Proposal Studio

Most companies don't have an AI problem. They have a work problem.

Value gets trapped in manual work, handoffs nobody owns, and systems that made sense in 2015. I find where it's trapped, put a number on it, and build the first fix so you can see it working before you commit to anything bigger.

20+ years of building and running digital businesses. This is the page where I show you how I'd approach yours — the engagement shapes, the documents you get, and the problems I keep being called in for.

Start with a diagnosticSee the engagement models
Years, since 2005
20+
Organisations worked with
1,000+
AI agents in production
72
Acquisitions
4+

What I actually do

Five moves. In this order. Every time.

Not a methodology with a trademark. This is just the sequence that stops people spending six figures on the wrong problem.

01

Find where the money leaks

I follow the work, not the org chart. Sit with the people doing it, watch the handoffs, read the spreadsheets everyone pretends are temporary. The drag is almost never where leadership thinks it is.

02

Put a number on it

Hours, headcount, error rate, days of delay, revenue sitting in a queue. If I can't cost it, I don't claim it. You need a figure you can defend in a board meeting, not a feeling.

03

Say what isn't worth fixing

Most of it isn't. I rank by value released against effort to change, and I tell you which items to leave alone. A shorter list is the whole point.

04

Build the first one

Working software or a working agent, in your environment, on your real data. Weeks, not quarters. You get evidence it works here before you commit to anything bigger.

05

Hand it over properly

Your team runs it. Documentation, training, and the freedom to change it without ringing me. If a system only works while I'm in the building, I haven't finished.

What gets measurably better

Cost out

Work that stops being done by hand

Capacity back

Hours returned to people you already pay

Cycle time

Days removed from the path to done

Risk down

Fewer manual steps, fewer ways to be wrong

Optionality

Things you can do now that you couldn't before

If an engagement can't move at least two of those, I'll say so and we can both spend the money somewhere better.

Engagement models

Four ways in. Pick the one that matches how certain you are.

Most people should start at the first one. Everything after it is easier to scope, cheaper to run, and far more likely to land once the diagnostic exists.

Start here

Discovery / Diagnostic

2–3 weeks. I map how work actually flows, cost the drag, and hand back a ranked list of what to fix first.

Right for you if

You know something is wrong. You can't yet name it, price it, or defend a budget for it.

Duration
2–3 weeks
Input
6–10 interviews, your systems, your numbers
Commitment
One engagement. No lock-in after it.
  • Interviews with the people doing the work, not just the people describing it
  • A walk through your actual systems and where data stops moving
  • Cost model for the current way of working
  • Ranked opportunity map with a recommended first project
  • 90-minute readout with your leadership team

You end up holding

A costed, ranked view of where value is trapped — and one first project specific enough to approve.

Start with a diagnostic
Prove it here

Strategic Pilot

6–10 weeks. One problem off the list, built and running in production. Not a proof of concept that dies in a demo.

Right for you if

You've agreed what's broken. Now you need evidence it can be fixed with your data, your rules and your people.

Duration
6–10 weeks
Scope
One workflow, end to end
Success
Agreed in writing before we start
  • Success criteria and a measured baseline agreed up front
  • Build in your environment, integrated with the tools you already pay for
  • Weekly working demo — no monthly status theatre
  • Adoption plan for the people whose job changes
  • Before-and-after numbers, published whichever way they fall

You end up holding

One workflow measurably faster or cheaper, live in your business, owned by your team.

Scope a pilot
Ongoing

Retained Advisory

Monthly. I'm the senior voice in the room when the decisions get expensive and everyone in the meeting is guessing.

Right for you if

You've got people who can build. You need someone who has watched this go wrong before and will say so early.

Cadence
Monthly, capped days
Access
Direct line between sessions
Term
Rolling, 3-month minimum
  • Architecture and build-versus-buy decisions, argued properly
  • Review of what your team and your vendors are actually shipping
  • Roadmap kept honest against cash and capacity
  • Vendor and agency pressure-testing before you sign
  • A named person your team can escalate to

You end up holding

Better decisions, made faster, with fewer expensive reversals six months later.

