Skip to content
Sales & quotingElectronics & SensorsMachinery & Plant EngineeringMetal & MachiningPlastics & Packaging+3

Early warning: a key account is getting unhappy

AI turns complaints, delivery performance, payments, response times and e-mail tone into one early-warning score per key account.

Today

Key accounts rarely leave out of the blue. Months before, complaints pile up, deliveries arrive later, payments slip, e-mails get shorter and suddenly the managing director is in copy. Each signal sits in a different system and a different department, so nobody puts them together. Sales finds out when the customer is already putting the business out to tender again.

An example

Nordhanse looks after a container terminal running 14 reach stackers RS 45 t, whose service contract ends in March. Within eight weeks come three complaints about twistlocks on the container spreaders, on-time delivery of spare parts drops from 96 to 81 %, two invoices are paid late for the first time, and the terminal's technical manager, who used to call every week, now sends only curt e-mails. On their own, none of these signals stands out. demi turns the customer red, shows every signal with its source and proposes a joint visit by the heads of sales and service with a spare-parts plan for the twistlocks, four months before the renewal.
Fictional company, typical workflow.

Your data, wherever it lives

The signs of how satisfied a key account is are scattered today: complaints in the QM system, delivery dates and payments in the ERP, service cases in the ticket system, visits and notes in the CRM, the tone in mailboxes and call notes. demi reads, understands and links them per customer, with your rights. The AI sees weeks earlier that something is tipping, and you decide how to respond.

Where the data lives today

  • CRM
  • ERP
  • QM system
  • E-mails
  • Tickets
  • Phone & voice
Data LayerOne view of everything, with your rights

How demi solves it

Three layers in this order: connect the data first, then the rights, then the AI.

  1. 1 · ConnectData Layer

    Data sources in this case: CRM, ERP, QM system, e-mail, ticket system and call notes. demi replaces nothing and joins complaints, delivery dates, spare-part orders, payments, service cases, e-mails and call notes into one record per customer. Next to every key account you see what has changed in recent weeks.

  2. 2 · PermitPermission Layer

    The early-warning score and its signals are visible to whoever looks after the customer, plus sales and service management. Payment data appears only for roles that see it in the ERP. demi reads only mailboxes and folders released for customer correspondence, and it rates the customer, never the work of your staff.

  3. 3 · ActIntelligence Layer

    The agent learns what is normal for each customer: order rhythm, complaints, payment behaviour, response times and tone. When something deviates, it weighs the signals into an early-warning score and shows each with its source. Figures are calculated from system data; the language model only rates the tone of e-mails and summarises.

How it runs

  1. 1

    demi keeps track of complaints, delivery dates, payments, service cases and e-mail traffic for each key account.

  2. 2

    The agent compares every signal with what is normal for this customer.

  3. 3

    When deviations pile up, the score rises and you get an alert with all signals and sources.

  4. 4

    The agent proposes countermeasures, such as a visit, a call from service management or a delivery plan for critical parts.

  5. 5

    You decide what happens; tasks and outcome go into the CRM and feed into the next rating.

What changes

  • Dissatisfaction shows while you can still turn it around

  • Sales, service and quality see the same picture of the customer

  • Fewer surprise cancellations and re-tenders

What does it bring you?

Check it with your own numbers

Every starting value is an assumption you can change. The calculation sits right below, step by step.

Assumption, editable

Your numbers

Result

+144 hfreed per year

Assumptions and calculation

demi gathers the warning signs per key account from every system; you decide how to respond.

60 min → 42 min per case

How does it run today?

What is already digital saves less: the factor lowers the share demi brings.

How we calculate

40 key account reviews × 60 min × 30% × 1.0 (manual) × 12 months = 8,640 min = 144 h

Value per year, before demi costs

Freed per year+144 h

Assumption: labour cost of €49.50 per hour in German manufacturing in 2025 (Destatis), plus workplace, IT and overheads.

= Value per year, before demi costs

+€9,360

gained per year

144 h ÷ 1,760 h per role = 0.1 full-time roles

An estimate, not a promise. The starting values are assumptions for an industrial company with 150 to 500 people.

Combine with more cases and get the package

Questions about this use case

Which signals go into the early-warning score?

The ones you define, typically: complaints and how long they stay open, on-time delivery, payment behaviour, order volumes, open service cases, response times and the tone of e-mails. Each signal is compared with what is normal for this customer. You and your sales team set how much each one counts.

Does demi read every e-mail for this?

No, only mailboxes and folders released for customer correspondence, such as the sales mailbox or e-mails filed in the CRM. It rates the customer's tone, not the work of your staff. Private mailboxes stay out.

What if the early warning goes off wrongly?

Then you see straight away why: every warning shows its signals with sources. You mark it as handled or as a false alarm, and the agent takes that into account next time. A person always decides what happens with the customer.

Does it work for customers who rarely order?

Yes, then other signals count more: service cases, complaints, reactions to quotes, the tone of e-mails and how often there is any contact at all. For project customers the agent also watches open points from the last project and upcoming tenders.

See this case with your data

In the demo we show you the workflow on your own systems.

Your configurator

No use cases yet

Your configurator is still empty

Add a use case from any card or case page and see what it brings per year, with the calculation. Free, no sign-up, no commitment: try it out and change any number.

Browse use cases