Touchpoint strategy
14 tasks, force-ranked. Click to jump.
UD — user-driven · designer
BD — business-driven · designer
Touchpoint Strategy
Task 1 [UD] · rank 1 — Understand my role vs. other roles.
Arriving: confused, worried, defensive → leaving: confident they can still add value.
Approach: name the fear before answering it — a defensive reader treats unearned optimism as a sales pitch.
Is AI killing design?

Yes. And also no.

Half of what we've been calling design is being automated — the wireframes, the mockups, the prototypes, the specs. That half is going, and pretending otherwise won't save it.

The other half — understanding humans, defining the problem, prioritising the tasks, providing the context — has never mattered more. It's also the half almost nobody has been spending their time on.

The whole industry is sprinting head-first into the half that's being automated.

Tasks 1 & 3 [UD/BD] · ranks 1, 3confused/defensive → confident they add value; and → validated and empowered. The model does the work: let the designer locate themselves on it and see which half is which.
This is design

You've been living in the wrong half

For the last decade or two, design has been defined by its outputs. Journey maps, storyboards, wireframes, prototypes, specs — we've made thousands of them. They're the currency of the profession.

But every one of those is a definition of the solution. And before any of them can be made well, a designer needs four layers of information about the opportunity.

Opportunity spaceFour layers of context — written, debated and refined by a human.
Solution spaceThe artefacts. Given the right context, AI produces these faster and cheaper than any human team.
Who

Who are we designing for? Demographics, psychographics — often represented as personas.

Why

Why is that audience using this product or feature — what's their intent? And what does the business want?

What

What is the customer doing? What information do they need, what decision are they making — and how will we know if it went well?Touchpoint strategy

How

How should it be designed? Principles, design system, brand identity, positioning, technical constraints, cultural context.

Journey mapsThe end-to-end walkthrough.
StoryboardsThe scene by scene.
Wireframes & mockupsThe screens.
PrototypesThe clickable version.
Specs & redlinesThe handoff documents.
Alive — and now the whole job
This is the half that's dead
Where our time has actually gone
10%
90%
Opportunity spaceSolution space
Task 2 [BD] · rank 2no arriving emotion → leaving reassured that AI isn't good at everything. Present capability as a boundary, not a hype list.
AI isn't omnipotent

AI can't do this half — and it needs you to

One — AI is bad at it

Models struggle to understand real human need and to prioritise according to human desires. They can generate a hundred layouts; they cannot tell you which task matters most to a frightened person at 11pm, or that the brief you handed them is wrong.

Two — AI needs it more than you did

Human designers could hold the opportunity space implicitly — in their heads, in hallway conversation, in judgement built over years. An AI model has none of that. It needs the context written down, explicitly, or it produces generic slop.

AI doesn't need designers to help it build faster.
It needs designers who can teach it about humans.
Task 4 [BD] · rank 4→ leaving excited by the possibilities. Excitement comes from specificity. Emotional peak of the designer path — names the new deliverable, which task 5 then shows.
The half that's growing

Your output changed

The Figma file is not your output any more. The wireframe is not your output any more. The spec is not your output any more.

Your output is the touchpoint strategy document. One per feature: the tasks that matter, force-ranked, with the emotional state behind each one and the measure that says whether it worked. Written by a human, debated by a human, refined by a human — and then handed to the machine that builds from it.

Task 5 [BD] · rank 5no arriving emotion → leaving clear on what they'd actually produce, and that it's a document they could write on Monday rather than a new discipline. Show the artefact itself before the demo shows what it produces. Added because the site was introducing a methodology without ever showing its output.
The new deliverable

So what is a touchpoint strategy?

One document, per feature. It replaces the wireframe set, and it's shorter than the spec it retires. Here are five rows from a real one — a pet store product page.

#
The task
They're feeling
Design & content approach
Success criteria
3
Find out how much the item costsUSER
Show prices for the item, including any configurations (size, colour). Possibly include a currency converter for Canadian customers.
Lack of price information accounts for <5% of missed sales
4
See if other people like the itemUSER
Mistrustful that the reviews shown are genuinely unbiased
Show customer ratings and reviews. Let reviewers tag their review with pros and cons, and attach images of their own pets using the item.
Lack of reviews accounts for <5% of missed sales
5
Find out shipping costs and timesUSER
Urgency — they may need the item quickly
Show shipping costs and delivery times.
Missing shipping info accounts for <5% of missed sales
6
Compare the item to others like itUSER
Fearful or unsure they've chosen the best product for their needs
Show recently viewed items and provide access to a "compare to similar items" tool.
20:1 ratio of in-page navigation vs. pogo-sticking back to the gallery
7
Believe the site is safe and secureBUSINESS
Nervous about buying online with a credit card
Show security certificates. Use language throughout that supports a safe online shopping experience.
Security concerns account for <5% of missed sales
Twelve rows in total — that's the whole document. These five are the ones you'll see at work further down the page.
One ranked listUser-driven and business-driven tasks force-ranked together, no ties — the only permitted tie is the same task appearing on both sides. The debate that produces the order is the method.
Emotion, not just taskPeople aren't machines working through a checklist. Two shoppers can both want to read the reviews — one curious, one suspecting they're fake. Same task, different design. The emotion is what tells you which.
Success criteria per taskA measure for every task, not one number for the whole page — and written before the screen exists, so it can't be invented afterwards to flatter the work.
Task 6 [UD] · rank 6Arriving apprehensive, skeptical → leaving confident and eager to try for real. Skepticism dies on contact with a working example, not an argument. Highest-leverage asset on the site.
Proof

