What is the Generative AI Adaptation Paradox (Gen-Dox)?

The Gen-Dox paradox describes the phenomenon where UK micro-businesses have unprecedented access to powerful AI tools like ChatGPT, Claude, and Gemini, but struggle to achieve meaningful commercial ROI due to the 'Bootstrap Paradox' of time and capability constraints.

What is the ALeM Framework?

The Adaptive Leverage Model (ALeM) is a PhD-researched, 4-phase framework grounded in Dynamic Capabilities Theory that provides a structured pathway for UK micro-businesses to transition from AI awareness to sustained automated leverage. It was developed using Design Science Research methodology.

How does the ALeM Framework help micro-businesses adopt AI?

ALeM guides business owners through four phases: Phase 0 delivers a first AI win in under 15 minutes, Phase 1 uses an 80/20 Audit to identify the highest-value AI opportunity, Phase 2 employs the CATS Matrix for tool selection and micro-learning, and Phase 3 builds Standard Operating Procedures for lasting integration.

Protocol Number:

1263 2025 Oct SSAH

Generative AI Adaptive Leverage Model (ALeM)

From Paradox to Practice

You've been invited to review the ALeM framework as an expert evaluator. This page contains everything you need: research consent, the full framework, and a simple way to share your feedback.

Estimated time: 2 min consent · 30–45 min framework review · 60–90 min interview

Before You Begin: Research Information & Consent

Click to expand

This takes about 2 minutes

This section covers the formal research information required by the University of Hertfordshire Ethics Committee. Once you've read the key points and confirmed your consent by email, you're all set to explore the ALeM framework.

About This Research

This research is part of a doctoral thesis at the University of Hertfordshire, developing the Adaptive Leverage Model (ALeM) — a practical framework to help UK micro-businesses move from AI overwhelm to simple, effective use of tools like ChatGPT, Claude, and Gemini.

You've been invited because of your expertise in policy/SME advocacy, academic research, or practitioner experience in AI/digital transformation. Your honest evaluation will directly shape and validate the framework.

Time Commitment

30–45 min review + 60–90 min interview

Interview Format

Microsoft Teams, at a time that suits you

Fully Anonymous

All data anonymised & GDPR compliant

Key Things to Know

Participation is entirely voluntary

You are free to withdraw at any time without giving a reason and without any negative consequences. If you withdraw, your data will be removed on request (before analysis is completed).

Your data is kept strictly confidential

All information is anonymised — your name and organisation are replaced with descriptors like "Policy Expert 1" or "Practitioner Expert 3." Recordings are stored on secure, encrypted University of Hertfordshire OneDrive. Data is retained for 5 years after thesis submission, then securely deleted. All processing complies with GDPR.

The interview will be audio recorded

With your permission, the interview is audio-recorded and transcribed for analysis. You can request to review your transcript before it's included. Anonymised extracts may be quoted in the thesis and related publications.

No risks — critical feedback is encouraged

This research poses minimal risk. There are no professional consequences for providing critical or negative feedback — in fact, constructive critique is exactly what's needed for rigorous validation.

What you get from participating

Early access to a novel AI adoption framework, your contribution directly shapes the final validated model, and upon completion you may reference your role as an expert evaluator in your professional profile. You'll also be offered the final validated framework and research findings.

Ethics approval & oversight

This study has been reviewed and approved by the Social Sciences, Arts and Humanities Ethics Committee (Protocol: 1263 2025 Jul SSAH). Any concerns can be directed to the University's Secretary and Registrar, College Lane, Hatfield, Herts, AL10 9AB.

Formal Documents

For your records, here are the full formal documents required by the University ethics process:

Participant Information Sheet

Full 7-page research information document

Consent Form

Full 3-page formal consent document

Confirm Your Consent

By confirming, you're acknowledging that you've read the participant information above, you understand the study, and you're happy to take part. You can do this simply by clicking the button below — it opens a pre-filled email you just need to send.

What you're confirming:

I've read and understood the Participant Information Sheet
I understand my role is to evaluate the ALeM framework
I know participation is voluntary and I can withdraw anytime
I consent to the interview being audio recorded and transcribed
I understand my data will be anonymised and kept confidential

I Confirm — Send Consent Email

This opens a pre-filled email in your mail app. Just check it and hit send.

Prefer the formal route? Download and return the Consent Form PDF instead.

Questions? Contact Bear McBlack at b.c.mcblack@herts.ac.uk or supervisors Dr Derek Ong (d.ong@herts.ac.uk) and Prof Amin Hosseinian-Far (a.hosseinian-far@herts.ac.uk).

Start Here: Watch This First

Click to expand

Start Here

A 2-Minute Introduction to the ALeM Framework

Before you dive in, this short video explains what the ALeM is, why it exists, and how it can help your business — in plain English.

From Paradox to Practice: The Problem & The Solution

Click to expand

"Imagine giving the keys to a Formula 1 car to a novice driver. Over 5 million UK micro-businesses have been handed powerful AI technology — but have no idea how to use it. They’re told it will make them faster, but they’re stuck in the pits."

The Problem: The "Gen-Dox"

This research discovered that this isn’t just a skills gap — it’s a paradox. The Generative AI Adaptation Paradox (Gen-Dox) is the gap between easy access to AI and the massive, hidden difficulty of adapting the business to get any real value from it.

The "Bootstrap Paradox"

A deeper barrier prevents the journey from starting:

Need Time/Capability to Learn GenAI

Can’t Justify Time Until Value is Seen

The Gen-Dox Paradox — Understanding the cyclical adoption barrier.

The Bootstrap Paradox — The catch-22 of needing time to learn AI, but needing AI to free up time.

The Solution: The "ALeM"

Most research stops at describing this problem. This project built a solution. The Adaptive Leverage Model (ALeM) is a practical, step-by-step framework — the driver’s manual, training program, and race strategy, all in one:

1. Sense

SENSE the real opportunities — where can AI actually help?

2. Seize

SEIZE them by targeting high-impact, low-effort tasks.

3. Reconfigure

RECONFIGURE the business to make that value last.

