When your AI agent negotiates with the local utility: How agent-to-agent communication is transforming energy services

From chatting to taking action: How customer service is changing right now

You’re on your way to work, headphones in, first coffee in hand. Whilst you’re getting ready for your day, something is happening in the background that you would previously have had to do yourself: your electricity contract is being reviewed and optimised.

Not by you. But by your digital agent.

What sounds like a pipe dream describes quite accurately the direction in which customer interactions are currently developing: away from manual processes – towards automated, yet controlled actions carried out by specialised AI agents.

The challenge: Complex processes, minimal interaction

Energy suppliers face several structural challenges:

• High process complexity (tariffs, regulatory requirements, individual consumption patterns)
• Low customer interaction outside of billing or faults
• High customer service workload (hotlines, manual enquiries, media breaks)
• Increasing competition from new providers and dynamic tariffs

At the same time, customers today expect:

• Proactive services
• Transparency regarding their contracts
• Simple and quick customisation options
• Digital, seamless interaction

The traditional response to this – apps, customer portals or chatbots – is increasingly reaching its limits. This is because they remain reactive.

The solution: Agent-to-Agent communication

A possible next step is direct interaction between two specialised systems:

• A customer agent (I Agent) that acts in the user’s best interests
• A company agent (E Agent) representing the energy supplier’s systems

These agents communicate directly with one another – based on clearly defined rules, interfaces and authorisations.

A possible scenario

To illustrate, here is a greatly simplified example:

• The customer agent identifies potential for optimisation in the electricity contract
• It enquires with the energy supplier’s agent about available options
• The enterprise agent checks tariffs, bundles or additional services
• The proposals are submitted to the customer for approval
• Implementation only takes place after approval

The crucial point: It’s not just about communication – it’s about taking action. But always under the user’s control.

Technological classification: What is an “agent” really?

The classification is important: An agent is not an autonomous “black box” system. The underlying structure is relatively straightforward:

LLM + Tools + clearly defined task = Agent

• LLM (Large Language Model): Understands and generates content
• Tools / APIs: Access backend systems such as CRM, billing or contract management
• Orchestration: Controls processes, decisions and interactions

In the energy sector in particular, this means:

• Integration into existing system landscapes
• Use of existing interfaces
• Extension via an orchestration layer for agent logic

Governance as a success factor: Why control is crucial

A key aspect in the introduction of such agents is trust.

And trust is only built through clear rules.

Therefore: Governance by design rather than governance as an afterthought.

This includes, amongst other things:

• Clear authorisation models (read vs. write)
• Earmarked data usage
• Full logging of all actions
• Explicit approvals for every change
• Transparency for users

Particularly in the energy sector – with sensitive contract and consumption data – this is not a ‘nice-to-have’ but a prerequisite.

adesso as your implementation partner

The implementation of such agent scenarios is not a theoretical vision of the future, but can already be realised step by step today.

Typical building blocks in the adesso context are:

• System integration: Connection of CRM, billing and service systems
• API management: Provision of standardised interfaces
• Orchestration platforms: Control of agent processes
• GenAI components: Integration of language models into existing processes
• Governance frameworks: Implementation of security and compliance requirements

adesso supports energy suppliers in combining precisely these building blocks in a meaningful way – and in gradually evolving existing processes towards agent-based interaction.

Opportunities and limitations

Advantages

• Significant reduction in customer service workload
• Proactive customer engagement instead of reactive handling
• Higher customer satisfaction through convenience and transparency
• More efficient use of existing data and systems

Challenges

• Integration into existing IT landscapes
• Ensuring data protection and compliance
• Customer acceptance
• Clear definition of responsibilities between humans and systems

Agents are not a sure-fire success. Their added value only arises when they are meaningfully embedded in processes.

Conclusion: From demo case to real impact

Agent-to-agent communication has the potential to fundamentally transform customer service in the energy sector.

Key takeaways:

• Automation does not replace customers’ decision-making authority
• Added value comes from integration into real business processes – not from isolated applications
• Governance and transparency are key success factors

Or to put it another way: those who simply let agents communicate end up with interesting demos.

Those who let them act under control create real added value.