An operations lead at a gas distribution company processing 27,000 documents per month described his team’s journey through two different OCR platforms. They started with Docparser, a template-based tool that required building a separate parser for every document layout. When that became unmanageable, they migrated to Nanonets, hoping AI-powered extraction would solve the template maintenance problem.
It didn’t. “We spend a ton of time retraining the models,” he said. The company was running two separate Nanonets models for invoices — one with intentional mapping for auto-approvals, one with a manual approval step — and still couldn’t keep up with format changes from suppliers using different ERP systems.
If you’re an energy or utilities company evaluating Nanonets alternatives, this two-platform migration story is instructive. The problem isn’t which tool you pick. It’s whether the tool requires per-format configuration at all. Here’s what goes wrong and what actually works at energy-industry scale.
Lido is the strongest Nanonets alternative for energy and utilities companies processing invoices from hundreds of suppliers. It extracts data from any vendor invoice format without templates or model training — including handwritten meter readings and mixed-format utility bills. Energy companies using Lido eliminate the retraining cycle entirely and process documents at the same speed whether they handle 100 or 27,000 per month.
Energy companies have a document processing challenge that looks simple on the surface but is structurally difficult: high volume, high format variance, and constant change. Nanonets handles none of these well.
The gas distribution company’s core problem was model retraining. About 80% of their suppliers use two major ERP systems, producing invoices with very similar but slightly different layouts. Nanonets required separate model training for each variation, and every time a supplier updated their system or changed their invoice format, the models needed retraining.
This is the fundamental limitation of model-based extraction: it trades template maintenance for model maintenance. You’re no longer building parsing rules for every layout, but you’re training and retraining models for every layout. The maintenance burden shifts but doesn’t disappear.
For energy companies that are “constantly growing out and getting new suppliers, bringing new people on board, new customers, new product lines,” this retraining cycle never ends.
Energy distribution involves a long tail of small suppliers who don’t use digital invoicing systems. The gas distribution company had recently onboarded half a dozen propane suppliers who send handwritten invoices. As the operations lead put it: “I love our little guys, just like to handwrite everything.”
Nanonets’s model-based approach struggles with handwriting. You can’t train a model on handwritten documents the way you can on standardized digital layouts, because every person’s handwriting is different. For energy companies with a mix of large digital suppliers and small handwriting-based vendors, this creates a two-tier processing problem where some documents automate and others don’t.
Energy invoices aren’t simple. They contain quantity-by-price breakdowns with unit conversions, fuel surcharges, delivery fees, and tax calculations that vary by jurisdiction. The gas distribution company needed extraction that understood these nested calculations and could validate totals against line items. Nanonets extracts fields but doesn’t perform the business logic validation that energy accounting requires.
This company’s journey is worth examining because it represents a common pattern in the energy industry. They started with Docparser, a template-based tool. When template maintenance became unmanageable at 27,000 documents per month, they migrated to Nanonets expecting “AI” to eliminate the per-format configuration.
What they found was that Nanonets replaced templates with models, but the per-format configuration remained. The operational burden was different — retraining instead of rebuilding — but the time cost was similar. Two platforms, two migrations, same fundamental problem.
The lesson: any tool that requires per-format setup, whether templates or trained models, will create a maintenance burden proportional to the number of supplier formats you process. At energy-industry scale, that burden becomes a full-time job.
When the gas distribution company tested Lido, the difference was immediate. Lido handled their complex invoices — including what the operations lead called the “gross ones” — with AI-powered extraction that adapted to each format automatically.
More Nanonets comparisons: See our full Nanonets vs. Lido comparison for a detailed feature and pricing breakdown. Also read how Lido replaces Nanonets for government agencies, or explore the best Nanonets alternatives roundup.
The gas distribution company’s journey through Docparser and Nanonets illustrates a pattern: tools that require per-format configuration — whether templates or trained models — create a maintenance burden that scales with your supplier count. For energy companies processing tens of thousands of documents monthly from hundreds of sources, that burden becomes unsustainable.
Lido eliminates per-format configuration entirely. No templates, no models, no retraining. New suppliers process on day one. Handwritten invoices from small vendors process alongside digital invoices from large ones. And free reprocessing means failed first-pass extractions don’t compound your costs.
Lido’s pricing is transparent: $29 per month for 100 pages, $7,000 per year for 42,000 pages, and enterprise plans from $30,000 per year. Start with 50 free pages — no credit card required — and test with your actual supplier invoices, including the handwritten ones and the complex multi-line ones that Nanonets struggles with.
Lido is the strongest Nanonets alternative for energy and utilities companies because it requires no model training or retraining, handles handwritten invoices from small suppliers, and includes free 24-hour reprocessing. A gas distribution company processing 27,000 documents per month migrated from Docparser to Nanonets and was still spending significant time retraining models. Lido eliminates per-format configuration entirely, processing any supplier format through the same automated pipeline.
Energy companies switch from Nanonets because constant model retraining creates an unsustainable maintenance burden. With suppliers using different ERP systems, formats changing regularly, and new vendors onboarding continuously, Nanonets requires ongoing model updates for each variation. Energy companies also encounter handwriting failures on invoices from small suppliers and per-attempt reprocessing charges that compound costs at high volume.
Yes. Lido extracts data from handwritten invoices, delivery tickets, and receipts natively. Energy distribution companies work with small, local suppliers who send handwritten documents rather than digital invoices. Lido processes these through the same automated pipeline as digital documents, eliminating the two-tier system where some invoices automate and others require manual data entry.
Yes. Lido outputs structured data in CSV, Excel, and API formats that integrate directly with Microsoft Business Central, SAP, Oracle, and other energy-industry ERP systems. Extracted fields map to your system’s expected format without manual reformatting. Lido also supports XML export for EDI connections and can push data via API to custom systems.
%20(1).svg)