The ALeM is built on a solid academic foundation — Dynamic Capabilities Theory — and bridges the gap between AI’s potential and the everyday reality of a small business owner.

Not Sure AI Is for You? Read This First.

You don’t need to be "tech-savvy." You just need to know what AI can do for your business — right now, today.

30 min

A coaching business owner used AI to complete research that previously took 2–3 weeks.

5 min

A sole trader used AI to reorganise an entire folder system that would have taken days manually.

10 min

A small business used AI to draft a client email, meeting agenda, and product description — tasks that used to eat up half a morning.

These are real business owners who discovered that AI takes the repetitive, time-consuming tasks off your plate so you can focus on what actually grows your business.

Common concerns from business owners (and honest answers)

"Is my data safe?" — Yes. All major AI tools have privacy settings and GDPR compliance. You control what you share.

"I’m not technical enough." — If you can write an email, you can use AI. The prompts in this framework are pre-written.

"I’ve heard AI makes things up." — It can. That’s why this framework teaches you to verify outputs. You stay in control.

"I don’t have time to learn something new." — Phase 0 takes under 15 minutes.

The Small Business AI Advantage — Why AI matters specifically for micro-businesses and sole traders.

Key Statistics: The Scale of the Problem

Click to expand

The Adoption Gap

Only 15% of small businesses have adopted at least one AI technology, despite making up 99.16% of the entire UK business population.

Source: British Chambers of Commerce (2024)

Micro-Business Reluctance

Over 5.2 million micro-businesses exist in the UK, yet 73% of sole traders are using AI in a limited way (31%) or have no plans to adopt it (42%).

Source: Moneypenny (2025); Money.co.uk (2025)

Shallow Adoption

Only 3% of small firms (approx. 49,300) are using 4-5 AI technologies simultaneously, compared to 20% of large firms.

Source: Dept. for Digital, Culture, Media & Sport (2022)

Why This Matters

UK micro-businesses aren't a niche — they are the economy. With 5.49 million SMEs employing 60% of the UK workforce and generating £2.75 trillion in turnover (over half of all UK business revenue), the stakes of getting AI adoption right are enormous. Yet the data above shows most of these businesses are either stuck at the starting line or barely past it.

The opportunity cost is staggering. McKinsey estimates that generative AI alone will drive $1.3 trillion in global economic impact annually by 2030, and firms that have adopted AI report that 63% saw an unexpected boost in revenue — even when their primary goal was simply cutting costs. Meanwhile, 91% of small businesses using AI say it directly boosts their revenue. The gap between these businesses and the millions still hesitating isn't just a missed opportunity — it's a widening competitive divide that threatens the backbone of the UK economy.

£2.75tn

Annual SME turnover — over half of all UK business revenue

$1.3tn

Projected annual global economic impact of GenAI by 2030

91%

Of SMBs with AI say it directly boosts their revenue

Sources: FSB UK Small Business Statistics (2024); Money.co.uk (2025); McKinsey Global Institute; British Chambers of Commerce (2025)

The Theoretical Anchor: Dynamic Capabilities Theory

Click to expand

To solve a problem of adaptation, we need a theory of adaptation. DCT provides the lens, explaining that advantage comes from three core competencies:

Sensing

The ability to identify and assess high-value AI opportunities, cutting through the hype.

Seizing

The ability to mobilize scarce resources (especially time) to capture value from an opportunity.

Reconfiguring

The ability to embed successful pilots into daily operations to build a sustainable advantage.

The Method: Design Science Research (DSR)

Click to expand

The goal of this research is to create the missing process that connects theory (DCT) to practice (GenAI) for micro-businesses. A 3-phase DSR journey was used:

Phase 1: Problem Definition

Empirically validated the Gen-Dox with N=3 interviews, analyzed via Gioia methodology.

Phase 2: Artifact Design

Systematically translated empirical findings and DCT principles into a practical solution.

Phase 3: Artifact Evaluation

Seeking critical, external evaluation from you, the expert panel.

The Artifact: The Generative AI Adaptive Leverage Model (ALeM)

Click to expand

This is the interactive artifact itself. It's a process, delivered as a tool, designed to solve the Bootstrap Paradox first, then systematically build capability. Select a phase to see the tools.

Figure: The ALeM Framework — A four-phase cyclical model grounded in Dynamic Capabilities Theory (Sensing, Seizing, Reconfiguring), preceded by Phase 0: Quick Wins to overcome the Bootstrap Paradox.

Figure 1 — ALeM Overview — A complete introduction to the Adaptive Leverage Model framework, what it is, why it exists, and how to navigate this tool.

Quick Check: Where Should You Start?

Answer these three quick questions to find your starting point:

I have signed up for or used at least one AI tool (ChatGPT, Claude, Gemini, Copilot, etc.)

I have used AI to complete a real work task (e.g., drafted an email, summarised a document, researched a topic)

I use AI at least once a week in my business and feel fairly comfortable with it

Your Starting Point:

0–1 Yes

Start at Phase 0: The Quick Win

Get your first AI result in under 15 minutes — no experience needed.

2–3 Yes

Start at Phase 1: Sensing & Planning

You've got the basics — now find the highest-value task to automate.

Phases 2 & 3 are part of the iterative cycle — you'll reach them naturally as you progress through the framework.

Your Journey at a Glance

Click any phase tab below to jump straight in. Each phase builds on the last.

Phase 0

The Quick Win

~15 min

Phase 1

Sensing & Planning

~20–30 min

Phase 2

Seizing & Learning

~20–30 min

Phase 3

Reconfiguring

~15–20 min

Total estimated time: 70–95 minutes across all phases (not in one sitting — take it at your own pace)

Phase 0: The Quick Win (Antidote to the Paradox)

Estimated time: ~15 minutes

Figure 2 — Phase 0 Walkthrough — A guided walkthrough of Phase 0: picking your first AI tool and getting your first quick win in under 15 minutes.

The Psychological Problem

The biggest barrier to GenAI isn't skill; it's the fear of the blank page and the feeling of being overwhelmed. You can't justify spending time to find value if you don't have the time to begin with.

