You have data.
Find the clarity.

01CRM, email and deals

The funnel as it is,
not as reported.

We’ll export the deal history and correspondence and rebuild the path of every request: stages, timing, reasons for loss. You’ll see where deals wait, where they get lost and what each stage costs.

What we take on

• We’ll export your CRM and email history read-only — nothing changes on your side.
• We’ll show the actual funnel: how many deals reached each stage and for what amount.
• We’ll find the bottleneck: the stage where deals wait longest.
• We’ll break down losses by the reasons managers enter when closing a deal.
• We’ll list stalled deals — untouched for a week or more — for each manager.
• We’ll link emails, quotes and drawings to deals, so the terms sit next to the number.

Example: in an estimating project’s CRM, the main queue turned out to be at the estimating stage, and most deals were lost after the quote had been sent.

Actual funnel · deals by stage

We start with read-only access. A CRM that is new to us is connected on a trial export first.

02Calls and voicemail

Calls into text.
Conversion, honestly.

We’ll transcribe call recordings and voicemail, sort calls by type and put a summary in the deal record. You’ll see how many calls are real requests and how many are suppliers and spam.

What we take on

• We’ll transcribe call recordings and voicemail: the text and a short summary go into the CRM record.
• We’ll sort calls by type: request, question about a deal, supplier offer, spam.
• We’ll measure channel conversion honestly: supplier calls are not mixed with requests.
• A voicemail becomes a separate lead with a transcript.

Example: in an estimating project, a noticeable share of “call leads” turned out to be supplier offers and spam, so call conversion could not be compared with email directly.

Call transcribed · summary in the deal record

We work with the recordings your phone system already keeps. We don’t set up PBX, numbers or outbound calling; transcription happens after the call, not in real time.

03Sales, customers and shipments

Ordered, accepted,
shipped — as it happened.

We’ll bring orders, acceptance and shipments into one picture: sales by customer, channel and period, the accepted share, unit price, delivery cost. No manual pivot tables.

What we take on

• We’ll combine orders from every channel — EDI, website, CRM, email: sales by customer, week and month.
• We’ll show the acceptance rate: how much was ordered, how much the customer accepted and where it diverged.
• We’ll calculate unit price by customer and the top items by order value.
• We’ll cost delivery: roubles per kilogram for each customer and each trip.
• We’ll compare the volume and value of every order — unusual orders stand out at once.
• We’ll show seasonality and trends from your sales history.

Example: orders and acceptance from national retail chains, shipments and transport are in the dashboard’s analytics section.

Sales by customer · by month

We calculate from data already in EDI, accounting and the shipment log; payments and accounting-system documents are connected as a separate step.

04Margin and cost

Margin per item —
down to the purchase row.

We’ll match sales with purchases for every item and calculate what you really earn, VAT included. For manufacturing — cost at your rates next to your specialist’s estimate.

What we take on

• We’ll calculate the actual margin: sale against purchase for every item and deal.
• We’ll put amounts on one basis: with and without VAT, suppliers on different rates.
• We’ll build supplier and contractor profiles from correspondence: prices, VAT, prepayment, minimum batch.
• We’ll build a price base from emails and quotes: every price has its source email and date.
• We’ll compare the system’s estimate with your specialist’s — differences show for every item.

Example: margin was calculated from sale and purchase rows, and a second check found a VAT error — the calculation and the report were fixed before any conclusions.

Sale − purchase = margin · by row

A person sets the price and chooses the supplier. We don’t promise accuracy in percent — we show the differences on your data.

05Reconciliations and cash

The money adds up —
or you see where it doesn’t.

We’ll bring together shipments, documents and payments: bank statements for each legal entity, reconciliations with customers, the carrier and the warehouse, debts by age. A discrepancy shows up as a row, not as a month-end total.

What we take on

• We’ll move bank statements from email into a transactions table for each legal entity.
• We’ll reconcile with customers and the carrier from the shipment log.
• We’ll calculate the warehouse fee by day — an unclosed day stands out at once.
• We’ll show customer debts by age: who, how much and since when.
• We’ll separate promised, shipped and paid — each with its own status.

Example: bank statements land in a shared table on their own, and reconciliations with the warehouse and customers are calculated from the shipment log.

Statement posted · transactions by legal entity

We don’t do bookkeeping or replace your accounting system: we show discrepancies, and the person in charge fixes them. Bank data is read-only.