Same model. Same product. Same prompt.

I gave Claude a touchpoint strategy document for a pet store product page and said, in plain English: "Use this as a design brief." Then I asked for an alternate version and told it specifically not to use the document.

Without the document
PetstoreShopDogsCatsBirdsFishSearch…Cart (0)
Home / Dogs / Beds / Orthopedic Dog Bed
Orthopedic Dog Bed
SKU: PET-BED-001
$89.99
A comfortable orthopedic dog bed with memory foam support. Removable, washable cover.
Add to Cart
Specifications
With the document
🐾 PetstoreDogsCatsBirdsDealsSearch…Cart (0)
Home / Dogs / Beds / Premium Orthopedic Dog Bed
Premium Orthopedic Dog Bed with Memory Foam
★★★★☆ 4.6 · 47 reviews ✓ verified purchases
$89.99 $119.99
Price includes taxes · 🇨🇦 View prices in CAD
4-inch memory foam that conforms to your pet's body, relieving pressure on joints.
🚚 Free delivery by Thursday — order within 2 hrs 14 mins
⇄ Compare with 3 similar beds · Recently viewed
Add to Cart
🔒 SSL secure · PCI compliant · BBB A+ rated
3Tax and currency note — find out how much it costs, and the Canadian currency question logged in the comments column.
4Verified-purchase reviews — see if other people like the item, arriving mistrustful the reviews are unbiased.
5Delivery date and countdown — shipping costs and times, with urgency as the emotion.
6Compare and recently viewed — compare the item to others, and its 20:1 in-page navigation measure.
7Trust marks — believe the site is safe and secure, where shoppers are nervous about paying online.

The version on the left is competent and generic. Everything on the right traces to a row of the document, and none of it was assumed or invented by the machine.

Task 7 [UD] · rank 7 — largely transactional, no emotional charge. Not gated. Ranks well above the email capture (task 13); gating would invert the force-rank and contradict task 13's exit emotion. Ship a filled example alongside the blank.
Take it away

The template

Touchpoint strategy template

The blank document. Tasks force-ranked, emotional state, design and content approach, success criteria.

Download .docx →

A completed example

The pet store product page, filled in end to end. Start here — a blank template is intimidating.

Download .docx →
Task 8 [UD] · rank 8Arriving apprehensive → leaving confident. Confidence comes from one completed rep, not from reading. Make the first attempt small and safe.

Try one, small

Don't start with your whole product. Take one feature and make four short passes over it. The whole thing is an hour.

1 · Tasks. List every task the feature has to serve — the ones the user came for, and the ones the business wants. Then force-rank them into a single list. Do it on post-its: put one up, add a second and ask "more or less important than that?", then a third. No ties — the only permitted tie is the identical task appearing on both the user and business sides. That debate is the method.

2 · Emotions. For each task, write the state the person arrives in and the state they should leave in.

3 · Approach. Then how — the design and content decisions that carry each task from its arriving state to its exit state.

4 · Success criteria. Last, one measure per task: the single thing you'd watch to know it worked.

Four passes over the same short list. That's the whole document.

Tasks 9 & 13 [BD] · ranks 9, 13→ confident Richard can help; confident it's worth the time. Invite me to do a talk (task 9), placed after the practice block. Register-interest framing keeps it low-commitment. The email capture doubles as task 13 (give me their email). Workshops (task 10) held for a later version.
If this is landing

Want a talk for your team?

Sixty minutes that gives a whole team the same language for what AI actually changes about design work — where the value moves, not which tools to learn. Register your interest and I'll follow up with formats and dates.

Task 11 [BD] · rank 11→ feeling included. Framed as co-development. Inline and per-section, not a global contact form.

How would you make this better?

This framework is still being built, and the corrections that have mattered all came from people running it against a real product. If something here doesn't hold up against yours, tell me.

Task 5 [BD] · rank 5→ clear on what a touchpoint strategy document actually is. Also served here: the whole page is a live TSD, and turning the mapping on shows the document applied to something real — which is the clearest way to grasp what one is.
About this page

This page was designed from its own touchpoint strategy document. Every section maps to a task and the emotional state it's meant to serve — turn the mapping on to see it, or download the full TSD.