The Solution

This phase is designed to give you an immediate, tangible win in under 15 minutes. It has two steps: picking a tool and executing a pre-built prompt. This builds your confidence and motivation before we ask you to do any procedural work.

Already familiar with AI tools? If you already use tools like ChatGPT, Claude, or Gemini regularly, you can to start the strategic planning process.

A Quick Word on Data Privacy

One of the most common concerns we hear is "Is my data safe?" Here's the honest answer:

Safe to share: General business questions, public information, draft content ideas, meeting agendas, generic emails.

Never share: Client names/personal data, financial account details, passwords, medical records, anything covered by GDPR.

All major AI tools (ChatGPT, Claude, Gemini) offer privacy settings. You can turn off chat history and opt out of training. You control what goes in. If in doubt, don't type it.

The AI Safety Net — How to use AI safely and responsibly — protecting your data, your clients, and your reputation.

What AI Does vs. What You Still Do

AI is not magic and it's not replacing you. Think of it as a very fast first-draft assistant. Here's the reality:

AI handles:

First drafts, research summaries, data organisation, repetitive formatting, brainstorming options, rewriting text in different tones.

You still handle:

Final decisions, fact-checking, personal relationships, creative direction, business strategy, anything requiring your judgement or lived experience.

"AI writes the first draft — you edit, verify, and decide. It's a very fast intern, not a replacement for your expertise."

You can do this on your phone!

Every AI tool in this framework works on your mobile phone — no laptop or desktop required. ChatGPT, Claude, Gemini, and Perplexity all have free mobile apps. You can complete Phase 0 on the bus, in a coffee shop, or between client meetings.

Step 1: The "Good Enough" Tool Selection Framework (TSF)

Don't get paralyzed by choice. Follow this 3-filter decision tree to pick a tool in 60 seconds.

The Tool Selection Framework (TSF) — How the 3-filter decision tree works — pick the right AI tool for your business in 60 seconds.

Filter 1: Data Governance (The Trust Boundary)
Do you handle sensitive client data (e.g., legal, medical, financial)?
YES: Start with a secure, local LLM (like LM Studio, Ollama) to keep data on your device.
NO: Proceed to Filter 2.
Filter 2: Ecosystem Integration (Path of Least Resistance)
What system do you already use every day?
Microsoft 365 (Word, Outlook): Start with Microsoft Copilot.
Google Workspace (Docs, Gmail): Start with Google Gemini.
Neither: Proceed to Filter 3.
Filter 3: Primary Use Case (Functional Fit)
What is your main need?
Content Creation (e.g., writing emails, blog posts, social media captions, marketing copy): Start with ChatGPT.
Web Research & Fact-Checking (e.g., competitor analysis, market trends, verifying claims): Start with Perplexity.
Long-Form Analysis & Reports (e.g., business reports, data analysis, document review): Start with Claude.
Image & Visual Generation (e.g., product photos, social media graphics, brand visuals): Start with Google Gemini.

Ready? Sign up here (all free to start):

ChatGPT

Free

Pro: £20/mo

Claude

Free

Pro: £18/mo

Perplexity

Free

Pro: £17/mo

Gemini

Free

Advanced: £19/mo

You do NOT need a paid plan. The free version of any tool is more than enough to get started and complete all phases.

Step 2: Execute One of Five Template Prompts

Now, use your chosen tool. Click a template below, choose "Simple" or "Advanced," and copy the prompt.

New to AI? Start with the GTC (Simple) version — it's just 3 lines: your Goal, Task, and Context. Think of it as telling a colleague what you need. Once you're comfortable, try ASPECT (Advanced) for more control over the output. You don't need the advanced version to get great results.

What if the result isn't great? That's completely normal — especially the first time. AI improves with better instructions, just like a new employee. Try rewording your request, adding more context, or being more specific about what you want. Every "bad" result teaches you how to get a better one next time.

Quick Fix: AI Gave You a Bad Result?

Too vague or generic?

Add more context. Tell the AI who you are, who it's for, and what tone you want. Example: "Rewrite this for a local bakery owner. Keep it friendly and under 100 words."

Too long or too formal?

Tell it exactly what you want: "Make this shorter — maximum 3 sentences" or "Rewrite this in a casual, conversational tone."

Completely wrong or off-topic?

Start a new chat and try the ASPECT (Advanced) version instead. The extra structure helps the AI understand exactly what you need. Or simply say: "That's not what I meant. I want [describe what you actually need]."

Golden rule: You can always reply to the AI and say "Try again, but this time..." — it learns from the conversation.

GTC vs ASPECT Prompting — The two prompt frameworks explained — start simple with GTC (Goal, Task, Context), then level up to ASPECT for more advanced results.

Template 1: The Content Creator Objective: Generate professional written content quickly. Best with: ChatGPT

**GOAL:** Write content for my business.
**TASK:** Draft a professional email to a client confirming our meeting next Tuesday at 2pm and outlining the three discussion points: project timeline, budget review, and next steps.
**CONTEXT:** I run a small graphic design studio. The client is a new local restaurant needing branding work.

**ACTION:** Draft a professional and reassuring client email that confirms a meeting and sets a clear agenda.
**STEPS:**
1. Open with a warm but professional greeting that acknowledges the client by name.
2. Confirm the meeting date, time, and location clearly.
3. Outline the three agenda items as a numbered list with a one-line description of each.
4. Close with an invitation for the client to add any items they'd like to discuss.
5. End with a confident, positive sign-off.
**PERSONA:** Act as an experienced, organised Account Manager. Your tone is professional, warm, and proactive — never casual or salesy.
**EXAMPLES:**
- Good Tone: "Hi Alex, thanks for confirming Tuesday. To make the most of our time, here's what I'd suggest we cover..."
- Bad Tone: "Hey! So we're meeting Tuesday right? Cool. We can chat about stuff."
**CONSTRAINTS:**
- The entire email must be under 150 words.
- Do not use exclamation marks.
- Use a clear subject line: "Meeting Confirmation — [Date] | [Your Business Name]"
**TEMPLATE:**
Subject: Meeting Confirmation — Tuesday 14th Jan | [Your Studio Name]
Hi [Client Name],
[Confirm meeting details]
[Numbered agenda items]
[Invitation to add items]
[Warm closing]
Best regards,
[Your Name]

Next: Open your AI tool → paste this prompt → read the result → edit anything that doesn't sound like you.

