An independent publicationEvery case sourced, dated & graded A–DGet the weekly Casefile →
Grade BproductionSource date Oct 17, 2026· Last verified Oct 17, 2026

How is SoFi using AI agents for member support?

SoFi · Banking & Financial Services

Header image: Hands holding a bank card at a table

SoFi (13.7 million members) uses a Sierra AI agent for member support. Sierra reports 61% containment across more than 50,000 conversations weekly.

Reported outcomes

containment
61%
Source [1] · self-reported
volume
more than 50,000 conversations weekly
Source [1] · self-reported

Outcomes are reported exactly as stated by the source; we have not independently audited them.

How it works

  1. AI agent answers member questions across banking, investing, and lending products.
  2. Escalates to human agents when needed.

Risks & compliance

  • Wrong or hallucinated answers to customers; needs guardrails and handoff
  • Vendor-published results; not independently audited
  • Regulated financial disclosures and identity verification
Build this yourself

Tools mentioned & how to build it

Disclosure: Some links may earn us a commission; it never affects grades. Tools are listed by fit to this case, never by payment.

Tools mentioned

  • Sierra (Sierra)
    Conversational AI agents for customer service, chat and voice.

Build a version yourself

  • n8n
    Open-source workflow automation with AI agent nodes; self-host or cloud.
  • HubSpot (Breeze agents)
    SMB-friendly CRM with Breeze AI agents for prospecting and support.
  • Relevance AI
    Build an 'AI workforce' of agents, e.g. BDR agents that research and book meetings.

All tools by category →

Sources

[1]Sierra· vendor customer story· Undated (accessed Oct 17, 2026)
“The AI agent achieved 61% containment, handling more than 50,000 conversations weekly.”
SoFi and Sierra

Grade rationale: Named organization with a quantified outcome from the vendor or the organization itself (self-reported). Automated check on Oct 17, 2026: 1/1 quotes and 2/2 outcome claims matched the source. Compiled by the Agent Casefile research desk (AI-assisted; see methodology). Spot an error? Request a correction.

Keep reading

Related cases