Talk about a retainer
Deepest

Embedded Partner

I take an operating role for a defined stretch. Accountable for the outcome, not just the advice.

Right for you if

The change is real, the timeline is fixed, and nobody internally has the capacity to own it without dropping their day job.

Commitment
Up to 2 days a week
Term
3–12 months, defined end
Role
Named owner of the programme
  • Programme owned end to end, with your name on the plan and mine on the delivery
  • Sits inside your leadership rhythm — your stand-ups, your board pack
  • Builds alongside your team so capability stays when I leave
  • Hiring and supplier decisions made with you, not for you
  • A written exit plan from week one

You end up holding

A programme with an owner, a plan, and someone whose reputation is attached to it landing.

Discuss an embedded role

Fixed fee, agreed before we start. Scoped to the engagement, not to hours, so nobody's watching a clock. Published day and retainer rates live on the Work with me page. If what you need is smaller than what you asked for, I'll tell you that before you sign anything.

Artifacts

What you actually get handed.

Every engagement produces defined documents, not a vague sense that something happened. These are the ones that come up most. Real client versions stay private — nothing from a live engagement is ever published here.

Diagnostic output

Opportunity Map

Every workflow in scope plotted by value released against effort to change. The sequencing argument, on a single page a board can read in two minutes.

  • Value and effort scored, with the scoring shown
  • Dependencies marked so nothing starts out of order
  • An explicit "leave this alone" list

One page

Diagnostic output

Cost-of-Drag Model

What the current way of working costs in hours, headcount, rework and delay. The number your finance director will interrogate, built so it survives that.

  • Assumptions listed separately so they can be challenged
  • Sensitivity range, not a single flattering figure
  • Traced back to observed work, not benchmarks

Spreadsheet + one-page summary

Design output

Workflow Redesign Blueprint

Current state, target state, and the specific steps that disappear. Written so an ops lead can build from it without a translation meeting.

  • Step-by-step current process, including the undocumented parts
  • Target process with the removed steps called out
  • Exceptions and edge cases, because those are what break rollouts

Document + diagrams

Design output

AI Opportunity Assessment

Where AI genuinely earns its place, where it's a liability, and where a database and a rule would do the job better and cheaper. Anti-hype by design.

  • Use cases ranked by value, not by novelty
  • Data, privacy and approval constraints stated plainly
  • Running cost estimated before anything is built

Document

Delivery output

Pilot Design Brief

One problem, scoped to weeks. Success criteria agreed before a line is written so nobody relitigates the definition of done in week nine.

  • Baseline measured before the build starts
  • Explicit out-of-scope list
  • Who owns it on day one after go-live

Short document

Design output

Operating Model Recommendation

Who owns what once the change lands. Roles, decision rights, escalation paths. The part most programmes skip, and the reason most of them quietly revert.

  • Decision rights written down, not implied
  • Named owners for every system that survives
  • What to stop doing, and who stops doing it

Document + role map

Strategy output

Transformation Roadmap

12–24 months, sequenced by dependency and by cash. Honest about what has to go first and what can wait until next year.

  • Phased against capacity you actually have
  • Cost and benefit staged per phase
  • Kill criteria for each phase, agreed in advance

Deck + one-page plan

Commercial output

Executive Proposal

The commercial case, written in the language of the person signing it. Delivered as a private link rather than a 40MB PDF nobody can open on a phone.

  • Situation, options and recommendation in that order
  • Commercials on one screen, no buried assumptions
  • Questions answered in the document, not in a follow-up call

Private web link

Proposals themselves arrive the same way — a private, tracked web page rather than an attachment, so the commercials stay on one screen and you can forward it to the person who signs without breaking the formatting.

Problems I solve

If one of these is your sentence, we should talk.

These are the six I get called in for most. The quote is what the first conversation sounds like. The paragraph under it is what I usually find.

Manual work and process drag

“We've always done it this way. It takes three people and a spreadsheet.”

What it usually turns out to be

Usually the process is fine and the handoffs are the problem. Work sits in inboxes waiting for a human to copy it somewhere else. Removing the copying is often the entire fix.

Overloaded teams, invisible bottlenecks

“Everyone's flat out and nothing's shipping.”

