Place des Énergies: Energy Bills Sorted by Client

Founded in 2013 and based in Saultain, in the Hauts-de-France region, Place des Énergies is a broker that negotiates and audits gas and electricity contracts for several hundred corporate clients. Every invoice received from a supplier must be verified and then filed under the correct client, even though nothing in the document itself indicates this classification. Place des Énergies currently manages several hundred corporate clients, each typically having around ten consumption sites, representing a continuous flow of documents that Koncile's data extraction platform collects, reads, and verifies.

The True Role of a Broker: Auditing Supplier Invoices

An energy broker does not simply negotiate a rate at the time of signing. Their added value comes into play afterwards, invoice by invoice: checking that the contracted capacity (subscribed power) corresponds to the billed power, identifying abnormally high consumption, recalculating consumption site by site based on meter readings, or ensuring that no pricing errors have slipped into line items. It is this auditing process, repeated across an entire client portfolio, that justifies the broker's role: without it, a supplier billing error remains at the client company's expense, often without them even realizing it.

The Problem: Invoices Designed by Supplier, to Be Filed by Client

A gas or electricity bill is a document constructed from the supplier’s perspective: it describes a site, a contract, and a consumption period, without ever specifying under which final client of the broker it should be filed. However, the same supplier often bills multiple clients in Place des Énergies' portfolio, and a single client may have multiple sites, multiple suppliers, and separate gas and electricity bills. The document's structure (by supplier, by site, by contract) is therefore the exact opposite of what the team needs for its tracking (by client).

I need to visualize distinct folders per client because the list of invoices can quickly become overwhelming for clients for whom we check dozens of invoices per month.
^{Laetitia Leclercq, Project Manager, Place des Énergies}

Without this organization, it is impossible to see at a glance which client has submitted their monthly invoices and which one needs a follow-up.

Two Sources of Ambiguity to Resolve for Reliable Classification

Implementing automated classification by client ran into two practical challenges rather than a simple document reading issue.

One Supplier, Multiple Clients to Avoid Confusion

In initial tests, a gas invoice ended up filed in the folder of a different client than the one it actually belonged to, as the extraction model primarily recognized the issuing supplier rather than the target final client. Explicit classification rules had to be added to each client folder (specifying which information on the invoice must be searched to identify the correct client, and when a document should be excluded from one folder over another) to make automated filing reliable, even for infrequent suppliers in the portfolio.

Two Import Streams That Must Not Overlap

Daily billing audits rely on email imports: invoices arrive continuously to be sorted and checked as they come in. A second, more industrialized use case relies on API imports to route each document directly to the correct folder and extraction model. Both streams rely on very similar models (a gas bill remains a gas bill regardless of the channel), creating a risk that documents arriving by email might be classified automatically as if they came from the API stream, or vice-versa. The configuration therefore had to include explicit rules at the folder level so that each channel remains isolated, even when identical models are used.

The Technical Solution Implemented with Koncile

Rather than multiplying models by supplier, Place des Énergies and Koncile built two generic models—Gas Invoice and Electricity Invoice—which were then duplicated for each client in the portfolio. In practice, an OCR engine capable of reading widely varying layouts from one supplier to another feeds a line-by-line configurable data extraction system: prices, quantities, VAT rates, subscribed power, and peak demand are reconstructed invoice after invoice, regardless of the originating supplier. Each client folder then carries its own classification rules, added as edge cases arose, so the tool knows where to file an ambiguous document rather than categorizing it at random.

The value of having a folder per client is being able to quickly spot which client hasn't sent us their invoices, so we can follow up with them right away.
^{Laetitia Leclercq, Project Manager, Place des Énergies}

It is this tracking logic, rather than document parsing alone, that guided the setup. Today, two import channels coexist to adapt to each user's workflow:

Who Does What Between Koncile and Place des Énergies

The division of responsibilities followed a simple logic: Koncile handles technical setup (duplicating models across the entire portfolio, configuring classification rules, and API support), while Place des Énergies defines the required business rules (how a client folder should be structured, which line items to verify) and performs the daily billing audit. This split between technical configuration and business validation is found in other high-volume document management sectors, such as at Toyota Assurances, where subscription files arrive already entered and checked.

Key outcome: Scalable control across the entire portfolio

The main result isn't just a number, but a fundamental shift in how the work is done: billing checks, which previously relied heavily on manual sorting, have now been structured around customer-based categorization across the entire portfolio. The emergence of a second, more automated use case, alongside routine email checks, is itself a clear sign that adoption has gone well beyond a simple trial.

Frankly, I can't live without it anymore.
^{Laetitia Leclercq, Project Manager, Place des Énergies}

Next Step: Pushing Automated Analysis Further

Beyond filing, Place des Énergies now possesses structured, line-by-line data for every single one of its energy invoices. This opens the door to advancing automated analysis, an objective envisioned since the initial discussions: detecting excessive consumption, identifying meter index anomalies, and performing systematic site-by-site recalculations across the entire portfolio rather than folder by folder.

Frequently Asked Questions