Prove the direction before you spend the build budget on it.
The expensive product mistakes I get called into usually start before design. Someone commits engineering quarters to a direction nobody tested, and the bill arrives at launch. I help you test that direction first, in two to four weeks, so the build starts from evidence.
- First step
- 30-minute callNo deck, no pitch
- You leave with
- A go, no-go, or pivot callPlus a written scope, in 2 to 4 weeks
- Fee
- Fixed, in writingAgreed before anything starts
- You work with
- Me, directlyNo account manager
Where I come in
Four situations I get called into
Each one is a decision someone is about to fund. That is when changing direction still costs the least.
01
A funded bet nobody has validated
The roadmap item is approved and the direction came from stakeholder requests, a competitor's screenshots, and three loud customer calls. Nobody has asked whether the thing you are about to build is the thing that gets bought.
02
Two directions and no tiebreaker
There is a real fork in the road, and the internal debate keeps resolving by seniority rather than evidence. Whichever way it goes, you will be defending it to a board a quarter from now.
03
An AI feature with no decided job
Leadership wants AI in the product this year. Nobody has named the decision it makes for the user, what happens when the model is wrong, or why a customer would trust it with work that matters.
04
A product that demos well and stalls in the account
Sales lands it and six weeks later the seats are quiet. Whatever is going wrong happens in the first two weeks of real use, and nobody owns finding it.
How engagements usually start
Start with one decision. Keep me on if it helps.
Start with the smallest engagement that answers the decision in front of you. Adding scope later is easy. Unwinding a large engagement that began before the question was clear is expensive.
01
Entry
AI Delivery Loop Sprint
For the product bet you are about to fund. I test the direction with your stakeholders and users, name what to cut, and leave your team with a scope engineering can start. AI speeds up synthesis, prototyping, and specs where the work is well defined. I make the calls and review everything you receive.
Stakeholder and customer research as needed, AI-assisted synthesis, a go / no-go / pivot call, and a written scope with stated assumptions.
Typically 2–4 weeks · from $8,000 · fixed fee
02
Expansion
Embedded product-experience retainer
For when the need is ongoing. Senior product-experience and UX strategy inside your cadence: roadmap input, research planning, design direction, and stakeholder alignment. A sprint first lets both of us check the fit before a monthly commitment.
Named weekly cadence, decision rights clarity, and continuity across the product org rather than a handoff at the file.
Monthly · three-month minimum
AI handles the mechanical parts, which is how two to four weeks covers this much ground. If a full-time hire would serve you better, I will say so on the call.
Six recurring risks
The risk you are carrying, and what replaces it
People bring me in when a specific risk is sitting on a budget they have to defend. These are the six that come up most.
Expensive uncertainty
You are about to commit engineering quarters to a direction nobody has put in front of a real customer.
OutcomeA go, no-go, or pivot call backed by evidence rather than the most senior opinion in the room.
A roadmap assembled from internal debate
The feature list came from stakeholder requests and competitor screenshots, not from what a customer would change a contract over.
OutcomeA prioritized direction you can defend to your board and hand to engineering without re-litigating it.
Scope that only grows
Every review adds a requirement and nothing ever comes off, because no one agreed what this release was supposed to prove.
OutcomeA scope with a stated hypothesis, explicit cuts, and the metric that will tell you it worked.
Adoption that stalls after the sale
The contract is signed and the seats go quiet. The trail goes cold somewhere in the first weeks of real use.
OutcomeNamed friction in the actual workflow, and specific changes tied to the number you are missing.
Handoffs that turn into rework
Design ships a file, engineering interprets it, and three sprints later what is in staging is not what was decided.
OutcomeStates, edge cases, and system-level components specified, so the build is a build rather than a translation.
The same decision, re-argued by every team
Patterns fork, components duplicate, and each new surface reopens a question that was already settled once.
OutcomeDecisions made centrally and encoded, so teams inherit them instead of negotiating them.
How I work
How the work runs
I use the Double Diamond, a common design method: widen the problem, narrow it, then do the same for the solution. It keeps what we know apart from what we have not proven yet. Every phase ends in a decision, and the documents follow from it.
- Diverge
Discover
What is actually going on?
Interviews with stakeholders and customers, as the question needs, plus a review of what you already have: analytics, support tickets, competitive notes. A sprint uses the parts of this phase your decision needs.
You getA research synthesis, prioritized problem themes, and a clear statement of what we do and do not yet know.
- Converge
Define
Which problem is worth solving?
Journey mapping, opportunity sizing, problem framing, and success metrics defined before anything gets designed.
You getA problem statement, the metrics that will tell us it worked, and an agreed scope. This is where projects get cheaper, because this is where things get cut.
- Diverge
Develop
What is the best way to solve it?
Concept exploration, information architecture, interaction design, and clickable prototypes at the fidelity the decision requires.
