AI-led operations products

The advisory pillar covers what good looks like in operational management. The products on this page are the AI-led implementations that make the back office viable at a small team size.

Product 5: Switch

AI validation and exception handling on the SIPS data exchange with the eight Spanish DSOs and Portuguese e-Redes.

The retailer's net position growth depends on switching working cleanly. Most retailers lose 1 to 3 percent of their monthly switching volume to technical rejections that nobody recovers. Switch is the validation and exception-handling layer that runs against each request before submission and works the exception backlog after.

What it does:

• Pre-submission validation against the latest DSO data: CUPS, postal code, tariff eligibility, P1-P6 period classification
• Retry logic for transient DSO failures with escalation for persistent ones
• Exception classification across the common rejection reasons: CUPS mismatch, postal code dispute, tariff code mismatch, period misclassification, cancellation race conditions
• Customer-facing notification: rejected switches surface to the customer service team within 48 hours, not on the next bill from the previous supplier
• Monthly switching analytics by DSO: gross add, gross loss, net position, time-in-switch, exception rate
• Reconciliation hook into the Reconcile product so the switching activity reconciles against the ATR settlement invoices

What it integrates with:

• The eight Spanish DSO SIPS endpoints (e-distribución, i-DE, UFD, Viesgo, Hidrocantábrico and the regional DSOs)
• Portuguese e-Redes
• The retailer's CRM
• The retailer's customer service workflow

Engagement: 8 to 12 weeks build. Strongly recommended managed-service follow-on; the SIPS formats are documented but not perfectly stable across the DSOs, so ongoing maintenance is real.

Best for: Retailers operating nationally across multiple DSOs, particularly those with monthly switching volume above 1,000 supply points.

Related reading: Switching and ATR with the Iberian DSOs.

Product 6: Office

AI-augmented back-office covering invoice generation, direct debit dunning, bad-debt management and CRM hygiene.

The back-office is the unsexy machinery that determines the cost-to-serve. Office is the AI augmentation that moves a typical retailer from €30 to €40 per customer per year to €15 to €25, with materially better operational discipline.

What it does:

• Invoice generation quality checks before dispatch: format compliance, calculation reconciliation, channel preference enforcement
• Direct debit failure handling: automated dunning cycle, escalation to customer service for resolution calls, defined write-off thresholds
• Bad-debt risk segmentation and collection workflow, with collection-agency hand-off for the right customers
• CRM hygiene: duplicate detection, address-change propagation to the DSO via the switching workflow, ownership transfer handling
• Document management: contract versioning, terms-of- service retrieval, marketing claims audit trail

What it integrates with:

• The retailer's billing engine
• The retailer's CRM
• SEPA direct debit infrastructure
• The Switch product (or any equivalent switching workflow)
• The retailer's customer service system

Engagement: Programme of 6 to 12 months, in phases. Each phase covers a defined workstream (invoice generation, then direct debit, then bad debt, then CRM hygiene). Phased delivery so the retailer sees operating-cost improvement at each milestone.

Best for: Retailers with cost-to-serve materially above €25 per residential customer or €80 per SME supply point per year. Also retailers with bad-debt write-offs above 2 percent of revenue.

Related reading: Back-office operations for Iberian retailers.

Product 7: Serve

Customer service co-pilot tuned to the three Iberian retail customer populations: residential PVPC and free-market, SME bilateral, corporate structured.

Customer service is operationally fragile in most retailers: high volume, low value per call, tight CNMC complaint-handling timeframes, three distinct customer populations with different needs. Serve is the agent layer that handles the routine queries directly and supports the human agent on the queries that need escalation.

What it does:

• Self-service for the routine queries (bill explanation, consumption history, switching status, tariff comparison)
• Triage and routing for queries that need human service, with proper segmentation by population
• Human-agent co-pilot: surfaces the customer history, the open issues, the relevant tariff terms, the regulatory complaint window
• Complaint root-cause feedback into the back office: recurring complaint patterns surface to operations for systemic fixes rather than being handled one at a time
• CNMC complaint pipeline management: escalations routed through the right path with the SLA tracking the regulator expects

What it integrates with:

• The retailer's CRM
• The retailer's billing engine
• The retailer's complaint-handling system
• The Switch product (or any equivalent switching workflow)
• The Office product (or any equivalent back-office)

Engagement: 8 to 16 weeks build. Optional managed-service operating-model takeover for retailers that prefer to outsource the function.

Best for: Retailers with high residential volume (above 30,000 customers), or retailers with material SME or corporate customer bases where the cost-to- serve has to absorb more sophisticated service delivery.

Related reading: Customer service operations in Iberian energy retail.

Product 8: File

Regulatory filing assistant covering the MITECO, CNMC, REE and OMIE submission calendar, plus the DGEG and ERSE equivalents.

The retail licence depends on the regulatory filings landing on time and in the right form. File is the agent layer that runs the filing calendar, prepares the submissions from canonical data sources, validates them against the regulator's published rules, and maintains the audit trail.

What it does:

• Filing calendar covering all mandatory submissions with named accountabilities, target delivery dates and upstream data dependencies
• Submission preparation from canonical data, with automated validation against the previous submission and the relevant financial statements
• Pre-submission review workflow with sign-off by the regulatory function and the finance function
• Versioned audit trail preserving the submission, the underlying data, and the methodology applied
• Exception handling for the cases where a filing cannot be made on time, with documented escalation paths
• Sanction-response playbook for the cases where a sanction notice arrives

What it integrates with:

• MITECO submission portals
• CNMC submission portals
• REE and OMIE settlement and reporting interfaces
• DGEG and ERSE for Portugal
• The retailer's billing engine and CRM (data sources)
• The retailer's finance close (data dependency)

Engagement: 6 to 10 weeks build. Strongly recommended ongoing managed-service support given the calendar weight and the sanction exposure.

Best for: Every retailer with an active Iberian licence. Sanction exposure is universal; the discipline of getting filings right pays back across the licence lifecycle.

Related reading: Submitting regulatory reports in Iberian energy retail.

Where these products sit together

The four operations products are designed to be implemented together but each can run standalone. The most common sequence:

1. File first. Regulatory exposure is the largest immediate risk and the filing assistant pays back quickly.
2. Switch next. Net position growth depends on this; the operational analytics feed the rest of the suite.
3. Office third. Back-office cost-to-serve improvement, phased over 6 to 12 months.
4. Serve last. Builds on the three above; the customer service function needs the other three to be solid before it can be augmented usefully.

Related

• Operational management advisory pillar (what good operations look like)
• AI-led purchasing products
• AI-led analytics products
• Products hub
• Cyber guardrails for AI agents in regulated workflows