Daniella's Customer Brain

UBS <> OneAI: Touch Base

3 min read

UBS <> OneAI: Touch Base

Date: 2026-03-24 | Duration: 30.440000534057617 min | Account: ubs-mailing

External attendees: mmorgan@ubsmailing.com, nancy@ubsmailing.com, vance@ubsmailing.com

Summary

The team faced significant challenges in contact discovery, hindering their ability to reach decision makers effectively. Calls often ended up with gatekeepers or non-decision makers, prompting a decision to pause outbound calls for two to four weeks while developing a new AI-driven contact discovery model tailored to the specific client’s needs. In parallel, they agreed to test 2,000 accounts with at least one relevant contact to evaluate the current model’s effectiveness. Current call answer rates fluctuate between 12% and 50%, with only about 20% resulting in identifying actual decision makers. Participants, including Daniella Block, Marc Morgan, and Yochai Levi, emphasized the importance of a collaborative approach and highlighted the potential market impact of resolving these issues through innovative AI solutions.

Key Points

🔍 Contact Discovery Challenges: Incomplete contact data prevented reaching decision makers, impacting call success rates and stalling progress. ⏸️ Pause on Outbound Calls: Outbound calls will pause for two to four weeks to build a better contact discovery model. 🧪 Testing 2,000 Accounts: A test with 2,000 accounts with at least one contact will evaluate current model effectiveness and improve outcomes. 📞 Call Identification Rates: Current call answer rates vary from 12% to 50%, with decision maker identification around 20%. 🤖 AI Model Development: The new AI model will use an AI-driven decision tree to adaptively navigate gatekeepers and identify decision makers.

Overview

  • Contact Discovery Challenges: Incomplete contact data prevented reaching decision makers, impacting call success rates and stalling progress.

  • Pause on Outbound Calls: Outbound calls will pause for two to four weeks to build a better contact discovery model.

  • Testing 2,000 Accounts: A test with 2,000 accounts with at least one contact will evaluate current model effectiveness and improve outcomes.

  • Call Identification Rates: Current call answer rates vary from 12% to 50%, with decision maker identification around 20%.

  • AI Model Development: The new AI model will use an AI-driven decision tree to adaptively navigate gatekeepers and identify decision makers.

  • Operational Setup: Contact lists will be validated before calls commence, ensuring accurate targeting to improve overall call efficiency.

Action Items

Marc Morgan Provide an updated list of approximately 2,000 accounts with accurate or appended contacts for test calls (16:41) Inform OneAI team when the updated contact list is ready to proceed with agent activation (17:09) Send 10-15 sample recorded calls of gatekeeper interactions to OneAI team to assist in AI training (20:50)

Daniella Block Coordinate with Michael to verify and finalize the updated test contact list for the 2,000 accounts (14:50) Activate the agent to commence test calls once client confirms contact list readiness (17:00) Provide ongoing updates on development progress, expected launch date, and details of the new contact discovery AI model (17:30)

Yochai Levi Continue supervising AI research and development focused on building the new contact discovery model tailored to handle complex gatekeeper and decision-maker identification (19:50) Support integration of client-provided recorded calls into the AI training dataset for improved model performance (20:50)

Graph View