You getMultiple viable directions, tested against each other rather than defended in a meeting.
- Converge
Deliver
Will it hold up, and can it be built?
Usability testing, iteration, accessibility review, high-fidelity design, and component and state specification. Building it is a separate engagement.
You getImplementation-ready design with edge cases and states documented, plus test evidence behind the decisions.
The sprint ends with a scope engineering can start. Staying through implementation is retainer or larger-engagement work, not part of the entry sprint fee. A validated direction nobody can build is still a failure mode, which is why the written scope names assumptions and cuts explicitly.
Full methodologyWhy me
What I bring to the decision
Most of the problems I see need all three at once.
- Product experience leadership
- Twenty-five years deciding what the experience should be when the system is complicated and the stakeholders disagree. Most of it inside Fortune 50 product organizations, where being wrong is expensive and slow to undo.
- Product operating model
- A surprising number of product problems turn out to be decision-rights problems. I work on how the call actually gets made, so the direction survives contact with delivery instead of drifting the moment I leave.
- AI-enabled execution
- AI compresses the mechanical parts of synthesis, prototyping, and specification, which is why a small engagement can cover more ground than it used to. The judgment stays mine and the delivery does not get fragile.
I also build and run my own products. Winzinvest is live commercial software for RIAs and family offices, and Foundpath and Casimir Systems are active. Founding things keeps me honest about what shipping actually costs, which is different knowledge from reviewing somebody else's roadmap.
Proof
Six projects, six different problems
Scope and decisions rather than headline numbers. Each one links to the full case study.
There are no client testimonials here yet. Each case study shows the scope I worked in and the decisions I made.
Before anything starts
What working together looks like
- You work with me
- The person on the first call is the person doing the work. No account manager and no junior bench.
- About four hours a week
- Plus one standing working session. That is usually enough if decisions can actually get made.
- One named decision-maker
- A committee that reviews is the most common way these engagements stall. Someone has to be able to say yes.
- Fixed fee, in writing
- 50% on signature, 50% on delivery. Net 15. Scope changes get a written change order before any extra work starts.
Other engagement shapes
When the question is narrower or larger than the entry sprint
The Entry and Expansion cards above cover most buyers. These four shapes apply when the decision on the table is different: a named product cost, a concept that needs testing, a full product build, or a design-system problem.
UX Diagnostic
Something in the product is clearly costing you and nobody can name it.
Expert review, analytics review, stakeholder interviews, prioritized recommendations.
2–3 weeks
Concept Validation
You need evidence behind a direction before engineering commits to it.
Ideation, flows, prototype, usability testing, recommendations.
3–6 weeks
End-to-End Product Engagement
A product or major feature needs experience leadership from the decision through the build.
Discover through delivery: research, strategy, design, testing, implementation support.
8–16 weeks
Design System Acceleration
The same decisions are being re-made by every team that ships a screen.
Component architecture, foundations, accessibility standards, documentation, governance.
4–8 weeks
Investment
AI Delivery Loop Sprints start from $8,000. Most larger product engagements land between $20,000 and $150,000 depending on scope. Embedded retainers are monthly.
Workshops, standalone usability studies, and audits are available when you have one specific question. If that work leads to a larger engagement within 60 days, the entry fee is credited.
Every engagement gets a fixed, scoped proposal with explicit assumptions before anything begins.
Not sure where to start
Product risk and UX maturity assessment
Twenty questions. About six minutes. A maturity score, your three largest product risks, and a recommended starting engagement. No email. Answers stay in this browser.
Take the assessmentIndustries
Where I already know the terrain
Domain familiarity means less of your budget spent explaining your business to me.
Healthcare and health insurance
Regulated workflows, member and provider experiences, claims and benefits, accessibility as a requirement. CVS, Aetna.
Financial services and fintech
Advisor and investor tools, compliance-constrained interfaces, rules-based execution, data-dense screens where a misread costs money. Winzinvest.
Telecom
Multi-product portfolios, large distributed design organizations, system governance at scale. Comcast.
B2B SaaS and enterprise software
Onboarding, activation, admin and permissions, analytics, and workflow tools built for people who use them eight hours a day.
Commercial real estate
Prospecting workflows and map-driven interfaces. Buildout.
AI-native products
Matching, synthesis, agent-assisted workflows, and the interface problems that come with probabilistic systems.
Platforms
Responsive web, native iOS and Android, design systems across product lines, admin tooling, data-heavy dashboards, and AI-assisted interfaces.
Tools
Figma and Figma variables, prototyping at the fidelity the decision needs, accessibility to WCAG 2.2 AA, and AI-augmented workflows including Claude Code and MCP integrations.
Working style
Colorado-based, working remotely with distributed teams across US time zones. On site when a workshop or research round genuinely needs it.
Before a call
Questions that come up before a call
Process, timing, research, ownership, and payment.
