Add-on · Brand Brain Build
We build the Brain. You walk into a working instance.
An optional engagement for teams who would rather hand the onboarding off. Our team reads the business, profiles the competitor set, writes the voice document, drafts the audience profiles, and enters every piece of it into your instance after you sign off. You log in to a Brain that is already loaded.
The math
Paid once at kickoff, on top of the annual subscription. Nothing recurring. No per-revision charge inside the revision window.
- About five weeks end to end
- Discovery, build, approval, load, revisions
- Done by a writer, not a prompt
- Brain handed over fully editable
The build costs more than a year of the tool because doing it yourself costs more than that.
The product is built to be self-serve. Most teams will run the onboarding themselves in an afternoon and never need this page. The teams that do need it have already tried, and the cost of that attempt is the reason this engagement exists.
The lost month on the voice doc
Two people block out a Friday afternoon to write the voice document. They make it through about half the brand before the quarterly planning cycle pulls them back. The half-written doc sits in a shared drive for nine more weeks before anyone opens it again. By then the brand has shifted and the work is partly stale.
The drift across sub-brands
The parent brand has a voice. The two sub-brands inherit it loosely and pull in opposite directions inside a quarter. Without a written split, the agency, the in-house team, and the founder each interpret the inheritance differently. Drafts come back sounding like three different companies.
The competitor work nobody gets to
Every quarter the team agrees the competitor file needs to be refreshed. Every quarter the work slips. The result is that the AI assistant in the product is reasoning against a competitor set that was current eighteen months ago, and the team is the last to notice.
The cold-start problem on day one
An empty Brand Brain produces drafts that read like every other AI tool. The system learns from edits, but the first hundred edits are paid for in the team's time. Teams that walked away from earlier AI tools usually walked away inside that first hundred.
The build is a forcing function. It puts the lost month on the calendar, in our hands, with the customer signing off at the end. The Brain that lands inside the instance is the one the team would have built themselves on the version of the schedule that never came.
About five weeks of work, phase by phase.
Discovery, two weeks of build, an approval gate, a load week, and two weeks of open revisions. The customer is on two calls and one review window. Everything else happens on our side until the Brain is ready to enter the instance.
Phase 01
Discovery
Kickoff calls and the shape of the engagement.
Two working sessions in the first week. The first call covers the business as it stands today; the second covers the brands, the verticals, and the people we will be reading and writing for. You hand over read-access to whatever the brand already has on paper.
- Two kickoff calls of about an hour each
- Read-access to existing brand assets and writing
- Alignment on the sub-brand list and the verticals we profile per audience
- Sign-off on the competitor list before any profiling work begins
- Optional interview list: customers, founders, sales leads
Phase 02
Build week one
Reading the business and drafting the voice.
Our team reads every public page, then works through the long tail of social posts, sifts the support archive, reads every published case study, listens to the recorded founder talks, and reviews the sales deck library. The voice document gets drafted from how the brand actually talks when it is not trying. Audience profiles get sketched per vertical. The refusal list comes out of what the brand has quietly stopped saying.
- Full read of public writing and any assets the customer shared
- First draft of the voice document
- Audience profile per vertical the brand serves
- Initial refusal list of words, claims, and frames the brand will not publish
- Proof points pulled from real customer evidence
Phase 03
Build week two
Competitor research and the first test draft.
The competitor set gets profiled in depth. Each profile is built from public material the same way the voice document was built — the live site, the blog archive, the social presence, any recorded talks, and the press coverage. Sub-brand splits land in this week as well. By the end of the week, a first piece of content has been generated against the in-progress Brain so the customer can see how the parts behave together.
- Profile of every approved competitor
- Sub-brand splits for groups running more than one brand
- Strategic posture and positioning notes per competitor
- First piece of content generated against the in-progress Brain
- Internal review pass before anything goes to the customer
Phase 04
Approval gate
The customer reads the full Brain and signs off.
Nothing touches the live instance until the customer has read the entire Brand Brain in a working doc and approved it. We walk through every section together on a call. The customer marks what is right, what is wrong, and what is missing from the draft. We rework anything flagged before the load week begins.
- Walkthrough call covering every section of the proposed Brain
- Customer-side review window of up to one week
- Marked edits returned to our team for rework
- Second walkthrough if the first pass needs material changes
- Final sign-off on the document before any data lands in the instance
Phase 05
Load week
We enter the entire approved Brain into the instance.
Voice document, audience profiles, refusal list, proof points, sub-brand setup, every competitor profile, technical context entries — all of it gets typed in by our team. The customer does not enter any of it. The instance is set up so that the first time the customer logs in after load week, every screen of the product is already populated.
- Voice document loaded into the brand studio
- Audience profiles created per vertical
- Refusal list entered with every banned word and frame
- Proof points filed against the claims they back
- Sub-brand records created with their own voice splits
- Competitor profiles loaded into the Competitors page
- Technical context entries created from the discovery materials
- A first publishable draft prepared inside Create Content
Phase 06
Revision weeks (two)
An open revision window after the Brain is live.
