brBuilt Revenue

Brian Kennett / Selected work

Built around
the work.

I spent two decades running revenue in media. Now I advise teams on how that work should run, and I build the software it runs on: the decisions, handoffs, and everyday work that keep a business moving.

A closer look at what I’ve been building at Built Revenue.

01
Local Media DeskFrom advertiser research to a signed order
↓02
Model DeskThe P&L and the drivers behind it
↓03
Built Revenue AgentAn assistant for the whole working day
↓04
SeichePublic records into useful local reporting
↓05
SABLEAn agency’s working record, ready to answer and draft
↓06
Lead to Cash Bash HubAn event platform from invitation to follow-up
↓
Two decades

Building, growing, and leading revenue operations in media.

12+ years, full P&L

Leading publisher-owned ad agencies, including one of the country’s top ones in revenue, profitability, and client retention.

Both sides of the buy

I’ve sold local media, bought it for a large regional health system, and run performance channels.

Online since 1989

Every generation of digital tools since my first online community. AI is the newest tool in that practice, not the beginning of it.

01 / Local Media Desk

Keep the seller
in the conversation.

The advertiser’s goal carries from first research to signed order.

Local Media Desk is a ground-up rebuild, shaped by what I learned building Compass and designed from the start to keep each customer’s data separate and secure. It carries the work from researching the business to shaping a brief, exploring a plan, and turning it into something a buyer can understand and act on, instead of asking sellers to assemble the story across disconnected tools. Acceptance, e-signature, and order handoff carry the plan into delivery.

Explore a sample proposal ↗View sample reporting ↗
Local Media Desk / The workflowProduct workflow illustration
RESEARCH

Know the business.

Turn business and market context into a useful starting point for the conversation.

BRIEF & PLAN

Make the choices clear.

Connect objectives, audiences, and the media plan. Explore options with the buyer.

PROPOSAL

Give the plan a home.

A shared, interactive proposal carries the recommendation through buyer review, acceptance, and e-signature.

REPORTING

Close the loop.

Bring performance back to the business goal, then explore the next campaign’s budget and channel mix.

Work from the assistant you already use.Claude connects today. Ask it to research an account or build the proposal from the plan, approve once, and the proposal goes to the buyer, with the same records and permissions as the desk.

Connected assistant

Research built for a local market.

Business, geography, audience, and market evidence shape the brief and the recommendation—not just a generic company summary.

See the AI-search picture.

When buyers ask AI search about a category, does the advertiser show up, and what does the answer say? AI Visibility checks Google AI Mode and ChatGPT Search, compares competitors, and turns the findings into next steps a seller can bring to the client.

Practice the hard part first.

Practice the conversation by voice or text before the buyer hears it. An AI buyer and coach work from the account’s own plan, and a short debrief follows. It’s practice, not evaluation: nothing scores the seller.

Workflow first

The brief, plan, proposal, and report share a coherent business story.

Alongside the CRM

A seller’s working layer can complement the existing system of record.

Useful beyond one screen

The desk and a connected assistant work from the same records and the same rules.

02 / Model Desk

Understand what
drives the business.

The whole operating model, from the P&L down to the drivers underneath it, in a form people can explore.

Model Desk takes the financial model a business already runs on and turns it into an interactive application. Follow the P&L back to the assumptions and drivers that shape it, change one, compare scenarios, and see what moves and why. The software is built from the existing workbook and checked against it, so the numbers stay the team’s numbers. Share a scenario, and ask questions against the model’s financial views.

Open the Northwind demo ↗About Model Desk ↗
Change an assumption. See the consequence.Illustrative calculator · synthetic data

A small example of the interaction pattern. Starting revenue: $5m. Operating cost: $4m. The live demo contains the fuller financial model.

Revenue$5.60m
Operating profit$1.36m
Margin24.3%
A repeatable refresh

Each new version of the workbook is checked against the original the same way, so the application stays connected to the model the team maintains.

Put the drivers in view

The assumptions that move the P&L sit in front of the people making the decision.

Separate math from narrative

A computation engine owns the numbers. AI can help explain what changed.

03 / Built Revenue Agent

An assistant for the
whole working day.

Walk into every meeting knowing what was decided and what I owe.

Built Revenue Agent is the connected working environment behind Pepper, my day-to-day assistant. It brings mail, calendar, meeting memory, commitments, briefs, and prepared work into the same conversation. I can ask what needs attention, prepare for a meeting, pick up an earlier decision, or see what I owe someone—by typing or talking.

