How it works: the journey of one order through a company

For example, a manufacturing company receives enquiries by email in every imaginable form – a spreadsheet one day, two sentences and a photo the next. The agent extracts the items, quantities and deadline from each one, adds the customer's history from the CRM and prepares a draft quote for the sales rep. The rep simply reviews it, adjusts it and sends it.

For example, a shared support inbox receives dozens of messages a day – order status questions, complaints, invoice requests. The agent categorises each one, answers order status questions itself from the data in your system (provided the customer writes from the email address on the order) and logs each complaint as a ticket with a suggested reply. A member of your support team simply reviews the draft and sends it.

For example, an online store with several thousand products receives price lists and product descriptions from suppliers in all sorts of formats. The agent extracts the price and availability changes, prepares product copy based on your template and submits everything for approval. The store manager confirms the changes in bulk instead of editing every item by hand.

For example, an accounting inbox receives hundreds of invoices a month – PDFs, scans and photos taken on a phone. The agent extracts the data from each one, checks the VAT number in VIES and the bank account against the tax authority's register, matches the invoice to its purchase order and prepares the accounting entry. The accountant approves the prepared documents instead of retyping them.

For example, a private practice receives dozens of emails, web form messages and text messages a day asking for an appointment, a prescription or a certificate. The agent sorts the requests, books routine appointments directly in the practice's scheduling calendar, prepares prescription requests for the doctor as a clear list with supporting details, and immediately hands anything that sounds like a health concern to the nurse. Patient data is processed only under a data processing agreement, within the EU or on the practice's own infrastructure, and is never used to train models.

For example, a small freight forwarder receives transport enquiries by email in dozens of different formats. The agent extracts the route, date, type and quantity of goods from each enquiry, checks them against the price list and prepares a draft quote. The dispatcher reviews the quote and sends it – instead of reading and retyping every email by hand.

For example, an engineering company receives dozens of delivery notes and material certificates a day – on paper, by email and through supplier portals. The agent matches each document to its purchase order, compares quantities, batches and certificates, and books clean receipts straight into the ERP. If something does not add up – a missing certificate, a different batch, a quantity mismatch – it puts the receipt on hold and alerts the buyer.

For example, a mid-sized company receives dozens of CVs in various formats for a single job ad. The agent reads each CV, extracts experience, education and how the candidate meets the requirements into a single table, prepares a short summary and confirms receipt to the candidates. The recruiter starts the selection with a clear overview instead of a folder full of PDFs – and decides for themselves who to invite to interview; the agent never rejects anyone automatically.

For example, a property management company with several hundred rental units receives dozens of emails from tenants a day – from reported faults to questions about service charge statements. The agent sorts every message, logs faults in the property management system and suggests a technician, and answers routine statement questions from the records. The property manager is left only with the cases that require a decision.

For example, a service company with a dozen field technicians receives fault reports by email, web form and text message, and the dispatcher would otherwise retype them into a scheduling spreadsheet by hand. The agent takes in every report, asks for any missing details, creates the job and suggests a technician based on location and qualifications. The dispatcher simply reviews and confirms the daily schedule.

For example, the marketing team at a manufacturing company gets technical updates from the engineering department every week that need to be turned into copy for the website, newsletter and social media, and every Monday it compiles a report from the ad accounts. The agent prepares drafts for all channels from a single source document in the brand's tone of voice and compiles the report itself. The team only edits and approves.

For example, a trading company receives contracts and amendments from suppliers in dozens of formats, and the lawyer would have to read every one in full before signing. The agent compares each draft with the company template, highlights deviations in payment terms, liability or notice periods and prepares a summary. The lawyer knows in advance where the contract departs from the template and decides on signing.