For the two weeks following load, the customer can request edits to anything inside the loaded Brain — voice, audience profiles, refusal list, competitor entries, proof points, sub-brand splits, every field of every screen. Our team applies the edits inside the instance so the customer keeps using the product without having to learn the editor first.
- Open revision window of two weeks from the load date
- Edits requested by email or shared doc
- Applied by our team directly in the customer's instance
- Turnaround inside two business days on any single request
- Unlimited revision requests inside the window
Phase 07
Handover
The engagement closes and the Brain is yours.
At the end of the revision window the engagement formally closes. Every field of the Brand Brain stays fully editable inside the product, the same way it works for every other tenant. The annual subscription continues. Nothing about the loaded data is locked, gated, or owned by us.
- Final walkthrough of the loaded Brain inside the instance
- Edit access confirmed across every section of the product
- Written handover note covering what was loaded and where
- Annual subscription continues as normal
- No retainer or recurring engagement attached
Concrete artifacts inside the product. Not a slide deck.
Every deliverable lives inside a screen of the product you already pay for. The handover is a working tool, not a PDF the team has to translate into the platform on its own time.
Voice document
Lands in · Brand Studio · Voice
A working voice document drawn from how the brand already writes. Rules, examples, and counter-examples loaded into the brand studio so every draft is scored against it before you see the output.
Audience profiles per vertical
Lands in · Audience
One audience profile per vertical the brand serves. Pains, goals, channel preferences, the language the segment actually uses. Loaded into the Audience page so they can be selected from every brief.
Refusal list
Lands in · Brand Studio · Refusals
Every word, claim, and frame the brand will not publish. Checked against every output before it lands in front of you. Editable later in the same screen.
Proof points
Lands in · Brand Studio · Proof
Customer evidence mapped to the claims the brand makes in market. Cited automatically when a draft references the claim, so the output does not invent supporting numbers.
Sub-brand setup
Lands in · Company · Sub-brands
A sub-brand record per brand under the parent group. Each one gets its own voice split, its own audience overlap with the parent, and the additions to the refusal list that only apply to that sub-brand. Selectable from every brief.
Competitor profiles
Lands in · Competitors
Every approved competitor loaded as a full profile inside the Competitors page. Positioning, voice notes, sample messaging, the claims they lean on. Available from the AI assistant and the brief screen.
Technical context entries
Lands in · Technical Context
Pricing, integrations, regulatory facts, product documentation references — every technical anchor the assistant needs. Loaded so the assistant stops inventing what the product does or how it is sold.
First publishable drafts
Lands in · Create Content
A first round of content generated against the loaded Brain and saved as drafts inside Create Content. Ready to review on the first day the customer logs in.
Brand voice, tone & direction guide
Lands in · Delivered as a portable document
The same Brand Brain that lives inside the product, also delivered as a standalone written guide. Voice rules, tone bands, audience profiles, refusal list, proof points, sub-brand splits, competitor read-outs. A document internal teams and outside agencies can read, share, and write against without ever opening the product.
Up to twelve competitors, profiled and loaded.
The competitor work is the largest single departure from an empty-Brain start. Most teams know who their competitors are, but the file inside the product is empty on day onebecause building it by hand is a slow, manual job that quietly gets pushed off every quarter. When the Brain reasons against an empty competitor set, the drafts read as if the customer is selling into a vacuum.
We profile up to twelve competitors as part of the standard engagement. The list is approved by the customer during discovery, and nobody is profiled who has not been approved. Each profile is built from public material thesame way the voice document is built — the live site, the blog archive, the social presence, any recorded talks, and the press coverage.
What each profile contains
- Positioning statement and our read on their strategic posture
- Voice notes — how they sound when they are not trying
- Sample messaging by channel with citations
- The claims they lean on and the proof they cite for each
- The language they avoid and what that reveals about who they fear losing
- The audience segments they appear to be writing to
Where it lands
The Competitors page of your instance, on day one.
Every profile lives inside the Competitors page. The AI assistant reads from the page when it drafts. The brief screen lets you select any competitor as context for a piece of content. Each profile is editable from the same screen, so a profile that drifts can be updated by the customer at any time after handover.
Profiles loaded
Up to 12
Per profile
~6 sections
Citations
Public sources
Editable after
Always
Customers with more than twelve relevant competitors can scope an addition on the kickoff call. Pricing for additional profiles is quoted at the same time.
Nothing touches the live instance until you sign off.
Approval means a working document, walked through on a call, marked up by the customer, reworked by our team, and confirmed in writing before any data lands inside the tenant. The point of the gate is that the customer reads the Brain in full and recognises the brand before any of it is wired into the place the team uses every day.
The walkthrough call covers every section: voice rules with examples, each audience profile, the refusal list in full, every proof point, every sub-brand split, every competitor profile, every technical context entry. Nothing is presented as a fait accompli. The customermarks what is right, what is wrong, and what is missing from the draft. We rework on the basis of those marks.
Once the marked-up Brain comes back without further changes, that constitutes sign-off. The load week begins the following Monday. If sign-off takes longer than a week, the load week shifts with it; the timeline is the customer's to pace.
How approval works
- 01
Working doc delivered
The full Brain in a shared document at the end of build week two.
