New research from Conviva shows that consumer-facing AI agents spend an average of 2 minutes and 32 seconds gathering basic context causing 8.5% of customers to abandon the session entirely.
The report shows purchase journeys are largely invisible to analytics tools and the problem compounds when AI agents act on the same data.
Conviva DPI reveals patterns behind every consumer interaction and conversation across apps, websites, and AI agents, uncovering growth opportunities, friction, inefficiencies, and confusion in experience, to drive predictable outcomes.
This retailer is deploying AI agents across 100 million weekly sessions but was blind to the two shopper segments driving $25 million in weekly revenue. Conviva changed that.
Keith Zubchevich, Conviva CEO, speaks with Total Retail about why customer-facing AI fails and how to measure accuracy and efficiency in automated systems.
Aditya Ganjam, Conviva Co-Founder and CPO, speaks with IT Nerd about how organizations can convert AI from novelty into predictable business performance.
Read Keith Zubevich's latest on why engagement with third-party agents should be selective and guarded, with clear identifiers and visibility.
Keith Zubchevich at Conviva explains why growing up depends on outcomes, not behaviours.
Keith Zubchevich, Conviva CEO, sat down with Neil C. Hughes on Tech Talks Daily podcast to unpack an honest analogy for today's AI rollout.
BetaNews spoke to Keith Zubchevich, Conviva CEO, about why today’s AI deployment is like sending a toddler out to get a job: bursting with potential, but nowhere near ready for the real world.
AI and robotics are heading into a pivotal year. After rapid advances in agentic systems, embodied intelligence and enterprise automation, 2026 will be defined less by experimentation and more by proving what works in the real world.
After years of experimentation, agentic AI has proven itself in sandboxes. In 2026, the focus shifts from capability to accountability: How do you govern an autonomous workforce? Who supervises the supervisors? And what happens when unsanctioned “shadow agents” start making decisions nobody approved? Industry leaders weigh in on what’s coming as autonomous systems move from proof-of-concept to production.
