform) and portfolio implementations (KG Electronics, Metal Roofing, a heat recovery center, and modular homes).
Custom pricing based on the process map. Fixed price once the scope is defined; 30/30/30/10 timeline. We price single-step automated workflows (e.g., just a follow-up after a form submission) separately—faster and cheaper than a full pipeline.
Make and Zapier are sufficient for simple, low-volume pipelines. We choose n8n self-hosted (or custom Python) when: volume increases significantly, the integration is non-standard (Polish APIs, ERP systems), or the data is sensitive (GDPR) and cannot leave the EU. We make our recommendation after the first phase—we don’t push for custom solutions where SaaS is sufficient.
Three options: (1) Your own infrastructure (AWS, your own server), (2) third-party hosting (recommended — VPS running n8n, approx. 50–20PLN 0/month), (3) we host it as part of a maintenance package. The decision is made after the initial discovery phase — it depends on GDPR requirements and infrastructure preferences.
Yes—n8n has a visual interface, and we provide documentation in Polish. Your accountant or marketing specialist can make simple changes (such as adding a new step to a follow-up) without needing a developer. For larger architectural changes, either your developer or our support team can assist on request.
Self-hosted n8n on your server = your data never leaves your infrastructure. Logs are anonymized after 30 days (configurable). A data processing agreement is signed before launch. For regulated industries (healthcare, finance), we recommend a custom Python solution with self-hosting—full control.
Every workflow includes monitoring (health checks or custom alerts to Slack/email). If an integration fails, you’ll get an alert within 5 minuteses—not two days later when the client calls. Minor maintenance fixes and major API changes (e.g., a new version of Allegro) are estimated separately.
Please describe the process you want to automate and which systems it will connect. We will respond within one business day.
