🎮 The Next Input — Issue #207

In partnership with

⚡ The Briefing — 60 sec

• OpenAI acquires presentation startup NextSlide Maybe now their design skills won’t actually suck. Jokes aside, presentations are one of those deceptively hard AI problems: everyone can generate 20 slides, but generating 20 slides that don’t look like a consulting intern discovered gradients yesterday? Different story.

• North Korean hackers are reportedly using AI for cyberattacks AI lowering the barrier to sophisticated cyber operations was always coming. The uncomfortable bit is that defenders and attackers get access to essentially the same capability curve. Welcome to the arms race.

• AI avatar to deliver lectures for University of Newcastle subject I mean… at this point why even go and get a degree? Slightly facetious, obviously, but universities are going to have to answer a serious question: what exactly are students paying for when the lecturer, content and increasingly the tutoring layer can all be generated?

🛠️ The Playbook — AI Value Chain Audit

Mission
Identify exactly where humans still create differentiated value in an AI-heavy workflow—and redesign everything else around it.

Difficulty
Intermediate

Build time
3–5 hours

ROI
Removes low-value work, clarifies where human expertise matters and prevents organisations from automating the wrong parts of the experience.

0) Why This Matters

AI is getting remarkably good at producing the thing.

Slides.

Lectures.

Code.

Reports.

Analysis.

Customer responses.

That forces organisations to ask a much harder question:

What were people actually paying us for?

If the answer was simply “someone creates the PowerPoint” or “someone delivers the lecture,” you may have a problem.

But most valuable services contain layers AI cannot replace merely by generating the output:

• judgement

• trust

• accountability

• contextual expertise

• relationships

• experience design

• interpretation

• intervention when things go wrong

The opportunity is to automate production while making the genuinely differentiated parts substantially better.

1) Architecture

2) Workflow

1. Select one service, product or internal workflow being affected by AI.

2. Map every step from initial request through final outcome.

3. Label each step as production, judgement, relationship, accountability or administration.

4. Test AI against production and administration tasks first.

5. Redesign human roles around judgement, intervention and differentiated expertise.

6. Measure whether the new experience is actually better—not merely cheaper.

3) Example Prompts

Value Chain Decomposition

You are an AI operating-model strategist.

Analyse the following service or workflow:

[DESCRIPTION]

Break every stage into:
- administration
- content production
- analysis
- judgement
- relationship building
- accountability
- specialised expertise
- exception handling

For each stage provide:
1. current human contribution
2. AI automation potential
3. AI augmentation potential
4. risk of removing the human
5. recommended future-state owner

Do not assume automation is desirable simply because it is technically possible.

Human Value Test

Review the following role in an AI-enabled workflow:

[ROLE]

Identify the activities where the human provides value through:
- judgement
- trust
- accountability
- domain expertise
- empathy
- negotiation
- contextual interpretation
- exception handling

Then identify activities where the person is primarily functioning as an information processor.

Recommend how the role should evolve.

Experience Redesign

Redesign the following customer or student experience assuming AI can produce most routine content instantly:

[EXPERIENCE]

The redesigned experience must:
- reduce repetitive production work
- increase access to expert human judgement
- provide personalised AI assistance
- clearly disclose AI involvement
- preserve accountability
- improve the outcome rather than merely reduce cost

Return the current state, future state and measurable differences.

4) Guardrails

• Never equate “AI can perform the task” with “AI should own the outcome.”

• Preserve named human accountability for consequential decisions.

• Tell users when meaningful parts of an experience are AI-generated.

• Measure outcome quality alongside cost reduction.

• Protect expert knowledge before automating the roles containing it.

• Create clear escalation routes when automated experiences fail.

• Test whether customers actually value the redesigned experience.

• Avoid charging premium-human prices for silently automated commodity delivery.

5) Pilot Rollout — 3 hours

1. Choose one service or workflow already experiencing AI disruption.

2. Map every step and identify where time and money are currently spent.

3. Classify each step by production, judgement, relationship or accountability.

4. Automate one high-volume production task while preserving human oversight.

5. Redirect the saved capacity into one higher-value human interaction.

6. Compare cost, quality and user satisfaction against the original workflow.

6) Metrics

• Percentage of workflow automated

• Human hours redirected to higher-value activity

• Cost per completed outcome

• Customer or user satisfaction

• Escalation rate

• Human intervention success rate

• Output quality

• Time-to-completion

• Expert utilisation

• Percentage of users aware of AI involvement

Pro Tip: When AI can generate the deliverable, your moat becomes everything surrounding the deliverable.

🎯 The Arsenal — Tools & Platforms

• Miro · maps where value, judgement and administrative work actually occur · Link

• Airtable · maintains task, role and automation-opportunity inventories · Link

• Azure AI Search · gives AI workflows access to approved expert knowledge · Link

• Microsoft Teams · routes exceptions and high-value interactions back to humans · Link

• Power BI · measures whether automation improves outcomes instead of merely reducing labour · Link

Copy-paste prompt block:

You are an AI operating-model architect.

Analyse the following business, service or team:

[DESCRIPTION]

Customers or users:
[LIST]

Current deliverables:
[LIST]

Roles involved:
[LIST]

AI tools available:
[LIST]

High-risk decisions:
[LIST]

Break the operating model into:
- administration
- production
- analysis
- judgement
- relationships
- accountability
- specialised expertise
- exception handling

Then design a future-state model that:
- automates commodity production
- augments expert judgement
- preserves human accountability
- clearly identifies where humans create differentiated value
- improves the customer or employee experience
- avoids automating simply for headcount reduction
- measures both efficiency and outcome quality

Return:
1. current value chain
2. automation matrix
3. future human roles
4. AI workflow architecture
5. escalation model
6. transparency requirements
7. pilot rollout
8. operational metrics

💡 Free Office Hours

As AI gets better at producing the visible output, businesses need to get much clearer about where their actual value lives. Often the answer is not the document, lecture, analysis or slide deck—it is the expertise and judgement surrounding it.

Book here: https://calendly.com

Domain Names + Web and Email Hosting Done Right

Tired of overpaying GoDaddy or Namecheap? There's a smarter option with Porkbun!

Porkbun is the domain registrar trusted by creators, developers, entrepreneurs, and anyone who wants low prices without the nonsense or the forced upsells.

Here's why so many people are making the move to the Bun:
• Most domains sold at cost
• Low, transparent registration and renewal pricing
• Free features like WHOIS privacy and SSL certificates
• Powerful web and email hosting options
• Real human support 24/7, 365 days a year
• Named the #1 domain registrar by Forbes Advisor and USA Today

Whether you're launching a business, building a personal brand, starting a side project, or creating your very first website, Porkbun makes the whole thing easy.

Get $1 off your next domain registration with Porkbun now.

Get Your Domain At Porkbun

🕹️ Game Over

AI can make your slides, hack your systems and apparently teach your university class.

Might be a decent time to figure out what exactly you bring to the party.

— Aaron Automating the boring. Amplifying the brilliant.

Subscribe: link