GTC: ~40 words | ASPECT: ~150 words

What Sarah Got Back (Example AI Output)

Sarah used the GTC prompt with ChatGPT. Here's what she received in under 30 seconds:

Subject: Meeting Confirmation — Tuesday 14th Jan | Sarah's Design Studio

Hi Alex,

Thank you for confirming our meeting on Tuesday 14th January at 2:00 PM. I'm looking forward to it.

To make the most of our time together, here's what I'd suggest we cover:

1. Project timeline — Agreeing key milestones for your branding package.
2. Budget review — Confirming scope and costs so there are no surprises.
3. Next steps — Mapping out what happens after our meeting.

If there's anything else you'd like to add to the agenda, just let me know.

Best regards,
Sarah

Time taken: ~30 seconds. Time saved: ~15–20 minutes of drafting and editing.

Template 2: The Web Researcher Objective: Quickly research a business question with verified sources. Best with: Perplexity

**GOAL:** Research a business question quickly.
**TASK:** Find the top three social media trends for small UK businesses in 2025. For each trend, give me a one-sentence summary and explain why it matters for a small business.
**CONTEXT:** I run a small bakery in Leeds and want to improve my Instagram strategy.

**ACTION:** Produce a concise research briefing on the top 3 social media trends relevant to small UK businesses in 2025.
**STEPS:**
1. Identify and rank the top 3 trends by relevance to small, local businesses (not enterprise).
2. For each trend, provide a one-sentence summary of what it is.
3. Explain in 2-3 sentences why it matters specifically for a small UK business.
4. For each trend, suggest one simple, actionable first step the business owner could take this week.
5. Cite or reference at least one credible source per trend.
**PERSONA:** Act as a Digital Marketing Consultant who specialises in helping independent, local UK businesses. Your tone is practical, encouraging, and jargon-free.
**EXAMPLES:**
- Good: "Short-form video (Reels, TikTok) — 73% of consumers say short video is their preferred way to discover new products. For a bakery, this could be as simple as a 15-second 'behind the scenes' clip of morning baking."
- Bad: "Social media is important. You should post more."
**CONSTRAINTS:**
- Maximum 300 words total.
- Focus only on trends relevant to a local, small business (not corporate/enterprise).
- All claims must reference a source (e.g., a report, article, or platform statistic).
**TEMPLATE:**
## Social Media Trends Briefing — Small UK Businesses (2025)
**Trend 1: [Name]**
- Summary: [One sentence]
- Why it matters: [2-3 sentences]
- Action step: [One sentence]
- Source: [Reference]
[Repeat for Trends 2 and 3]

Next: Open Perplexity (or your tool) → paste this prompt → check the sources it cites → save any useful findings.

GTC: ~45 words | ASPECT: ~160 words

What Sarah Got Back (Example AI Output)

Sarah used the GTC prompt with Perplexity. Here's a summary of what she received:

Trend 1: Short-Form Video (Reels & TikTok)
73% of consumers prefer short video to discover products. For Sarah's bakery, a 15-second "morning bake" clip could reach hundreds of local customers.

Trend 2: AI-Powered Personalisation
Small businesses using personalised content see 26% higher engagement. Sarah could use AI to tailor captions to different customer segments.

Trend 3: Community-Led Content
User-generated content (customer photos, reviews) drives 4x more trust than brand content. Sarah could repost customer stories featuring her products.

Time taken: ~45 seconds. Time saved: ~2–3 hours of manual research.

Template 3: The Report Analyst Objective: Analyse data or documents and produce a structured summary. Best with: Claude

**GOAL:** Analyse my business performance.
**TASK:** Review the following monthly sales figures and tell me which product is performing best, which is declining, and suggest one action I should take next month.
**CONTEXT:** I sell three products: handmade candles (£1,200 this month, £900 last month), wax melts (£450 this month, £600 last month), and gift sets (£800 this month, £750 last month). My busiest month is usually December.

**ACTION:** Produce a structured monthly performance summary based on the provided sales data, with a clear recommendation.
**STEPS:**
1. Present the data in a clear comparison table (Product | This Month | Last Month | Change).
2. Identify the top-performing product and explain the likely reason for its growth.
3. Identify the declining product and suggest a possible cause.
4. Provide one specific, actionable recommendation for next month backed by the data.
5. Conclude with a one-sentence overall assessment of business health.
**PERSONA:** Act as a friendly but sharp Business Analyst who works with small independent retailers. Your tone is clear, data-driven, and encouraging.
**EXAMPLES:**
- Good: "Candles are up 33% month-on-month — this likely reflects seasonal demand ahead of December. Consider increasing stock by 20%."
- Bad: "Sales are fine. Keep going."
**CONSTRAINTS:**
- Output must be under 250 words.
- Use a simple table format — no charts.
- Recommendations must be specific and actionable (not vague advice).
- Products: Handmade candles (£1,200/£900), Wax melts (£450/£600), Gift sets (£800/£750).
**TEMPLATE:**
## Monthly Sales Summary — [Month]
| Product | This Month | Last Month | Change |
|---------|-----------|-----------|--------|
| [Product] | [£] | [£] | [+/-] |
**Top Performer:** [Product] — [Reason]
**Needs Attention:** [Product] — [Reason]
**Recommendation:** [Specific action]
**Overall:** [One-sentence assessment]

Next: Open Claude (or your tool) → paste this prompt → replace the example data with your real numbers → review the analysis.

GTC: ~50 words | ASPECT: ~170 words

What Sarah Got Back (Example AI Output)

Sarah used the GTC prompt with Claude to analyse her candle business sales data:

Monthly Sales Summary

| Product | This Month | Last Month | Change |
| Candles | £1,200 | £900 | +33% |
| Wax Melts | £450 | £600 | -25% |
| Gift Sets | £800 | £750 | +7% |

Top Performer: Handmade Candles (+33%) — likely driven by seasonal demand ahead of December.
Needs Attention: Wax Melts (-25%) — consider a promotional bundle or social media push.
Recommendation: Increase candle stock by 20% for December and create a "Candle + Wax Melt" gift bundle to revive melt sales.
Overall: Business is healthy with strong growth in your core product.