Pepper, the Built Revenue Agent assistant, preparing a fictional meeting brief with upcoming meetings, drafts, and commitments alongside the conversation.
Built Revenue Agent / Pepper · Actual application interface with a fictional conversation, appointments, drafts, and commitments.

A clearer start to the day.

Scheduled briefings bring mail, tasks, calendar context, and open commitments together to surface what needs attention and what can move next.

Carry the conversation forward.

Record meeting audio and visual context, retain key screen images, create speaker-attributed transcripts, and turn the discussion into briefs, commitments, and follow-up drafts.

Keep track of the promises.

See what I owe, what I’m waiting on, and what needs a reminder. End-of-day review carries open work into the next day with a clear disposition.

One part of the larger system

From conversation to shipped work.

When a conversation turns into a change to the software, Pepper hands it to a coding agent, picks the right model for the job, and brings the finished work back for review.

Pep Talks · morale, with a raised eyebrow A slightly less serious capability. Open for encouragement.

Pepper has range.

Sometimes the next action
is a well-timed card.

Pep Talks turns a moment into a designed greeting card, with an editable personal note and a choice of tone. Pepper can suggest a card, render the artwork, prepare the email, and remember what has already been sent. Apparently the assistant also has opinions about my project plans.

Pep Talk card: I believe in you. My opinion of the project plan is currently under review.

Actual Pep Talks card renderer. Example card only; no recipient or private correspondence.

See the schedule · application screenshot with sample dataBuilt Revenue Agent Schedule showing fictional product reviews and planned work across a sample week.

Actual calendar renderer and application styling. All appointments and planned work in this image are fictional.

Context that carries forward

Source-backed meeting and project memory supports briefs, commitments, and the next piece of work.

Ready for review

Finished work comes back with its context, so the review starts where the work left off.

One place to run the day

Briefs, commitments, schedule, and finished work for review share a common home.

04 / Seiche · formerly Civic Desk

Follow the matter.
Keep the sources.

A local story rarely begins and ends with one meeting.

Seiche follows public matters across agendas, minutes, and recordings, then helps turn that material into resident-facing briefs. The editorial desk and resident experience share a persistent record of the matter, with claims tied back to sources. Residents can follow the history, search published material, and ask source-grounded questions. In St. Croix County, Wisconsin, it also publishes a voter guide: the races on the ballot, the candidates, and the official source behind every fact.

01Build the recordConnect documents and meetings around the same matter.
02Write from evidenceDraft a useful account of what happened and why it matters.
03Verify the claimsSeparate source checking from the writing step.
Explore Seiche ↗See the voter guide ↗

05 / SABLE · with Indigo Trigger

Put project memory
where work happens.

An agency’s whole working record, ready to answer and draft.

Built in partnership with Indigo Trigger, SABLE is the operating platform for an agency’s client work. It brings meetings, Asana tasks, Toggl time, Drive documents, email, and CRM context into one record, organized by account and project. Staff ask scoped questions, prepare weekly client updates, flag scope movement, and estimate statements of work from what similar projects actually took. AI drafts; a person approves before anything reaches a client.

01Connect the working recordMeeting, project, time, document, and CRM context.
02Move from context to actionDecisions, promises, task proposals, client updates, scope flags, and draft deliverables.
03Work with the customer in the same systemA client portal connects the shared plan, requests, deliverable reviews, and released resources.
Explore SABLE ↗

Connected tools / Practical AI economics

Make the intelligence
fit the work.

AI choices, workflow design, and the customer experience belong in the same product decision.

Customers shouldn’t have to change how they work to get value from AI. The products work in their own screens and inside the assistants people already use, and the records and business logic stay the same either way.

Give customers a choice.

Bring your own AI account, use one Built Revenue provides, or connect the assistant you already pay for: whatever fits how a customer’s team works and buys.

Pay for intelligence where it counts.

Each task gets the right model for the job, so AI cost stays in proportion to the work and becomes something the business can plan.

Keep the math in code.

Tested calculation engines handle financial and planning arithmetic. Models help interpret, write, and reason about the result. The Northwind demo even answers supported model questions locally in the browser. The customer gets a useful workflow with understandable economics.

How I build

The workflow
is the starting point.

The throughline is taking a process I understand, finding where it loses context or momentum, and building software that makes the next step easier.

Start with the handoff.

What does the next person need to know, decide, or do? That question shapes the interface and the data model.

Keep logic explicit.

Financial math belongs in a calculation engine. Source relationships belong in the record. AI works with that structure.

Make tools composable.

The same capability can serve a screen, a connected assistant, or a background workflow without rebuilding the product for each one.

Build the whole loop.

Inputs, useful output, review, and the next action belong to the same product experience.