System
/
Memory
/
Tone of Voice

Tone of Voice

One definition of how your company sounds: the style principles, the language preferences and the patterns that are never acceptable. Held as your GTM system's memory, applied by every capability that writes and inspectable when something sounds wrong.

Learning from every outcome
One prohibited pattern keeps reappearing
Edited out of 47% of generated messages
Not listed in any of 23 prohibited patterns
Human edits cluster on openings
68% of edits touched the first two lines
Body copy edited 1 time in 8
Upgrade proposedv22v23
LEARNING
The memory upgrades
Memory
Tone of Voice
v23
Style principles stated with examples
Opening line constraints added from edit evidencev23
Language preferences recorded per context
Reappearing pattern added to the prohibited listv23
Prohibited patterns listed and enforced
Approve upgrade to Tone of Voice?
The current way > AI added on > AI-native

The moments something gets written, and what shapes it

Tone is the part everyone can hear and nobody has written down. It survives as long as the people who absorbed it stay, which is not a system.

01 | The Current Way

02 | AI Added On

03 | AI-Native

A message gets generated

Whatever the writer absorbed

Style comes from having read enough of the company's writing. New people approximate it and tools do not.

A style prompt, per tool

Each tool carries its own fragment of guidance, and everyone interprets it a little differently, so there is no real consistency and the voice drifts further apart over time.

One consistent style, everywhere it writes

Every capability that writes reads the same approved rules, so the company sounds consistently like itself, whichever one produced the line.

Someone edits the output

Fixed silently, every time

The same correction gets made by hand repeatedly and never becomes a rule.

Edited, not learned from

Nothing captures what was wrong, so the drafting assistant repeats the same slip tomorrow, and tone drifts further still from person to person and model to model.

Edits become rule evidence

A correction made repeatedly is visible as a pattern, and becomes a candidate rule rather than permanent manual work.

A phrase should never be used

Known by some, not all

The people who know why it is banned are not always the people writing.

No concept of prohibited

Even when a banned phrase list gets updated, the update doesn't cascade to every tool or every person using one, so the same banned phrases keep cropping up.

Prohibited patterns are explicit

The list of what is never acceptable is part of the memory and applies across the whole organisation from the moment it is approved, not caught after the fact in review.

The style needs to change

Updated in one of many places

The brand document changes and the prompts scattered across tools do not.

Change everything by hand

Every tool holding its own style fragment must be found and edited individually.

Changed once, versioned

The rules are updated as a version, so every capability writes to the new style from the same moment.

Something sounds wrong

No way to trace it

The output is off and there is nothing to inspect that would explain why.

Opaque by construction

Style lives inside a prompt nobody reads, so diagnosing tone means guessing.

The active rules are inspectable

You can read the style rules that produced any piece of work, which is what makes tone something you can actually govern.

Referenced by

Capabilities that reason with this memory

Prospecting

Daily Brief | Prospecting

Active
Six accounts moved in-market overnight
Flagged with what changed and why they're worth chasing
The right contact at Fenwick Systems, not the switchboard title
Picked from who's actually engaging, not the org chart
First email and LinkedIn message already drafted
Written to the pain they're showing right now, ready to send
+
Two prospects from last week gone quiet
Flagged to follow up before the window closes
Today's list ready before your first call
SEE HOW IT WORKS

Prospecting

Propensity

Active
Three accounts just moved into in-market
New funding and hires in the roles that feel this pain
Meridian Logistics jumps to the top of the list
Same profile as your last five closed-won deals, and hiring for the gap you fill
A contact at Northbridge is showing real intent
Visited pricing and a comparison page twice this week
+
Your ranked account list, ready for outreach
Who to call first, and the reason why
Priority accounts, ranked, before you start prospecting
SEE HOW IT WORKS

Sales

Meeting Prep

Active
Brief built for tomorrow's Acme renewal call
Pulled live deal state 1 hour before the meeting
Flagged two deals with no next step
Both stalled 14+ days since last activity
New stakeholder added to the buying group
CFO joined — added to the account brief
+
Objection playbook matched to the call
Pricing pushback: three proven responses surfaced
2 briefs ready for your review before 9am
SEE HOW IT WORKS
SEE ALL CAPABILITIES
Used by

Skills that use this memory

No items found.
SEE ALL SKILLS
The Platform

One system that understands, decides, acts and learns.

Every GTM signal flows through an AI-native operating layer into a system that runs on the surfaces your team already uses.

Explore the GTM System →
GTM Data & Knowledge
CRM · Emails · Calls · Marketing · Product · Support · Documents · Research
AI-Native Operating Layer
Context · Memory · Skills · Agents · Decision Traces
AI-Native GTM System
Understand
ICP Fit
Deal & Account State
Stakeholders
Risk & Health
Decide
Targeting
Pipeline & Forecast
Campaign Planning
Renewals & Expansion
Act
Outreach & Follow-up
Meeting Prep
CRM Updates
Alerts & Escalation
Learn
Upgrade ICP
Upgrade Messaging
Upgrade Playbooks
Upgrade Forecasting
Surfaces
CRM · Slack · Teams · ChatGPT · Claude · MCP · API
Output
Briefings · Artifacts · Alerts · Recommendations · Approvals · Actions

SEE WHY REVOPS + MARKETING LEADERS CHOOSE REVENUE LABS

The GTM teams that learn fastest will win. Build yours a system that learns.

An advantage competitors cannot buy back: years of success and failure, codified.

FAQ

What buyers ask

If it's not here, we'll answer it live.

Talk to us
What is tone of voice memory?

One approved definition of how your company writes: the style principles, the language preferences, the patterns that must never appear, and worked examples. Held as a version that every capability generating customer-facing communication applies.

How is tone of voice memory different from a prompt?

A prompt is a fragment held inside one tool by whoever configured it. Memory is one governed definition applied across every capability, with versions, owners and an audit trail. Prompts drift independently; a memory changes once and everything follows.

How do you give AI a company tone of voice?

By making the implicit rules explicit, which is most of the work. Style usually survives as taste rather than as a document, so the useful exercise is extracting the actual principles, the preferred language and the prohibited patterns, then holding them where every capability reads.

How do you govern AI writing style?

By treating edits as evidence. When people correct generated communication in the same way repeatedly, that pattern is a missing rule. Capturing it turns permanent manual editing into a proposed change that someone approves, and the rules stay inspectable throughout.

How is this different from messaging?

Messaging is what you say: the claims, the narrative, the pains and the objections. Tone of voice is how it sounds. They are separate because a message can be entirely on-message and still read as though a different company wrote it.