Time taken: ~20 seconds. Time saved: ~1–2 hours of spreadsheet analysis.

You're halfway through the templates! If you've already tried one prompt and got a result back — congratulations, you've just used AI for your business. That's the hardest part done. The rest is just exploring what else it can do.

Template 4: The Image Creator Objective: Generate a professional image for your business. Best with: Google Gemini

**GOAL:** Create an image for my business.
**TASK:** Generate a professional, warm-toned photograph of a freshly baked sourdough loaf on a rustic wooden cutting board, with soft natural morning light coming from the left. The background should be a clean, modern kitchen — slightly blurred.
**CONTEXT:** I need this for my bakery's Instagram page. My brand style is warm, artisan, and inviting.

**ACTION:** Generate one high-quality, on-brand product photograph suitable for an Instagram post.
**STEPS:**
1. The subject is a single freshly baked sourdough loaf, positioned slightly off-centre on a rustic wooden cutting board.
2. The lighting should be soft, natural morning light from the left side, creating gentle shadows.
3. The background is a clean, modern kitchen — lightly blurred (shallow depth of field).
4. Include subtle props: a linen cloth and a small bowl of flour nearby.
5. The overall mood should feel warm, artisan, and inviting.
**PERSONA:** Act as a professional Food Photographer with 10 years of experience shooting for independent bakeries and cafes.
**EXAMPLES:**
- Good visual direction: "Warm tones, natural light, shallow depth of field, styled but not overly staged — it should feel like a real kitchen, not a studio."
- Bad visual direction: "A picture of bread."
**CONSTRAINTS:**
- Aspect ratio: 4:5 (portrait, optimised for Instagram).
- Colour palette: Warm earthy tones (browns, creams, soft golds).
- No text overlays or logos.
- No people or hands in the frame.
- Photography style: editorial food photography, NOT stock photo or clip art.
**TEMPLATE:**
[Subject]: Sourdough loaf on wooden board
[Lighting]: Soft natural light, left side
[Background]: Modern kitchen, blurred
[Props]: Linen cloth, flour bowl
[Mood]: Warm, artisan, inviting
[Aspect ratio]: 4:5 portrait

Next: Open Gemini (or your tool) → paste this prompt → download the image → post it or save it for later.

GTC: ~45 words | ASPECT: ~140 words

What Sarah Got Back (Example AI Output)

Sarah used the GTC prompt with Google Gemini to generate a product image for Instagram:

Gemini generated a warm-toned, professionally lit photograph of a sourdough loaf on a rustic wooden board with soft morning light. The image was styled with a linen cloth and flour bowl, matching Sarah's artisan brand aesthetic — ready to post directly to Instagram.

Time taken: ~10 seconds. Cost saved: £50–150 vs hiring a food photographer for one shot.

Template 5: The Meeting Agenda Creator Objective: Reduce admin burden and improve meeting effectiveness. Works with: Any AI tool

**GOAL:** Create an agenda for my weekly team meeting.
**TASK:** Make a 1-hour agenda.
**CONTEXT:** We need to discuss the Q3 marketing report, the upcoming website launch, and the new client proposal.

**ACTION:** Create a detailed, timed, and outcome-driven agenda for a 60-minute weekly team meeting.
**STEPS:**
1. Allocate a specific time slot for each of the three main discussion points.
2. Begin with a 5-min review of previous action items.
3. For each agenda item, define a clear desired outcome (e.g., "Decision on...").
4. Assign a clear owner responsible for leading each discussion point.
5. Conclude with a 5-min summary of new action items.
6. Tackle the most complex topic (website launch) first.
**PERSONA:** Act as a highly experienced Chief Operating Officer (COO) known for running efficient meetings.
**EXAMPLES:**
- Good Item: "10:15 - 10:35 (20 mins) | Website Launch Final Review (Owner: Sarah) | Desired Outcome: Final go/no-go decision."
- Bad Item: "Talk about the website."
**CONSTRAINTS:**
- Total time must not exceed 60 minutes.
- Each item must have a duration, owner, and outcome.
**TEMPLATE:**
[Provide table structure for the agenda]

Next: Open your AI tool → paste this prompt → replace the example topics with your real agenda items → send to attendees.

GTC: ~30 words | ASPECT: ~150 words

What Sarah Got Back (Example AI Output)

Sarah used the GTC prompt with ChatGPT to plan her weekly team catch-up:

Weekly Team Meeting Agenda — 60 Minutes

10:00–10:05 | Action Item Review (Sarah) — Check progress on last week's items.
10:05–10:25 | Website Launch Final Review (Dev Lead) — Go/no-go decision needed.
10:25–10:45 | Q3 Marketing Report (Marketing) — Key metrics and next quarter plan.
10:45–10:55 | New Client Proposal (Sarah) — Scope, timeline, and pricing review.
10:55–11:00 | Action Items & Close — Confirm owners and deadlines.

Time taken: ~15 seconds. Time saved: ~20–30 minutes of planning and formatting.

What this saves you: Completing Phase 0 gives you a working AI tool and a tested prompt. Business owners report saving 2–5 hours per week on routine tasks like email drafting, research, and content creation — starting from day one.

Nice work! You've completed Phase 0 — you now have a working AI tool and you've seen what it can do. Phase 1 takes you from that first win to a strategic plan. You'll identify the single best task in your business for AI to handle. This is where the real time savings begin.

Phase 1: Sensing & Planning (From Audit to Action)

Estimated time: ~20–30 minutes

Figure 3 — Phase 1 Walkthrough — A guided walkthrough of Phase 1: identifying your highest-value AI opportunity using the 80/20 Audit.

Goal: Move from a "quick win" to a strategic plan. This phase walks you through a 5-step process to identify your single highest-value opportunity and commit resources to it.

