Use case · Complaints handling
Agent Serve logs every complaint with its deadline, drafts a reply from your policies for a handler to approve, and shows the causes behind them.
A complaints handling use case for Agent Serve, the AI workspace where teams build, deploy, and manage AI agents: a workflow classifies each complaint, logs it in a table with its deadline and drafts a reply from the knowledge base for a handler to approve, while rules and dashboards show root causes.
At a glance
The problem it removes
Complaints come in by email, web form and phone, and some are never logged. Handlers retype each one into a case log, look up the policy, and write the reply from scratch. Deadlines are tracked in a spreadsheet, and the monthly root cause report is a manual count.
How it runs
Agent Serve logs, classifies and drafts. A handler reads the draft and decides what the customer receives.
A new email in the complaints inbox starts the workflow. Web forms can start it too, through a webhook.
An agent tags the product, cause and severity, and the Table block logs the case with the date its response is due.
The Knowledge block finds the policy passage that applies, and the agent drafts a reply that cites it.
A Human step sends the draft to the handler, who edits it, approves it, or takes the case over.
Before and after
| Work | Before | With Agent Serve |
|---|---|---|
| Logging | BeforeHandlers retype each complaint into a spreadsheet log. | With Agent ServeAgent Serve logs each complaint with its product, cause and deadline. |
| Deadlines | BeforeSomeone filters the log each week to find what is due. | With Agent ServeA morning run lists cases close to their deadline for the team. |
| Replies | BeforeWritten from scratch, with policy wording copied by hand. | With Agent ServeDrafted from the current policy and approved by a handler. |
| Root cause | BeforeA monthly manual count of categories in a pivot table. | With Agent ServeA dashboard and Chat show causes by product and channel. |
In depth
Agent Serve gives handlers the draft and the deadline, and gives managers the causes.
A manager mentions the complaints table in Agent Serve Chat, asks which causes grew this month, and gets a chart built from the rows.
An Agent Serve classification sorts complaints into labels such as Regulatory, Service and Billing, with General as the fallback.
When an Agent Serve Relay drafts a reply, it shows the exact message and waits for a handler to approve or decline it.
One step of the process: the Agent Serve Call Compliance dashboard scores uploaded call transcripts against your checks, and a check can flag calls where a customer complained.
What does the work
A workflow classifies each complaint, logs it and drafts the reply.
A complaints table holds every case with its deadline and owner.
Policies and procedures sit in a knowledge base the reply cites.
A classification rule sorts complaints the same way everywhere.
A dashboard shows volume and causes by product and channel.
Examples
Three complaints processes financial services teams run on Agent Serve, each with a person approving the reply.
An Agent Serve workflow classifies each complaint, logs it with its deadline and drafts a reply for a handler.
An Agent Serve dashboard shows complaints by product, cause and channel, and arrives as a PDF before the monthly review.
An Agent Serve schedule runs every morning, lists open complaints close to their deadline and posts them to the team’s channel.
Examples show what teams can build in Agent Serve. They describe patterns, not named customer deployments.
Governance
Agent Serve keeps a handler on every reply and keeps complaint records to the people who need them.
Security and governanceAn Agent Serve Relay can be limited to internal recipients, an allowlist or anyone, and never sends email without approval.
The Agent Serve Guardrails block masks personal data in a draft before it moves to the next step.
Relays never send email or book meetings without an approval, under any autonomy setting. An approval covers the exact payload the reviewer saw and expires if nobody decides.
Proposals, approvals, declines and executed actions are written to the workspace audit log, which admins can export.
Row-level policies narrow what each person sees in tables, dashboards, exports and scheduled reports.
Admin, write and read roles per workspace, plus permission groups that limit which models, blocks and integrations a team can use.
Sign in through your identity provider with SAML 2.0 or OIDC, and provision people with SCIM.
Run Agent Serve in your own cloud or data centre, next to the systems it reads.
Zero data retention with supported model providers, backed by your own provider keys and agreements or self-hosted models. Some provider features keep data by design, such as Gemini Deep Research, so they are outside zero data retention.
Every workflow run records each block’s inputs, outputs and errors, so a reviewer can trace how a result was reached.
Further reading
The deep dive models the cost of missed complaint deadlines and what changes, with every figure labelled.
Read the blogMost complaint breaches are visible weeks before they happen, in a queue nobody has time to read. Here is a setup that computes every deadline, flags risk early and puts the at-risk list in front of the right person.
Retail banking service · 12 min readCounting complaints tells you volume. Grouping them by root cause, product and journey step tells you what to change. Here is a dashboard built for the second question.
Retail banking · 5 min readCustomers rarely say the word complaint. This architecture transcribes each call, flags likely complaints with the exact words used, and lets a complaint handler confirm every entry in the register.
Contact centres · 5 min readQuestions
No. In this Agent Serve process a handler approves every reply. A Relay never sends email without approval of the exact message.
Agent Serve workflows start from a new email in Gmail or Outlook, a web form through a webhook, a new row in a table, or events from tools such as Zendesk.
Each complaint is a row in an Agent Serve table with its due date, and a scheduled workflow lists the cases close to their deadline every morning.
Only as one step. Call checks live in Agent Serve Dashboards, where the Call Compliance template scores uploaded transcripts against your checks.
A team can start with the Agent Serve intake workflow and complaints table, then add reply drafting, the root cause dashboard and the morning deadline run.
Last updated
Enterprise access
Agent Serve is available to teams by invitation. Tell us what you want to build and we will set up your workspace with you.