Hiver

Hiver is a customer service platform that organizes shared inboxes, assigns conversations, supports team collaboration, and adds AI for answering and handling requests. It currently offers two experiences: Hiver in Gmail, which brings support workflows into Gmail, and Hiver Omni, a dedicated omnichannel workspace. The official website presents both. That distinction matters when choosing the product: a team committed to working inside Gmail has a different starting point from a support operation that wants a standalone interface across channels.

Hiver's core problem is familiar to anyone sharing an address such as support@ or billing@. Several people can see the message, but it is unclear who owns the response, what has already happened, and whether someone else is replying. Hiver turns that shared traffic into an organized service workflow. AI assistance builds on the inbox, knowledge sources, and configured procedures rather than requiring every teammate to copy a conversation into a separate chat tool.

Shared inbox ownership and everyday collaboration

The Gmail product comparison documents shared inboxes, assignment, tags, custom views, templates, customer information, and custom fields. These features serve different parts of the work. Assignment establishes ownership; a tag helps organize or route a request; a custom field records structured information; a view helps a teammate find the conversations they need to handle. Together, they provide more context than an unread message count.

A useful setup starts with the service process, not a long list of tags. Define who reviews new conversations, how ownership changes, and which status marks a request as finished. Use categories that support an actual action, such as billing review or technical escalation. If two teams regularly collaborate, decide which team remains accountable for communicating with the customer while another provides an internal answer.

Hiver's email conversation guide describes shared drafts, templates, AI assistance, and scheduled replies. Shared drafts can help when a response needs review from another person before it is sent. A template is appropriate for stable policy wording, while a drafted answer needs to account for the specific customer's circumstances. Reviewers should check both the policy and the facts of the case instead of approving a fluent paragraph on tone alone.

Choosing Hiver in Gmail or Hiver Omni

Hiver in Gmail suits teams that want support management inside an existing Gmail workflow. Hiver Omni is a standalone product that works with Gmail, Outlook, and other email providers, according to the Omni pricing page. It brings email, live chat, and self-service into a dedicated workspace, with additional capabilities depending on the plan. Voice is presented as an add-on, so buyers should check their actual channel requirements instead of assuming every channel has identical packaging.

The choice should be tested with the people doing the work. A finance team handling mostly email may value staying in Gmail. A customer support group with several channels may prefer a central service interface. Ask how existing addresses connect, where an agent reads and replies, and how customer context appears. Also check the mobile and administrator experiences if those are important to the team.

Both versions sit within the same Hiver product family, but tutorials and plan names can describe different experiences. When evaluating a feature, confirm which product the documentation applies to. A screenshot from the Gmail interface is not evidence that an identical control appears in Omni. The useful question is whether the proposed version supports your request flow, permissions, channels, and reporting needs.

AI Copilot for replies and conversation context

The Hiver AI overview distinguishes teammate assistance from customer-facing automation. AI Copilot helps agents find information, summarize conversations, and draft responses. This can be useful when a teammate inherits a long thread or needs policy information before replying. The output should help them understand and resolve the case, not obscure the fact that the customer has already asked the same question several times.

Hiver's Summarizer guide describes extracting key points from written email content and saving an editable summary as an internal note. It also states that this summarizer does not read attachments or comments. That boundary matters in a handoff: if the decisive evidence sits in an attached invoice or an internal discussion, the next agent still needs to review it. A concise summary is an entry point to the thread, not a complete case file.

For reply assistance, ask a focused question such as which policy applies and what information is still missing. Then verify the answer against the approved knowledge source and the customer's record. If the customer needs a refund decision, a polite response is only one part of the work. The team must also establish eligibility and confirm whether the requested action was completed. Keep that operational state separate from the text of a proposed answer.

AI Agents and procedures for multi-step requests

Hiver documents AI Agents that resolve customer requests and hand conversations to people when needed. Its AI overview describes configuring their scope, communication, guardrails, and handoff behavior. The Omni product page illustrates multi-system support tasks, but a buyer should ask to see the exact integration and procedure for their own request type. An example on a product page is not a configured workflow in your account.

The Skills documentation explains how to describe a process, generate and review a draft, define triggers and steps, test it, and activate it for a shared inbox. Actions include tagging, assignment, internal notes, knowledge lookup, and interaction with connected apps. Some actions can wait for a member's approval. This creates a way to encode a support process that requires more context than a fixed matching rule.

A practical first Skill might classify an incoming billing question, look up the relevant guidance, ask for a missing invoice number, and prepare a response for review. Keep the first version narrow. Describe when it should stop and which cases require a human: conflicting account details, an ambiguous request, or a policy exception. Once the team understands the results, it can consider adding further steps and app actions.

Test with cases that differ from the happy path. Include a customer who supplies the wrong identifier, a thread with two separate requests, and a case that needs escalation. Review whether the Skill interpreted the conversation correctly and whether its action was appropriate. Automation success is more than producing an answer; it includes knowing when the process should pause and what context reaches the person taking over.