Step 1: The Internal 80/20 Audit

List your top 5 most frustrating or time-consuming tasks. Rate them on a scale of 1–5 for each column. This helps you find your "20%" — the tasks where AI can make the biggest impact.

The 80/20 AI Audit — How to score your daily business tasks and find the single best candidate for AI automation.

How to Score (1–5):

Time Investment: 1 = Very little time spent → 5 = Takes up most of my day
Value Creation: 1 = Low business impact → 5 = Directly drives revenue or clients
AI Suitability: 1 = Highly personal/creative, hard for AI → 5 = Repetitive/structured, ideal for AI. (Not sure? Score it a 3 and revisit later.)

Example: Sarah runs a small graphic design studio. Here is how she scored her tasks:

Task Description	Time Investment (1=Low, 5=High)	Value Creation (1=Low, 5=High)	AI Suitability (1=Low, 5=High)
Responding to client emails	5	2	5
Writing social media captions	4	3	4
Preparing project proposals	5	5	4
Sending invoices and payment reminders	3	2	5
Designing client mood boards	4	5	2

Sarah's Insight: "Responding to client emails" scored highest for Time Investment (5) and AI Suitability (5) but lowest for Value Creation (2). This is a high-time, low-value task that AI can handle well — a perfect candidate for her first AI automation. Meanwhile, "Designing mood boards" scores high for Value Creation but low for AI Suitability — this is a task Sarah should keep doing herself.

Step 2: The External Opportunity Audit

Briefly research what others are doing. Use your AI tool to speed this up! Ask it: "How are businesses like [YOUR BUSINESS TYPE] using AI for [YOUR TASK FROM STEP 1]?"

External Opportunity Audit — How to research what other businesses in your industry are doing with AI — and steal the best ideas.

You're making real progress! You've audited your tasks and researched what's possible. Now comes the most important decision — picking your ONE task. This is the step that separates people who "try AI" from people who actually save hours every week.

Step 3: The "One-Task" Decision

You must pick only one task to focus on. Resisting the urge to multitask is key. Choose your single, highest-leverage task to pilot.

The One-Task Rule — Why focusing on a single task is the secret to AI success — and how to resist the temptation to do everything at once.

My "One-Task" is:

For your own reflection only — this isn't saved or submitted anywhere.

Step 4: Identify Your Tool Category

Based on your "One-Task," which type of AI tool do you need?

Written Content (emails, reports, social media posts) → General text-based AI (e.g., ChatGPT, Claude)
Visual Content (product images, social graphics, design concepts) → Image generation AI (e.g., Gemini, DALL-E, Midjourney)
Research & Analysis (competitor research, data analysis, fact-checking) → Research-focused AI (e.g., Perplexity, Claude)

Not sure which category? Use your AI tool to help you decide. Try this prompt: "I run a [YOUR BUSINESS TYPE]. I want to use AI to help with [YOUR ONE-TASK]. What category of AI tool would be most suitable, and can you suggest 2-3 specific options?"

Step 5: Resource Allocation Plan

Commit to it. A plan turns a vague idea into a real project.

My Goal is:

I will commit:

Success looks like:

For your own reflection only — these fields aren't saved or submitted anywhere.

What this saves you: After Phase 1, you'll know exactly which task to automate first — the one that wastes the most time for the least value. Owners who complete the 80/20 Audit typically identify 5–10 hours per week of tasks that AI could handle, and have a clear action plan to start.

Great — you have a plan! You've identified your highest-value task and know what type of AI tool you need. Phase 2 is where you pick the exact tool and learn just enough to start using it. No courses, no certifications — just targeted, 30-minute learning.

Phase 2: Seizing & Learning (Selecting & Learning)

Estimated time: ~20–30 minutes

Figure 4 — Phase 2 Walkthrough — A guided walkthrough of Phase 2: comparing AI tools with the CATS Matrix and beginning your micro-learning journey.

Goal: To make a strategic choice and acquire the minimum necessary skills to act on your plan from Phase 1.

Step 1: The 'CATS' Matrix (Tool Selection)

You've chosen your category (e.g., "Written Content"). Now, compare 2–3 specific tools in that category. Score each from 1 (Low/Poor) to 5 (High/Excellent).

The CATS Matrix — How to compare AI tools using Cost, Accessibility, Time to Value, and Skills Required — with Sarah's worked example.

What Each Score Means:

C — Cost: 1 = Very expensive → 5 = Free or already paying for it
A — Accessibility: 1 = Hard to access or set up → 5 = Can start using immediately
T — Time to Value: 1 = Long learning curve → 5 = Useful results within minutes
S — Skills Required: 1 = Needs technical knowledge → 5 = No special skills needed

Example: Sarah's "One-Task" is automating client email responses. She compares three text-based AI tools:

Tool	C (Cost) 5=Cheapest	A (Accessibility) 5=Easiest	T (Time to Value) 5=Fastest	S (Skills Required) 5=Least	Total
ChatGPT (Free)	5	5	4	5	19
Claude (Free tier)	5	4	5	4	18
Microsoft Copilot	3	4	4	5	16

Sarah's Decision: ChatGPT scored highest (19/20) — it is free, she can access it immediately on her phone or laptop, and it does not require any technical skills. She will start with ChatGPT for her email automation task. Decision rule: Choose the tool scoring 15 or above with the highest total.

You've picked your tool! That's a huge step — most people never get past "I should try AI someday." You've chosen a specific tool backed by data. Now let's make sure you know just enough to use it brilliantly.

Step 2: Targeted Micro-Learning

Don't become an "AI expert." Learn only what you need for your "One-Task." This is "Just-in-Time" learning.

Micro-Learning: AI Skills — Why you don't need to become an AI expert — and how to learn just enough to get real results in 30 minutes.

Identify Core Skill: e.g., "Writing a prompt for Instagram captions."
Find Resource: Search for a 10-minute YouTube video or blog post on that exact skill.
Time-box: Allocate 30-60 minutes for this. No more.
Practice Immediately: Apply what you learned to your "One-Task" right away.

Use AI to help you learn. Try this prompt in your chosen tool: "I'm new to using AI for [YOUR ONE-TASK]. What are the 3 most important things I need to learn to get good results? Keep it simple and practical — I have 30 minutes."

