TOTVS: SAP says human oversight is key to scaling AI agents; AMX/TEF-BZ/TIM: AI reshapes networks, costs, and revenue models

Research evidence excerpt

TOTVS: SAP says human oversight is key to scaling AI agents; AMX/TEF-BZ/TIM: AI reshapes networks, costs, and revenue models

Global Research

16 June 2026ab

TMT Daily Equities

Emerging MarketsTOTVS: SAP says human oversight is key to

scaling AI agents; AMX/TEF-BZ/TIM: AI Telecommunications

reshapes networks, costs, and revenue models Leonardo Olmos,AnalystCFA

leonardo.olmos@ubs.com

+55-11-9658 7533

TOTVS: SAP says human oversight is key to scaling AI agents Gustavo Farias

Associate Analyst

What happened: According to Digitimes Asia, SAP said agents are moving into gustavo.farias@ubs.com

deployment as companies embed genAI in workflows such as finance and supply +55-11-2767 6154

chain, instead of chat and summarization. The firm said many AI applications

reached about 80% accuracy at the proof of concept stage, but that core workflows

require substantially higher reliability and human oversight for safety and

compliance.

POC-to-production is the biggest challenge to enterprise AI adoption. SAP

called bridging the gap to full production the biggest challenge for enterprise

AI adoption. It disclosed it has released 51 AI assistants and 224 AI agents and

expects the total to exceed 1,000 over the next year under its Autonomous

Enterprise vision, which delegates repetitive tasks to agents while preserving

managerial approval for final decisions.

Agents are designed around a human-in-the-loop model. SAP described its

agents as designed around a human-in-the-loop model where AI performs

recommendations, information organization, or preparatory work, and

managers provide final approval to protect supply chain integrity. It said

repeated validation and fine-tuning with real production data is required

because an 80% accuracy rate acceptable in testing is often insufficient for