Process
How do we start?
A 30-minute call. Bring the problem, not a brief. Most teams start with an AI Delivery Loop Sprint: a fixed engagement that puts evidence under a product bet before the build is funded. If the gap is ongoing senior judgment, we talk about an embedded product-experience retainer instead. If there is a fit, you get a scoped proposal with a fixed price and dates within three business days.
What is not included in the AI Delivery Loop Sprint?
Build and implementation, design-system construction, and ongoing embedded leadership. Those are separate shapes. The sprint delivers the decision artifacts: research as needed, synthesis, a go / no-go / pivot recommendation, and a written scope with stated assumptions.
Do you work alone?
Yes. You work directly with me on every engagement. No account manager, no ramp, and no junior bench learning on your budget. The person you meet on the first call is the person doing the work.
What if we do not know which service we need?
That is normal, and it is what the first call is for. If you would rather do something before talking to anyone, the product risk and UX maturity assessment takes about six minutes and tells you where your risk sits.
Take the six-minute assessmentHow involved does our team need to be?
Roughly four hours a week, plus one standing working session. The largest single predictor of whether an engagement goes well is whether you have one named decision-maker who can say yes, rather than a committee that reviews.
What if the research says we should not build it?
You get that in writing, with the evidence. It is a legitimate outcome, and usually the cheapest one available.
Do you work on site?
Remotely by default, across US time zones. On site when a workshop or a research round genuinely warrants the travel, billed at cost with travel time at half rate.
Timing
How quickly can you start?
Entry engagements can usually begin within two weeks. Larger engagements, three to four. If timing is tight, say so on the first call rather than at proposal.
How long do engagements take?
AI Delivery Loop Sprints are typically two to four weeks. A UX Diagnostic runs two to three weeks. Concept validation is three to six. End-to-end product work is eight to sixteen. Embedded retainers are ongoing with a three-month minimum.
What happens if our side slips?
Some slack is built in. Delays beyond ten business days may shift the schedule, and past that may trigger a re-quote. Everything about this is written into the proposal in advance, so nothing about it is a surprise.
Research
Who recruits the research participants?
Either of us. Recruiting through your existing customers is faster, cheaper, and produces better participants, so I recommend it where possible. Where you do not have a usable list, I recruit through a panel and that cost is quoted separately.
How many participants do we need?
Five interviews usually show a direction. For a decision the board will question, we may need more, across segments. We agree the number before you pay for it.
How do you handle participant privacy?
Written informed consent before every session, in plain language. Consent to be recorded and consent to any published use are asked separately, and participation is never conditioned on the second. Identity is stored separately from responses, findings never identify individuals, and recordings are deleted on a date agreed in advance, with written confirmation.
Can we watch the sessions?
Yes, and you should. One observer per round is built into every engagement. Watching a user struggle with your product changes internal conversations in a way no readout can.
Ownership and payment
Who owns the work?
You do. All deliverables transfer to you on final payment, including working files rather than flattened output. I keep the methods, templates, research instruments, and internal tooling I walked in with. You get a perpetual license to use them as embedded in your deliverables.
Can you show our work in your portfolio?
That is agreed at contract, with confidentiality protected. You control timing, you can require anonymization, and you can ask for anything specific to be removed. If your situation means the answer is no, anonymized use usually works for both of us.
Do we have to give you a testimonial?
No. Nothing in any agreement requires you to say anything about the work.
How does pricing work?
Fixed fee for defined work, monthly for ongoing work, hourly only for approved changes in scope. The number in the proposal is the number on the invoice, and it does not vary with how long the work takes me. AI Delivery Loop Sprints start from $8,000. Most larger product engagements land between $20,000 and $150,000 depending on scope. Embedded retainers are monthly.
Do you work hourly?
Not for project work. I price a defined outcome as a fixed fee. Hourly billing is only for approved changes outside that scope, written and signed before the extra work starts.
How do contracts work?
A scoped proposal with a fixed fee comes first. Smaller first engagements usually combine the proposal and the statement of work. Larger work, or anything going through legal or procurement, uses a master agreement once, then a short statement of work per project. Change orders are written and signed before extra work starts. Deliverables transfer on final payment.
What are the payment terms?
50% on signature, 50% on delivery. Net 15. Monthly engagements billed monthly in advance. Scope changes go through a written change order, priced and approved before any work begins. Participant incentives, recruiting, and research tooling are billed at cost with no markup, estimated in the proposal, and anything above $500 is pre-approved in writing.
Bring me the bet you are about to fund
Thirty minutes, no deck. Tell me what you are about to commit to and which part is still unproven. I will tell you what I would test first and what shape of work fits, if any. If there is a fit, you get a fixed-fee proposal within three business days.
The form comes straight to me. I reply within a day and we find a time.
Prefer email? ryan@winzenburg.comEntry engagements can usually begin within two weeks.