What this saves you: Phase 2 gives you a tested, working AI workflow for your highest-priority task. One participant described work that "used to take three days now takes three hours." You'll also have a CATS Matrix score to validate your tool choice.

You're in the home stretch! You've chosen your tool, learned the basics, and you're getting results. Phase 3 is about making it permanent. You'll create a simple, repeatable process (SOP) so this isn't a one-off experiment — it becomes part of how your business works, every single week.

Phase 3: Reconfiguring & Adapting (Building Your SOP)

Estimated time: ~15–20 minutes

Figure 5 — Phase 3 Walkthrough — A guided walkthrough of Phase 3: building your Standard Operating Procedure and making AI a permanent part of your workflow.

Goal: To make your successful pilot permanent. This phase turns your individual "win" into a simple, repeatable Standard Operating Procedure (SOP) for your business. This documented process is your new, reconfigured capability.

Building Your AI Asset — How to turn a one-off AI experiment into a permanent, documented business asset that saves you time every week.

Step 1: Create Your 'Minimal Viable Routine' (SOP)

Document your new process. This turns it from a personal habit into a business asset.

SOP-MKT-001: Weekly Instagram Content Generation

**SOP Title:** SOP-MKT-001: Weekly Instagram Content Generation
**OBJECTIVE:** To generate one week of social media content in under 30 minutes.
**TOOL(S) USED:** Google Gemini

**MASTER PROMPT (ASPECT):**
- ACTION: Generate 5 Instagram posts.
- STEPS: 1. Create a theme for each post. 2. Write a 50-word caption. 3. Suggest a visual.
- PERSONA: Act as a social media expert for graphic designers.
- EXAMPLES: [Your own best example]
- CONSTRAINTS: Tone must be "creative, professional, insightful."
- TEMPLATE: "Day 1: [Caption]..."

**STEP-BY-STEP PROCESS:**
1. Open Google Gemini.
2. Paste the MASTER PROMPT.
3. Review outputs for tone, accuracy.
4. Schedule posts in [Your Tool].

**SUCCESS METRIC:** Time spent per week < 30 minutes.
**VERSION HISTORY:** v1.0 (06/11/2025) - Initial creation.

Next: Paste into a Word doc or Google Doc → replace all [bracketed text] with your real details → save as your first official SOP.

You've created your first SOP! That document is now a business asset. Every week, you can follow it to save hours of work — and you can improve it over time. You're no longer "trying AI" — you're using AI.

Step 2: Understand the Automation Gradient

Your new SOP is just the start. You must make a strategic choice about how far to automate it.

Level 1: Manual MVR (You are here): You copy/paste the prompt. (Low effort, good value)
Level 2: Semi-Automated: You use a "no-code" tool like Zapier to connect your AI to another app (e.g., auto-draft a post from a blog RSS feed). (Medium effort, great value)
Level 3: Fully Autonomous: An "agent" does the whole process for you. (High effort, highest value)

Key insight: You don't need Level 3. Level 1 might be "good enough" and the right strategic choice for your business right now.

Step 3: Continuous Adaptation (PDCA / Review)

Your new capability must adapt or it will become obsolete. Use two feedback loops:

PDCA Cycle (Operational): On a weekly basis, ask "Can I improve the prompt?" (Plan-Do-Check-Act). This refines your current SOP.
Quarterly Review (Strategic): On a quarterly basis, ask "Is this SOP still the right process?" or "Is it time to move to Level 2 automation?" This challenges the whole SOP.

The Infinite Cycle — Why the ALeM framework is designed to be cyclical — once you master one workflow, you return to Phase 1 to tackle the next opportunity.

Future-Proof Your Business — Why the businesses that start using AI now will be the ones that thrive in 5 years — and how to stay ahead of the curve.

What this saves you: Phase 3 turns your one-off AI win into a repeatable, documented process — an SOP you can rely on every week. The iterative cycle means each pass gets faster. Owners who complete all four phases report an overall time saving of 8–15 hours per week and describe AI as feeling like "an extra team member."

Your ALeM Cheat Sheet (One-Page Summary) Print Me!

This is everything you need on one page. Print it, stick it on your wall, or save it as a PDF.

ALeM Quick Reference Cheat Sheet

The Adaptive Leverage Model — From Paradox to Practice

Phase 0: Quick Win (~15 min)

Pick a tool using TSF (3 filters)
Copy a template prompt (GTC or ASPECT)
Paste, run, get your first result

Tools: ChatGPT, Claude, Perplexity, Gemini

Phase 1: Plan (~20–30 min)

80/20 Audit — score your top 5 tasks
External Audit — research what others do
Pick ONE task (the "One-Task Rule")
Identify tool category
Set time & success metric

Phase 2: Learn (~20–30 min)

CATS Matrix — compare 2–3 tools
Score: Cost, Access, Time, Skills
Pick tool scoring 15+ (highest wins)
30-min micro-learning session

Phase 3: Build (~15–20 min)

Write your SOP (copy the template)
Choose automation level (1–3)
Set PDCA weekly review
Schedule quarterly strategic review

Then: Return to Phase 1 for next task!

Key Prompting Tips:

GTC = Goal + Task + Context (simple, 3 lines) | ASPECT = Action + Steps + Persona + Examples + Constraints + Template (advanced, full control)

If it's bad: Add more context → Specify length/tone → Try ASPECT version → Reply "Try again but..."

Research & Academic Context

Click to expand

The following sections provide the academic and practical context underpinning the ALeM framework. Click each title to expand the content.

The "Golden Thread": Academic Rigor

The ALeM's practical steps are a direct operationalization of academic theory. This "golden thread" connects each phase to a core DCT capability.

In Plain English:

Every step of this framework is backed by a proven theory about how organisations adapt to change. In simple terms: Phase 0 gets you motivated by showing you a quick result. Phase 1 helps you spot the right opportunities in your own business. Phase 2 helps you pick the best tool and learn just enough to use it. Phase 3 helps you turn what worked into a repeatable process you can rely on. The theory ensures this is not just a random checklist — each step builds on the last.

