Real Simplit projects: each covers the problem, the solution and the outcome.
Our own companies
IT consulting and IT services
Problem
Projects, messages, meetings, invoices and contracts all came down to one person. Promises and deadlines had to be remembered, while assistants started every task without context.
Solution
We built a shared database of projects and tasks, with assistants connected to it, each with different roles and rules: where to get facts, where to record results and what to ask a person. In the morning, the system prepares a daily plan; in the evening, it checks the changes.
Outcome
The system prepares invoices, contracts, proposals and meeting minutes; Nikita checks them and makes decisions. Promises and deadlines are visible in the daily plan.
Client numbers were growing, while manual document entry, emails and calls took up a significant share of the accountants’ time. Online shops brought hundreds of invoices and payouts through several payment systems.
Solution
We described how the firm works and brought clients, documents, deadlines and correspondence together in one place. Of six possible assistants, we started with two that together cover the full cycle: receiving and entering documents, and handling email and client correspondence. We kept the accounting software: the team knows it.
Outcome
According to the project calculations, six AI assistants will free up around 250 working hours a month. The team will be able to use this time to serve new clients. Implementation is phased, and the project is currently in progress.
Projects and tasks were scattered across spreadsheets and correspondence. Income and expenses for each project had to be compiled manually from the accounts.
Solution
A shared project database in Notion: incoming documents are linked to projects, and each project manager has their own workspace. We chose Notion because the team is small and needs a simple solution they can learn quickly. We provided training and put the system into use.
Outcome
Companies, projects, tasks and money are brought together in one system instead of spreadsheets. Profit on each project is visible straight away, rather than at the end of the year.
Everyone used their own ChatGPT in their own way, or did not use it at all. The manager saw no benefit and did not know which company data was going into personal chats.
Solution
We identified participants’ tasks, prepared company accounts and rules, provided training with ChatGPT and NotebookLM using work documents, and ran a workshop where everyone built an assistant for their own task.
Outcome
AI became a working tool for the team, with common rules. Participants have assistants for their own tasks, the administrator has instructions, and the manager can see where AI is already useful.
A single transaction passed through several programs, while stock was tracked in Excel without a movement history. Stock figures did not match, reports were compiled manually, and some of the hundreds of formulas in the management spreadsheet were broken.
Solution
We started with organised data rather than AI: we described 12 processes and mapped the data and risks. Using this map, we replaced disconnected programs with one business system — Odoo — and added four AI assistants on top.
Outcome
Stock figures match and are visible straight away, reports are compiled without manual work, and AI assistants rely on accurate data.
Orders, purchasing, manufacturing and stock were kept in separate spreadsheets and in people’s heads. The accounting software did not show the whole operation.
Solution
We reviewed ten real working situations — from an order cancelled after assembly had begun to a shortage of parts in stock. We chose a shared system for orders, manufacturing, stock and logistics. We decided to keep the accounts in the familiar software. The documented processes will form the basis for connecting AI assistants.
Outcome
In the chosen model, the entire order journey — from enquiry to dispatch — is brought together in one place. Every operation has a responsible person and a history of completion. The manager can more easily see where work has stopped and what needs doing next.
A couple of sentences is enough. I will reply personally and suggest a time to meet. You can pick a time in the calendar right away or message us on WhatsApp or Telegram.
