Summary of findings from the Gradient Labs website (product, pricing, security/compliance, terms, customers, and resources). Company snapshot: Gradient Labs is positioned as a fintech-first autonomous customer operations AI agent, headquartered in London and founded in 2023. The company announced a $13M Series A in 2025 led by Redpoint Ventures (press release on site). Product & capabilities: Gradient offers a single procedural AI agent spanning chat, phone, email, SMS and social channels. Agents are authored in plain language (no coding required) to define reasoning and actions. Capabilities include omnichannel automation for frontline responses, proactive outreach, back-office case building and evidence collection, and automated handoffs to humans. Financial-services-focused guardrails and controls are described (approval requirements for sensitive decisions, role-based access, SSO, audit logs, automated QA, regulatory nuance awareness). The product uses multiple LLM providers (OpenAI, Anthropic, Google models) and supports failover across cloud/LLM providers. Integration is API-first and can run on top of existing platforms (Intercom, Zendesk, Freshdesk) without replatforming. Testing and monitoring features include scenario testing in the web app and real-time monitoring of resolution rates and CSAT. Emphasis on continuous learning from human handling and iterative improvement is stated. Pricing & commercial model: Pricing is outcomes-based (pay for successful query resolutions delivered by the AI agent); the site indicates there is no free trial. No fixed public price points were listed; pricing and implementation vary by customer and are often run as customer-specific projects with delivery teams. Onboarding outcomes cited: 20–50% day-one resolution for chat out-of-the-box, commonly 40–60% quickly, and potential to exceed ~80% as integrations, SOPs and data are added. Positioning emphasizes end-to-end automation rather than a human-in-the-loop co-pilot. Security, data handling & compliance: The product page claims SOC 2 / "banking-grade" controls and enterprise-ready controls. Terms reference GDPR and a DPA/App Privacy Policy (Customer Personal Data processed per a DPA and App Privacy Policy). Access and governance controls include SSO, role-based access, audit logs, approval workflows and automated QA. Multi-LLM strategy is noted to reduce vendor/ops risk. The site’s standalone privacy URL currently returns a 404; the Terms reference an App Privacy Policy/DPA but that standalone privacy page appears missing. Terms of Service highlights: Relationship governed by Order Form + Terms of Service. SLAs apply only if included in the Order Form; Gradient may modify SLAs with notice. Term starts per Order Form; 30-day notice for convenience termination at term end; early termination may leave certain fees payable. Fees are non-refundable unless stated; 30-day payment terms; 60 days’ prior notice for fee changes with a 30-day right to terminate if you disagree. IP: each party retains its own IP; Gradient owns Service IPR; Customer grants licenses to use Customer Data to deliver the Service. Data & model use: Gradient may fine-tune LLMs with Customer Data for the Service; explicit consent required to fine-tune models used by other customers; fine-tuned LLMs are to be deletable at term end per terms. Data processed per DPA/App Privacy Policy; training/re-use beyond the Term is prohibited by the Terms. Liability is capped to fees paid in prior 12 months with carve-outs for statutory liabilities. Governing law: England & Wales. Customers & case studies: Named customers/case studies listed on site include Sling Money, Zego, Plum, Lendable, Yonder and Nala. Example outcomes cited on the site: Sling Money — 50% out-of-the-box resolution day one; 78% automated resolution after enhancement; CSAT up to ~86%. Zego referenced for CSAT improvements. The site emphasizes measurable CSAT and resolution rate improvements; these claims should be validated in procurement/POC. Resources: White paper "Practical AI Customer Support Playbook" (tactical guidance from >100 implementations). Blog posts on vertical AI for financial services, implementation lessons, and product thinking. Press post announcing Series A and product milestones. Site observations & missing items: The privacy and contact pages returned 404s; Terms references a DPA and App Privacy Policy but the standalone privacy page was not available. Media contact appears as [email protected] on some pages. Blog (blog.gradient-labs.ai) includes About and Careers pointers. Caveats & verification items: Request current SOC 2 report or security assessment; request the DPA and App Privacy Policy directly (site references but standalone page 404); ensure required SLAs are written into Order Form; verify data use and model fine-tuning limits, opt-in/opt-out and deletion guarantees in writing; confirm whether the liability cap (prior 12 months fees) is acceptable and negotiate if needed. Practical next steps suggested: Request demo, request SOC 2 and security architecture documentation, request explicit DPA/App Privacy Policy and confirmation on data location/processing, request sample Order Form showing fees/SLAs/termination/outcomes, request references and raw metrics/QA reports, run a small POC using the provided playbook, confirm integration specifics for your stack (API details and supported connectors). Links: Home/product/pricing/terms/white paper/customers/press/blog/about and note that contact/privacy currently return 404 on the site.