The commercial problem: First -line agents drown in the tabs, tilles and the fatigue of the tool. The leaders want a faster resolution, deeper relationships and lower costs, but the complexity flies the oxygen of great service. Meanwhile, IA’s media threshing may transform the strategy into a demonstration theater.
HAS Webexone 2025 In San Diego, we sat with Vinod MuthukrishnanVP & COO, Webex Customer Experience (Cisco), and Matt KravitzVice-president of the product and customer chief, Salesforce Service Cloud, to probe a simple question in a true journalistic way A screen, a truth– And give managers a safer ramp at the agentic AI?
Start with people, then the platform
The two leaders argue that while AI accelerates, Humans still manage the vast majority of interactions. Complexity – more than the volume of call – is the silent killer of CX. The immediate price: reduce the work of swivel flesh, surface the context when needed and let people look at conversations that matter.
“As simple things become automated, what remains is complex and significant commitments. This is where people shine. ”
Intelligence connected in practice
The partnership focuses on Integration of the CRM + deep contact center: Think of data, automation and service console of Salesforce merged with telephony, routing and Webex contact telephony. For the representative who lives in Salesforce, calls and channels natively appear; For operations, travel and policies remain consistent between systems.
- A workspace: Reduce contextual switching and cognitive load.
- A truth: Unify customer data, history and the next best action.
- A path to AI: Present automation where it helps, not where it hurts.
“AI is a marathon, not a sprint. It’s a team sport—Milestone by milestone with customers, railings and real value. »»
Original AI: Deployment demos
He has Never was easier to build a smoking AI agent demo – and never more difficult to deploy in a large scale. The framework of the AFETIC IA pair as an operating model, not a gadget: start with the truth (your data), choose the RIGHT Metrics (customer results on “confinement”) and the automation of the phases where it increases people.
- Could vs should: Just because you can automate a step, especially for moments that benefit from human empathy (for example, cancellations or complex changes).
- Good vs bad deviation: Remove the friction for trivial tasks, but invite conversations when there is value in dialogue.
- Guarotler first: Security, security and governance are prerequisites, not fixes.
What integration means daily
For agents: fewer systems, a richer context, a faster resolution time. For leaders: Clearer visibility of marketing at the service, cleaner data for AI and a pragmatic track to automate the right trips in the right order.
Top tips for CX and Tech leaders in 2025
- Master the basics: Stabilize data, travel and KPI before AI scale.
- Apply the test “could vs should”: Prioritize man in the loop where confidence and value are at stake.
- Correction of the agent’s experience first: A unified console beats a dozen dashboards.
- Instrument results: Optimize customer value, not confinement.
- AI of the phase agency: Milestone by step, with cooked governance.
