May the Agentforce Be With You: AI in Energy Services
Generative AI has moved past being a shiny demo and into the messy reality of enterprise operations—where data lives in different systems, customers expect instant answers, and security teams (rightfully) say “prove it.” In energy services specifically, even small efficiency gains matter: many retail energy providers operate on thin margins, and operational blind spots—billing confusion, outage updates, fraud, or pricing risk. This can quickly erode profitability. As a result, the conversation around AI in energy has shifted from chatbots to agents: systems that don’t just answer questions, but can take action across tools, workflows, and roles.
So here’s the real question organizations are asking right now: How do you turn “AI hype” into secure, practical agent-driven work that improves customer experience and operations—without creating chaos, risk, or broken trust?
That’s what this episode of The CG Hour tackles. Hosted by Fanny Dunagan, the conversation features Mike Hatch, a Certified Principal Enterprise Architect at Salesforce, and Jonathan Goldstein, Senior Vice President at CG Infinity. Together, they break down what Agentforce is, how “agentic maturity” is evolving (from basic retrieval to multi-agent orchestration), and why energy services may be one of the clearest proving grounds for AI that actually delivers value.
Top insights from the talk…
-
Agentforce reframes the UI: from clicking screens to “conversing with systems.” The panel explains how agents sit on top of existing Salesforce capabilities (sales, service, industry clouds) and make them accessible through chat—and increasingly voice—without needing a traditional interface.
-
Energy use cases get concrete fast: billing, outages, field service, and clean-energy programs. The episode highlights how agents can answer “why is my bill so high?” using contextual usage patterns and external factors like weather, proactively manage outage communication, and help technicians use policy/equipment knowledge without digging through manuals.
-
Trust isn’t optional: guardrails, permissions, masking, and zero-retention are central to adoption. Mike outlines Salesforce’s approach via the Einstein/Agentforce trust layer—grounding responses in enterprise data, enforcing access controls, masking sensitive information, and addressing toxicity/hallucination risk—while Jonathan emphasizes that many utilities can still adopt agents using private or approved models, not just public LLMs.
Mike Hatch is a Certified Principal Enterprise Architect at Salesforce with deep experience leading enterprise architecture, cloud adoption, and large-scale CRM/ERP integration across retail, manufacturing, and supply chain. Over an 18-year career at Microsoft and subsequent roles at REI and Fossil Group, he delivered global supply chain and direct-to-consumer solutions, modernization programs, and API-first integration strategies (including major investments in commerce platforms and SAP landscapes). He’s known for bridging business strategy with technology transformation—driving data governance, security/privacy compliance, and innovation initiatives spanning analytics/AI and emerging tech—while building and mentoring high-performing architecture teams.
Jonathan Goldstein is an MBA-credentialed AI management consultant who helps operations leaders at multi-entity organizations turn fragmented, post-acquisition data into decision-ready “executive clarity,” with a focus on industries like energy, manufacturing, distribution, and CPG. He emphasizes rapid operational assessments, proof-of-concepts delivered in weeks, and implementation partnerships that take AI from pilot to production—recently enabling a $3B industrial holding company to gain visibility across dozens of acquired entities. He brings 25+ years of technology consulting and energy-sector delivery leadership, including senior roles at CoPoint AI and CG Infinity, where he led value-driven transformations across strategy, delivery, and operational execution.
Article written by MarketScale.
