Octopus Energy Automates 40% of B2B Emails via Kraken AI Platform

Octopus Energy announced this week that its Kraken platform now automates the bulk of its B2B supplier communications using generative AI. For UK SMEs, this proves that AI is moving out of customer service chatbots and into the messy reality of supply chain automation. The system is already handling 40% of inbound emails from vendors and partners without human intervention.

Octopus Energy automates 40% of B2B emails

Octopus Energy has integrated generative AI into its Kraken platform to manage vendor management and procurement queries, according to Kraken's latest release. The energy supplier is now using large language models to read, categorise, and respond to inbound emails from its B2B supply chain.

As reported by the Financial Times, the AI system handles 40% of all customer and supplier emails. It drafts responses, updates internal databases, and flags exceptions for human review. The system doesn't just send generic replies. It extracts structured data from unstructured emails, such as invoice queries, delivery updates, and compliance documents, and matches them against Kraken's internal records.

The company states the AI performs the work of hundreds of staff members, allowing their human procurement team to focus on complex negotiations rather than routine inbox management. This marks a significant shift from using AI purely for consumer-facing support to deploying it deep within back-office operations.

The shift from customer service to back-office operations

This deployment changes how you should think about AI in your own operations. Until now, most businesses viewed AI as a tool for writing marketing copy or answering basic customer FAQs. I think this view is far too narrow. The Octopus Energy case shows that the real value lies in automating the messy, unstructured data flowing through your SME procurement channels.

When a supplier emails you about a delayed shipment or an invoice discrepancy, that message usually sits in an inbox waiting for a human to read it, check a spreadsheet, and type a reply. Kraken AI proves that language models are now reliable enough to handle this exact workflow. They can read the unstructured text, query your database, and draft the response.

For a UK SME with 50 or 100 employees, you don't need to build a proprietary platform like Kraken to get these benefits. Off-the-shelf tools are starting to offer similar capabilities for standard email clients and ERP systems. The second-order effect here is that your own suppliers will soon expect faster resolutions. If your competitors are automating their vendor management, they'll negotiate better terms and resolve supply chain bottlenecks days before you do.

Three steps to audit your supplier communications

You don't need a massive tech budget to start automating your back office. Take these steps this week:

1. Audit your shared inboxes. Look at your accounts payable or procurement email accounts. Identify the three most common types of routine queries your team handles from suppliers.
2. Test a pilot workflow. Use tools you already have, like Microsoft Copilot or Google Workspace AI, to draft responses to these common queries. See how accurately the AI extracts the necessary details from the emails.
3. Standardise your vendor data. AI systems need clean data to cross-reference against inbound emails. Ensure your core supplier details, contracts, and payment terms are stored in a central, accessible database rather than scattered across individual spreadsheets.