What it usually turns out to be

Capacity is rarely the constraint. One or two people are the single point of approval for everything, and the queue in front of them is invisible because it lives in chat.

Scaling by headcount instead of systems

“The answer to more volume keeps being another hire.”

What it usually turns out to be

Cost per unit of work never falls, so growth makes the margin worse. The fix is boring: standardise the work first, then automate the standard version. Automating chaos just makes faster chaos.

AI enthusiasm with nothing in production

“We've got eleven pilots and none of them are live.”

What it usually turns out to be

Pilots were picked for how interesting they were, not for whether anyone would have to change their day. Pick one workflow with a named owner and ship that. Interest isn't a business case.

Approvals, compliance and content grinding slowly

“Legal takes two weeks and nobody can tell me why.”

What it usually turns out to be

The delay is usually queueing, not reviewing. Actual review time is minutes. Making the queue visible and pre-structuring the submission removes most of the two weeks without touching the policy.

Operational readiness before growth or exit

“Diligence is coming and our processes won't survive it.”

What it usually turns out to be

Undocumented process is a valuation discount. The work is making the operation legible — who does what, on what system, with what evidence — and that work is worth doing whether or not you sell.

Why me

Twenty years of shipping, not twenty years of advising.

I run my own ventures on the same operating layer I build for clients. When something I recommend is painful to live with, I find out first.

Years building digital, since 2005
20+
Organisations worked with
1,000+
Acquisitions
4+
Domains
850+
AI agents in production
72
Web apps shipped
80+
Base44 apps built
375+
The ventures behind it

Who I've worked with

Over 1,000 organisations across sport, charity and fundraising, healthcare supply, travel, media and founder-led scale-ups — plus my own portfolio of ventures. Some of it under my own name, some of it through businesses that have since been acquired.

I don't publish client results, named or anonymised. Commercial numbers belong to the people who earned them. Ask me in the room and I'll make introductions to people who will tell you what it was like.

What this is not

  • —Not a deck factory. You get a working thing, not a hundred slides about the working thing.
  • —Not a reseller. I don't have a vendor quota or a licence kickback, so I'll tell you when the cheap tool is the right tool.
  • —Not an AI evangelist. Plenty of the problems I find are fixed with a form, a rule and someone owning it.
  • —Not a body shop. If your team can do it, my job is to make sure they can, then leave.
AI Advisory

The build side — agents, integrations, the tooling itself.

Ventures

The businesses I run, where this way of working was learned.

Work with me

Published rates and the smaller ways to start.

Straight answers

The questions I get asked first.

What does a diagnostic actually involve?

2–3 weeks. Six to ten interviews with the people doing the work, a walk through your systems, and a look at your own numbers. You get a cost model for the current way of working, a ranked opportunity map, and one recommended first project — plus a 90-minute readout with your leadership team.

How long before we see something actually working?

A strategic pilot runs 6–10 weeks and ends with one workflow live in production, on your data, owned by your team. Success criteria and a measured baseline are agreed in writing before the build starts.

Do you build it, or just advise?

Both, and I prefer building. 20+ years of shipping digital products, 80+ web apps and 72 agents running in production. Advice that has never survived contact with a deployment isn't worth much.

We bought an AI tool and nobody uses it. Can you help?

Yes, and it's common. Adoption failure is usually a workflow problem rather than a tool problem — the tool asks people to change their day and nothing made that change easier. I look at the work first and the licence second.

Who owns what you build?

You do. It runs on your infrastructure, with your accounts and your data, documented so your team can change it without me. If a system only works while I'm in the building, the engagement isn't finished.

What does it cost?

A fixed fee, agreed before anything starts, scoped to the engagement rather than to hours. Published day and retainer rates are on the Work with me page, and I'll tell you early if what you need is smaller than what you asked for.

Next step

Send me the messy version.

You don't need a brief. A paragraph about what's slow, what keeps breaking, or what you're about to throw headcount at is enough to have a useful first conversation. If it's not something I should take on, I'll say so in the reply and point you at someone better.

Start a diagnosticdean@deanlynn.comLinkedIn

Keep it short. Voice notes under two minutes are welcome.