Industry analyst Joe McKendrick broke-down Brian Solis’ latest ServiceNow research report, “Work Reimagined: The Human + AI Blueprint for Exponential Performance,” in his latest Forbes column.

Business leaders need to consider the possibilities of human-AI collaboration beyond simple task replacement, urges Brian Solis, global head of innovation for ServiceNow and a highly regarded digital analyst and anthropologist, in a compelling thought-leadership piece. He makes the point that he sees many companies attempting to simply slap on AI or associated agents on processes without considering the bigger picture – and opportunity.

“Left to their own devices, executives gravitate toward eliminating costs as AI’s primary use case,” Solis writes. “If cost-cutting and automation are the priorities for C-suites, boards, investors, and shareholders, then AI will deliver those very well – and not much else.”

To assure more productive human-AI collaboration, Solis walks through a seven-phase process:

• Phase 1: Define intent and create a business case. The problem is companies are simply attempting to repeat automation endeavors that may have worked in the past. Instead, business leaders need to “rethink business-as-usual approaches that focus on efficiency gains, cost reduction, and machine-driven job displacement,” Solis says. “Instead, shift your mindset to value creation powered by the unprecedented capacity of AI to augment human capability.”
• Phase 2: Clarify which activities require human attention and which can be delegated to machines. Ask: “‘How frequently does the activity occur and how much capacity does it consume in hours, handoffs, bottlenecks?’’ Solis states. ”’How much unique human value, creativity, judgment, and trust does it generate for the business?’”
• Phase 3: Designate human and agent roles. Build job descriptions pairing people and agents. “Estimate the ‘human-to-AI ratio’ for each role,” Solis advises. Ask: “’How many agents are needed for which roles and tasks?’ and ‘How many humans are needed to guide them?’ ‘Which KPIs will show that both the human and the agent are succeeding?’”
• Phase 4: Build AI fluency. Encourage programs, sessions and coaching to help employees understand AI better.
• Phase 5: Design strategic pilots. Test the arrangement to see if it actually delivers positive results. “Partner an agent with a person, outline metrics, and run a 30-day A/B comparison against the old process,” says Solis. Metrics that should be tracked include time saved, quality enhancement and new capacity redeployed to higher-value work.
• Phase 6: Scale and govern. Here, agents should be managed by an AI resources office, consisting of IT and HR.
• Phase 7: Once AI agents are deployed, shift focus to performance.“Agents should be managed, not installed and forgotten,” says Solis. “Monitor, measure, and improve agent performance over time through a recurring orchestration and management system that assesses what’s working, what’s not, and where to optimize or retire agents.”

The bottom line is that humans can only do so much, and AI can only do so much, But together, they can deliver, as Solis puts it, “exponential outcomes that neither humans nor AI can achieve alone.”