Chicago, IL, Nov. 02, 2026 (GLOBE NEWSWIRE) — warrantini, AI-native warranty infrastructure for brands and manufacturers, today announced that its latest platform release connects warrantini’s full functionality directly to Claude, ChatGPT, and Microsoft Copilot through an MCP server. A brand connects warrantini to its assistant one time. After that, its team asks in plain English, and the assistant works against live warranty data instead of a spreadsheet export.

warrantini puts warranty operations inside Claude, ChatGPT, and Microsoft Copilot

warrantini

The assistant can do the work, not just answer questions about it. A product team can load 400 SKUs from a spreadsheet, find every product with no warranty rule covering it, or ask which supplier is over-represented in this quarter’s claims. An automated agent can watch for new claims and escalate the ones that matter. Nineteen tools cover orders, products, registrations, claims, and coverage rules, and every write passes through the same validation, permissions, and audit trail as the dashboard. Account administration and billing stay in the dashboard on purpose.

AI in the decision path, not a chat window

In warranty software, AI usually arrives as a chatbot that answers customers faster. That solves a queue problem. The hard part of a claim is judgment: defect or misuse, covered or not, repair or replace, and whether it’s the fourth one this month on the same product line. warrantini built its AI into the workflow where those calls are made.

Six decision points sit inside it. Triage classifies every claim on arrival. Assessment pre-fills the defect analysis, choosing only from the brand’s own categories. Resolution recommendations cite up to five of the brand’s comparable past claims. Escalation, rule recommendations, and defect pattern detection cover the rest. Auto-approval ships off, and a brand that turns it on sets the confidence threshold.

Every decision is written to an immutable record of its inputs, reasoning, confidence, and model, and every human override is appended to it. If the model is unavailable, the manual workflow runs exactly as before. That is the difference between AI built into a workflow and AI added to one.

A development platform, not an endpoint

The MCP server is one of three ways into the same platform, alongside a versioned REST API and signed webhooks. Around them sits what engineers check first: test and live keys separated at the credential, a sandbox on an isolated database branch with seeded data, documentation generated from the code and readable without signing up, a written 90-day deprecation commitment, stable error types, idempotency keys on every write, and Postgres row-level security enabled and forced on every table.

warrantini’s own customer portal runs on the same public API. A brand can use the portal as it ships, replace it with its own front end, or do both, and the warranty engine underneath doesn’t change.

“Warranty is a series of judgment calls, and the usual place for AI in this category is the front desk, answering customers faster,” said Brian Carpizo, Managing Partner of MGP Labs, which develops and operates warrantini. “We put it where the calls get made, and we recorded every one, so a person can see why it said what it said and overrule it. The platform underneath follows the model Stripe set for payments: a product you can use on day one, and an API developers would choose to build on.”

Carpizo, a former McKinsey & Company leader, advised manufacturers on applying AI to warranty operations while at the firm, which is where warrantini started.

Availability

warrantini is available now at $49–399 a month, flat, with a 14-day trial (cancel any time) and setup assistance at no charge. Developer documentation is public at warrantini.com/developers, with no signup required.

About warrantini

warrantini is AI-native warranty infrastructure for brands and manufacturers. It turns coverage into a system of record: a branded portal where customers register products and file claims, a workflow where your team resolves them, and a versioned API and MCP server so engineers and AI agents can drive the same platform. Flat pricing from $49/mo. warrantini is developed and operated by MGP Labs, a division of Mahusai Global Partners. For more information, visit https://warrantini.com/

About MGP Labs

MGP Labs builds, owns, and operates AI-native vertical software companies. It is a division of Mahusai Global Partners LLC, a U.S.-headquartered consulting firm of fifteen years, serving Fortune 1000 industrial companies and private-equity clients, with staff across Southeast Asia, South America, and Europe. It is led by Brian Carpizo and Ed Humphries, both former McKinsey & Company leaders. warrantini is its first product. For more information, visit https://www.mgplabs.ai/ 

Press Inquiries

Jamie Dayagbil
jamie [at] warrantini.com
https://warrantini.com/

A video accompanying this announcement is available here: https://youtube.com/watch?v=MdSUS24X42U


Primary Logo

About The Author