Your team's
shared response brain.
tonebase combines RAG over your docs, your approved templates, and your team's voice right inside the inbox tools your agents already use.
The problem
Your best agents know
what to say.
The rest are guessing.
Support teams spend enormous time on questions that have been answered before, just inconsistently, slowly, and without the company's best language attached.
- Tribal knowledge trapped in top performers' heads
- Inconsistent tone across agents and shifts
- AI drafts that hallucinate or ignore policy
- No easy way to surface the best past replies
- Long review chains for sensitive responses
- Docs that are scattered, stale, or hard to find
- No visibility into where quality breaks down
Core capabilities
Everything your team needs
to reply confidently.
From first draft to final send, grounded in your knowledge, governed by your policy, built for team collaboration.
The AI reads the full thread, classifies the issue, detects escalation risk, and drafts a reply that fits the actual situation, not just a generic answer.
Retrieves from your help docs, approved macros, and historical replies. Every draft shows exactly which sources and chunks were used.
Personal snippets, team templates, and workspace-locked responses. Templates include use-when guidance, required phrases, and version history.
One click to make it shorter, warmer, firmer, or more formal without changing the factual content. Policy phrases are preserved automatically.
Request a policy check, wording review, or approval right from the thread. Reviewers see the draft snapshot and can edit or approve inline.
Set banned phrases, required phrases, escalation triggers, and approval rules. The AI can never override your compliance or legal constraints.
Track insertion rate, edit distance, adoption by agent, low-confidence categories, and missing template opportunities so quality improves over time.
Every draft shows confidence score and source references. Agents always know what the AI used and can verify before they send.
AI drafts internal notes from thread context, including escalation notes, handoff summaries, and decision logs.
In-workflow plugin
Stays inside the inbox.
Never a new tab.
tonebase lives as a sidebar inside Gmail, Front, Zendesk, and Intercom. Agents never leave their workflow to get an AI-powered draft.
I took a look at your account and can see the charge in question is from [invoice_date]. Based on our billing policy, subscription renewals process on the original sign-up date each month. If you did not receive an advance reminder, that is definitely something we want to fix going forward.
I'd be happy to walk through the details with you. Could you confirm which subscription tier you're on so I can pull the right information?
Use cases
Built for how support teams
actually work.
Three workflows that transform response quality from day one.
Fast, consistent replies at scale
The sidebar classifies the thread, surfaces the right template, and generates a grounded draft in seconds, with citations your agent can verify before sending.
- Category detected automatically on thread open
- Best template suggested from workspace library
- Draft grounded in actual policy docs and macros
- Tone adjustable with one click
- Sources and confidence always visible
Human review before it goes wrong
tonebase flags escalation risk, surfaces the relevant policy constraint, and prompts the agent to request lead review without auto-sending anything it should not.
- Escalation and churn risk detected in thread
- Refund language flagged before it is sent
- Review request sent to lead in one click
- Lead edits and approves from their queue
- Audit log records approver and version
The system improves from your team's work
Low-confidence categories surface in analytics. Knowledge managers can see exactly where quality breaks down and fill the gap with templates and docs.
- Analytics flag low-confidence categories automatically
- Sent replies surfaced as template candidates
- Knowledge manager builds and publishes template
- New docs indexed and linked to category
- Future threads immediately better-served
Architecture
Template-first. Fact-grounded.
Human-reviewed.
tonebase uses a constrained generation approach: templates first, docs second, and freeform generation never without grounding.
The plugin reads the full thread and classifies intent, urgency, escalation risk, and missing context using your workspace's category taxonomy.
Semantic search over your chunked docs plus lexical search for exact policy phrasing. Template candidates are retrieved separately and ranked by relevance.
Drafts are built from template structure and retrieved facts. Required phrases are enforced, banned phrases are blocked, and uncertain claims are flagged.
Agents edit, transform, request review, or insert. Usage is logged with sources used, template used, edit distance, review status, and final outcome.
Integrations
Plugs into the tools
your team already uses.
Native plugins for the most common support inboxes, with the web app as your central control plane independent of any one provider.
Pricing
Predictable pricing for teams
that grow together.
Workspace subscription with included seats and pooled assists. No per-call surprises. Overages are transparent.
- 3 seats
- 1 inbox connection
- 500 assists / month
- Personal + shared templates
- Basic knowledge sources
- Basic analytics
- 10 seats
- 5 inbox connections
- 3,000 assists / month
- Review requests & approvals
- Full template library
- Policy & voice settings
- Analytics dashboard
- 25 seats
- 15 inbox connections
- 15,000 assists / month
- Multiple teams
- Audit logs
- Advanced permissions
- Template versioning
- Low-confidence analytics
- Priority support
- SSO & security review
- Custom deployment options
- Private retrieval controls
- Advanced governance
- SLA support
- Custom integrations
- Custom usage pool
Ready to give your team
a shared response brain?
Start with your docs, your templates, and your first inbox. First draft in under 10 minutes.
"We've tried three AI tools for support. tonebase is the first one that our leads actually trust because they can see exactly what it pulled and block anything that should not be there."
"The template library is the real unlock. We've been building this knowledge in spreadsheets for years. Now it is actually usable by every agent, every time."
"Our compliance team was skeptical of any AI in customer-facing replies. The citation panel and locked templates changed their mind. They can see what the AI used and they trust it."