- 02
Walkthrough call
Section by section, with our team on the line answering questions live.
- 03
Review window
Up to one week for the customer to mark the document up.
- 04
Rework pass
Our team applies the marks and returns a revised draft.
- 05
Written sign-off
Confirmed by email or in the shared document. Load week scheduled.
Teams who have already paid the price of doing it themselves.
- Companies whose voice lives in twelve heads and zero documents
- Multi-brand groups who want every brand loaded to the same standard on the same week
- Marketing leads whose first attempt at writing the voice doc stalled inside week two
- Operators who have tried other AI tools and bounced off the cold start
- Customers whose competitor file has been on the to-do list for a year
Things this engagement is not.
- —A retainer
- —A monthly engagement
- —A revision-round contract
- —A reason you cannot edit the Brain after handover
- —A gated layer on top of the product
The build runs about five weeks and ends. The annual subscription continues at the standard rate. The Brain belongs to the customer, editable forever, the same as if it had been filled in-house.
$7,500 one-time. No second invoice.
Paid once at kickoff. Sits on top of the annual subscription and does not change its rate. There is no per-revision fee inside the two-week window. Multi-brand groups are priced per brand built and quoted on the kickoff call.
One-time engagement fee | $7,500 Paid once at kickoff. Nothing recurring. |
Annual subscription | Separate Continues at the standard rate (see pricing). |
Revisions inside the window | No charge Two weeks from load date. No per-edit fee. |
Per-brand pricing | Per brand built Multi-brand groups quoted on the kickoff call. |
Engagement length | About 5 weeks Discovery, then two build weeks, then approval, then load week, then two revision weeks. |
The build is more than a year of the subscription because the work of loading a Brain by hand is more than a year of the cost of operating the software around it. The price is not a markup on the software. It is the price of the work.
The questions every operator asks on the kickoff call.
Why does the build cost more than a full year of the subscription?
Because someone on our side is reading the business end to end and writing the entire Brain by hand: voice document, audience profiles, refusal list, proof points, sub-brand splits, the full competitor set, technical context. The subscription pays for software you operate yourself. The build pays for the work of getting the software loaded with the right material before you ever sit down in front of it.
What does discovery actually look like?
Two calls in the first week. The first is an hour on the business as it stands today. The second is an hour on the brand list, the verticals you sell into, and the competitor set you want profiled. Between the calls we want read-access to whatever already exists on paper: the site, the blog archive, the last quarter of social posts, sales decks, customer call recordings if you have them.
How many competitors do you load?
We profile and load up to twelve competitors as part of the standard engagement. If your set is larger, we scope the addition on the kickoff call. The customer signs off on the list before we begin profiling, and we will not load a competitor that has not been approved.
What does each competitor profile contain?
Positioning, voice notes, sample messaging by channel, the claims they lean on, the language they avoid, the audience they appear to target, and our read on their strategic posture. The whole profile lands inside the Competitors page in your instance and is editable from there.
What counts as approval before load?
We send the full Brand Brain as a working document, walk through it together on a call, and you mark what is wrong or missing. We rework anything flagged. Once the marked-up version comes back without further changes, that is approval. Until that moment, nothing has touched the live instance.
What happens if more revisions are needed after the two-week window?
The Brain stays fully editable inside the product after the window closes. Every field, every section, every competitor profile can be changed by the customer at any time. If you would rather have us continue making edits after the window, we can quote a small follow-on block on the spot. There is no automatic retainer.
What if the voice still feels wrong after load?
Flag it during the revision window and our team rewrites the voice document with you. Voice work tends to land closest when the customer reacts to a draft rather than to the document; we expect to iterate once or twice on the basis of a real piece of content generated against the Brain.
How does multi-brand pricing work?
Per brand built. The discovery and approval flow runs once for the group, then the build phases run per brand. We quote the total on the kickoff call once we know how many brands and how many competitors per brand are in scope.
What do you need from the customer during the build?
Access to existing writing, the competitor list, sign-off on the approved Brain, and a short window of attention during the revision weeks to react to drafts. No new writing is required from the customer side.
What happens after handover?
The annual subscription continues at the standard rate. The Brand Brain belongs to the customer, fully editable from inside the product. We are no longer in the room day to day. Standard customer support continues the way it does for every other tenant.
Is the Brain editable forever?
Yes. Every field of the loaded Brain lives inside the same editors every self-serve customer uses. You can change voice rules, rewrite a refusal list, add a sub-brand, retire a competitor profile, or rewrite an audience segment from inside the product at any time.
Is anything about the build gated?
No. The Brain that ships through the build is the same data model that every other tenant fills themselves. None of it requires our involvement to change after handover, none of it is held server-side outside of your tenant, and none of it is licensed back to us.
Hand us the brand. Walk into a loaded instance.
About five weeks from kickoff to handover. $7,500 paid once at kickoff, on top of the annual subscription. Everything loaded into your instance by the time the engagement closes.
Quick recap
- Format
- Done-for-you
- Duration
- About 5 weeks
- Price
- $7,500 one-time
- Output
- Loaded Brain + competitor set + first drafts