CX Automation Performance Analyst

About the role

The CX Tooling & Innovation team at Betterment is responsible for the systems, AI programs, and automation that power our customer experience. Shipping AI is only half the job — keeping it performing is the other half. As our CX Automation Performance Analyst, you own the post-launch life of our AI programs. You ensure that what gets launched keeps working, keeps improving, and is well-documented enough that the whole organization can benefit from it.

• You ensure that what gets launched keeps working, keeps improving, and is well-documented enough that the whole organization can benefit from it.
• You identify performance degradation, intent gaps, hallucination, model drift, and guardrail failures across Zendesk AI, chatbots, ticketbots, and automation workflows.
• You own the incident response process for AI quality issues — when performance degrades or a guardrail fails, you triage, escalate, and drive resolution with a clear post-mortem.
• You serve as the primary admin for net-new AI tooling post-launch — including VoiceAI platforms, Zendesk AI agents, and any other tools handed off from the implementation phase.
• You maintain active optimization backlogs for launched programs, ensuring improvements don't stall after initial implementation.
• You drive post-launch iteration in partnership with the Chatbot Strategy Manager and tooling teams, tracking outcomes rigorously.
• You own the AI program documentation library: runbooks, flow documentation, QA procedures, training guides, and internal knowledge base articles.
• You maintain accurate records of AI program behavior, guardrails, and change history to support compliance and audit readiness.
• You partner with the CX AI & Automation Manager during launch phases to establish performance baselines, define QA criteria, and ensure clean handoff documentation.
• You build feedback loops from AI performance data to inform product, engineering, and CX leadership decisions — your insights should shape what gets built next, not just what gets fixed.

What We're looking for

• Hands-on experience analyzing performance data from AI or automation tools in a support environment.
• Understanding of common AI failure modes — hallucination, model drift, guardrail gaps, and intent misclassification — and the ability to design QA processes that catch them before they reach customers.
• Proficiency with data visualization and reporting tools (e.g., Zendesk Explore, Looker, Tableau, or similar) and comfortable writing queries to validate AI performance data.
• Strong documentation skills — you write clearly, organize well, and build resources people actually use.
• Analytical and detail-oriented: comfortable building dashboards, running QA analysis, and spotting patterns in messy data.
• Experience with Zendesk (or comparable CX platforms) and AI tools such as Zendesk AI, Decagon, or similar.
• Experience in fintech, financial services, or another regulated industry where AI outputs are subject to compliance review and audit.
• A continuous improvement mindset and systems thinker — you see how knowledge structure, prompt design, monitoring, and feedback loops connect, and you're never satisfied with "it launched."

Core Values

• We change lives
• We set audacious goals
• We don't just show up; we give our all, then celebrate our wins.
• We value all perspectives
• We simplify financial services

What Happens Next

• We'll take a few weeks to review all applications.
• If we'd like to spend more time with you, we'll reach out to arrange next steps, which will include 3-4 sets of meetings with your future colleagues.
• In the interview process, we'll look to learn more about your skills, experiences, capabilities, and motivators.
• Depending on the role, we may ask you to complete a case study exercise or technical assessments, as we want to collect a robust set of data points to better inform our decisions.
• On average, it takes us around 3-5 weeks to make a hiring decision, depending on your availability and sense of urgency.