Knowledge Hub and self-service answers

Knowledge Hub stores and connects the sources Hiver AI uses. The documentation lists websites, snippets, uploaded documents, and connections to Notion, Google Drive, and Confluence. Administrators can map a source to particular inboxes or agents and choose whether it is available to Copilot, an AI Agent, or both. This matters when internal operational guidance should help employees but should not appear in customer-facing answers.

Source maintenance is part of the service workflow. The guide says uploaded documents must be replaced when their content changes, and only text is processed in those uploads. A policy screenshot or an outdated PDF can therefore leave important knowledge missing or stale. Keep the current authoritative version, remove superseded files, and give each source a clear owner. The same principle applies to connected pages whose permissions or contents change.

Hiver also documents AI search for Help Centers. Customers can ask natural-language questions, receive a summary, and open the supporting articles. The answer is grounded in the Help Center's content; the guide says it cannot fully answer information absent from that collection. Good self-service therefore begins with clear articles covering the questions customers actually ask. Generated wording cannot repair a missing policy.

The AI Help Center generation guide describes building content from useful closed conversations, with a documented minimum conversation requirement. Treat generated material as a draft for the knowledge owner to review. A response that was correct for one customer may reflect an exception rather than the standard policy. Publishing that exception as general guidance can create new support problems.

Routing, sentiment, and support reporting

Hiver's round-robin assignment documentation explains distributing new inbound conversations among available team members. It also covers exclusions, tag-based routing, and assignment limits on specified plans. The guide notes that custom assignment automations run before round robin. That ordering is useful to understand when a conversation goes to a specialist rather than joining the general assignment pool.

Capacity settings are operational controls, not a guarantee that every person has the same workload. Two open conversations can require very different effort. Review unassigned requests, complex escalations, and availability alongside the count. If the entire assignment pool is at capacity, the documentation says new conversations can remain unassigned until capacity returns. Someone needs to own that queue and respond when the rule cannot distribute more work.

The sentiment report guide describes AI classifications and reporting by inbox, time range, and user. This can help managers find conversations that deserve inspection. It should not be used as an automatic judgment of agent quality: a teammate handling billing disputes may naturally receive more negative messages. Read representative conversations and consider assignment patterns before drawing a conclusion about an individual.

Current AI documentation also describes using sentiment changes in automations. A negative shift can trigger attention or escalation where configured. The setup guide encourages correcting classifications when they are wrong. That makes the system's intended role clearer: sentiment is a signal for investigation, and the team remains responsible for interpreting tone in the actual customer context.

A workflow for a delivery-related billing question

Imagine a customer emailing a shared billing address because an order arrived late and they want to know whether a charge will be refunded. Hiver assigns the conversation to the appropriate queue and preserves ownership. Copilot summarizes the existing thread, while the agent checks the order record and any attachments separately. Knowledge Hub provides the current delivery and refund guidance.

If the organization has configured a suitable Skill and integration, the process can gather the missing order identifier, look up the approved information, and prepare the next step. A refund outside the standard policy is handed to a person with the right authority. The team leaves an internal note explaining the decision, reviews the draft reply, and confirms the operational action before telling the customer it is complete.

The conversation then contributes to reporting. A manager can examine whether similar requests are increasing, whether the handoff is causing delays, and whether the Help Center needs clearer guidance. This example illustrates a configured support workflow, not a tested result or a default automation that every subscription includes. A useful demonstration should show both the straightforward request and the policy exception.

Access, limitations, and buying considerations

Hiver publishes separate pricing comparisons for Omni and the Gmail product, with paid per-user tiers and a free trial route. AI capabilities, workflow limits, integrations, and support reporting vary by plan. The signup page also mentions a free plan, while the product pricing pages emphasize paid tiers. Confirm the current free eligibility and limitations directly rather than assuming the trial's AI features remain available indefinitely.

Before choosing a plan, count the people who need Hiver access, shared inboxes, required channels, and workflow volume. Ask how AI usage is measured and whether the functions in the demonstration are included or separately metered. Also verify the commercial treatment of voice and any advanced integration. A headline per-user price does not by itself establish the total cost of a support setup.

The most important practical limitation is the quality of the process Hiver automates. Unclear escalation rules, inconsistent policy sources, and incomplete customer records create ambiguity for both AI and people. Start with a manageable inbox, keep knowledgeable reviewers involved, and check actual conversations after activation. Hiver can coordinate and assist the workflow; the organization still needs to decide what a correct resolution means.

Frequently asked questions

Must the team use Gmail?

Hiver in Gmail works inside Gmail. Hiver Omni is the standalone option documented for other email providers. Choose and evaluate the specific experience you intend to deploy.

Can AI replies use internal documents?

Knowledge Hub can connect approved internal sources. Map them deliberately to Copilot or customer-facing agents, and review whether the content is appropriate for its intended audience.

Does the email summary include attachments?

The dedicated Summarizer documentation says it uses written email messages, not attachments or comments. Other AI features have their own scope, so check the relevant guide.

Can a workflow require human approval?

Yes. Skills documentation describes approval controls for actions. Configure and test the review points for the procedures you want to automate.