06Tenders and procurement

Tenders and prices —
in numbers, not from memory.

We’ll gather any tenders and procurement into one database — selling on platforms and buying from suppliers: schedule, items, prices, results. Your base price and markup sit alongside, so you can see where bidding pays and where it doesn’t.

What we take on

• We’ll put tenders into a database: schedule, item specifications, the result of each procedure.
• We’ll show platform prices next to your base price and markup.
• We’ll collect the results: which items were won and at what price.
• We’ll compare supplier offers in your purchasing: price, lead time, terms — in one table.
• We’ll build a customer directory: how they are named in email and contracts, terms, distribution centres.
• We’ll sort tender email: which platforms write, where the terms are, where the claims are.

Example: retail tenders are gathered in the dashboard — schedule, items, results and base prices sit next to shipments and documents.

Tender window · items in the plan

We don’t write bots that place bids on their own. A person decides whether to bid and at what price; nobody can guarantee a win.

07Fields and farm

The field from a satellite —
the plan next to reality.

We’ll show your plots on a map: cadastral boundaries, area, vegetation condition from the season’s images. Alongside — the work and irrigation plan, to compare what was planned with what the field shows.

What we take on

• We’ll show plots on a map: cadastral boundaries, area, a satellite layer.
• We’ll build a vegetation index series from the season’s images — for each plot.
• We’ll lay out the work and irrigation plan by field block: zones, shifts, dates.
• We’ll simulate machinery work by block to estimate the load on the fleet.
• We’ll bring plan and actuals together in the farm dashboard: layers, devices, notes.

Example: the farm dashboard shows plots, satellite layers and the irrigation plan, and a plot check by cadastral number is open on the website.

Plots on the map · boundaries and area

The images are open satellite data, with clouds and delays: a person on site confirms any conclusion about the field. We don’t give agronomic advice or yield forecasts.

08Data quality

First we check the data.
Then the conclusions.

Before building a report, we check the data itself: duplicates, gaps, different spellings of one customer, amounts with and without VAT. Every number keeps its source and date; anything disputed goes to a person.

What we take on

• We’ll find duplicates and different spellings of one company and merge them into one record.
• We’ll enrich company records with open data: registration details from the tax service, court cases, enforcement proceedings.
• We’ll collect contacts from email signatures: clear ones at once, doubtful ones into a review queue.
• We’ll re-check conclusions against the originals before they reach a report.
• We’ll show how fresh the data is: what was updated when and where a source has fallen behind.

Example: retailer contacts are collected from email signatures and doubtful ones wait for a person’s decision; company records are enriched from open data every morning.

Two records of one company · merged into one

Open registers can be incomplete — disputed data is never fixed silently; it goes to the person in charge.

09Reports and dashboards

Answers for management —
in the dashboard and in the chat.

Findings live where people work, not in a slide deck: a management screen in the dashboard, an evening summary in your messenger, an alert when a metric crosses a limit. Every metric has a source and a date.

What we take on

• We’ll build a management screen: sales, shipments, documents and deadlines — by period and customer.
• We’ll send summaries to your messenger: daily results, manager metrics, stalled deals.
• We’ll set up alerts: the queue has grown, a customer accepted less than ordered, data hasn’t updated.
• We’ll gather metrics for any department: sales, purchasing, warehouse, finance, HR, service.
• We’ll show AI spending by project and operation.
• We’ll build a specialist’s screen: the system’s estimate next to theirs, with rates edited in the same place.

Example: a wholesale business, an estimating project and a farm share one dashboard with analytics sections, and evening summaries arrive in work chats.

Management screen · by period and customer

The dashboard is web-based and opens in a phone browser. We build dashboards on your data, not to a spec without access to it.

10Management questions

Your question —
its own route through the data.

Analytics starts with a question. Pick one, and the data that has to be connected to answer it lights up.

The routes are examples. We’ll go through your questions and data at the first meeting.

11Getting started

One question.
A good beginning.

Where is time being lost? What is happening to the margin? Which orders are stuck? Pick one important question — we’ll agree on the data, the timeline and a verifiable result.

1. Read-only access and one question
2. Audit: the picture as it is, checked against the originals
3. A report, monitoring or automation of the findings

Useful to have: an export or read-only access to the CRM, email or accounting, and a person who knows the process.

We agree access, confidentiality and the scope of work in advance.