ALeM Practical Phase	Core Mechanism	DCT Capability Operationalised
Phase 0: Quick Wins	Bootstrap Paradox Mitigation	A pre-capability "nudge" that enables Sensing by providing initial motivation.
Phase 1: Sensing & Planning	80/20 Audit & Resource Plan	Sensing: A practical tool to analyse internal processes, not external hype.
Phase 2: Seizing & Learning	CATS Matrix & Micro-Learning	Seizing: A resource-allocation mechanism to commit scarce resources to high-ROI tasks.
Phase 3: Reconfiguring	MVR / SOP & PDCA Cycle	Reconfiguring: The engine to move from ad-hoc use to systemic, learned adaptation.

Practical Challenges & Realities

This framework is practical, not magic. You will face real-world challenges. Here is how ALeM is designed to handle them.

Digital Divide & Infrastructure

Challenge: The ALeM assumes you have reliable internet and a modern computer.
Mitigation: This is a hard prerequisite. The TSF in Phase 0 guides users to low-bandwidth or even local/on-premise tools (like Ollama) if data security or poor connectivity is a major issue.

Failure Management (e.g., Bad Outputs)

Challenge: Your AI pilot might fail, or the AI will "hallucinate" (make things up). This is demoralizing.
Mitigation: ALeM reframes this as a learning event, not a personal failure. We use a Failure Taxonomy:
- Was it a Prompt Failure? (Go back and refine your ASPECT prompt)
- Was it a Tool Failure? (Go back to the CATS matrix and try a different tool)
- Was it a Strategic Failure? (Go back to the 80/20 Audit; you picked the wrong task to automate)

Ethical & Legal Risks (GDPR, Copyright)

Challenge: As a business owner, you are legally responsible for your AI's output, including data privacy (GDPR) and copyright.
Mitigation: This is built directly into the framework:
- TSF (Phase 0): The first filter is Data Governance, forcing you to think about security before you start.
- ASPECT (Phase 0): The "C" (Constraints) in the advanced prompt is where you build in rules like "Do not use personal data" or "Ensure output is original."
- MVR (Phase 3): The "Human Review" checkpoint is a mandatory step for all customer-facing content to check for accuracy, bias, and brand safety.

AI Literacy Barriers

Challenge: Users may not know how to use AI, evaluate outputs, or understand what it's doing.
Mitigation: The framework is a "scaffolded complexity" system.
- Phase 0 uses simple templates (GTC) to build confidence.
- Phase 2 introduces "Just-in-Time" Micro-Learning.
- ASPECT framework is offered as an advanced option for users who are ready for it.

Limitations of this Framework

To be a rigorous artifact, ALeM must be transparent about its own limitations and boundary conditions.

The Time and Resource Paradox

Limitation: The framework requires an upfront investment of time and cognitive energy, the very resources a stressed owner-manager lacks.
Built-in Mitigation: Phase 0 (Quick Win) is designed specifically to solve this. It delivers a tangible ROI (e.g., 15 minutes saved) before asking for the 1-hour commitment for Phase 1. It creates motivational capital.

The Cognitive & Metacognitive Burden

Limitation: The model's structure can feel analytical or "academic" to an intuitive, action-oriented entrepreneur.
Built-in Mitigation: Scaffolded Complexity. The user starts with the simple GTC prompt framework. Only when they are ready for more power do they "unlock" the advanced ASPECT framework. This progressive disclosure respects the user's learning curve.

Boundary Conditions (Who is this for?)

Limitation: ALeM is not a universal solution.
Exclusion Criteria: This framework is not for:
- Businesses in acute financial crisis (it's not a rescue plan).
- Users who lack foundational digital literacy (e.g., cannot use a web browser).
- Users who are unwilling to commit 2-3 hours over a few weeks to learn.

Sector-Specific Limitations (e.g., Healthcare, Legal)

Limitation: Businesses in highly regulated sectors (healthcare, legal, finance) face data security risks that make public cloud AI tools (like ChatGPT) inappropriate for core tasks.
Built-in Mitigation: The TSF (Phase 0, Step 1) makes Data Governance the first decision. It explicitly directs these users to Local LLM solutions (Ollama, LM Studio) that run on their own hardware, ensuring no sensitive data ever leaves their control.

References

This artifact is built on established academic theory and contemporary market research. Key sources are listed below.

Academic Journal Articles & Books

Baker, T. and Nelson, R. E. (2005) 'Creating something from nothing: Resource construction through entrepreneurial bricolage', Administrative Science Quarterly, 50(3), pp. 329-366.
Brownlee, J. (2020) How to Choose a Machine Learning Model. Machine Learning Mastery.
Cohen, W. M. and Levinthal, D. A. (1990) 'Absorptive Capacity: A New Perspective on Learning and Innovation', Administrative Science Quarterly, 35(1), pp. 128-152.
D'Aveni, R. A. (1994) Hypercompetition: Managing the dynamics of strategic maneuvering. New York: Free Press.
Davenport, T. H. and Ronanki, R. (2018) 'Artificial intelligence for the real world', Harvard Business Review, 96(1), pp. 108-116.
Davis, F. D. (1989) 'Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology', MIS Quarterly, 13(3), pp. 319-340.
Deming, W. E. (1986) Out of the Crisis. Cambridge, MA: MIT Press.
Eisenhardt, K. M. and Martin, J. A. (2000) 'Dynamic capabilities: what are they?', Strategic Management Journal, 21(10‐11), pp. 1105-1121.
Eloundou, T., Manning, S., Mishkin, P. and Rock, D. (2023) 'GPTs are GPTs: An early look at the labor market impact potential of large language models', arXiv preprint arXiv:2303.10130.
Flavell, J. H. (1979) 'Metacognition and cognitive monitoring: A new area of cognitive–developmental inquiry', American Psychologist, 34(10), pp. 906–911.
Ghobadian, A. and O'Regan, N. (2006) 'The impact of ownership on small firm behaviour and performance', International Small Business Journal, 24(6), pp. 555-586.
Goswami, S. (2025) Choosing the Right LLM for Your Use Case. Towards Data Science.
Gregor, S. and Hevner, A. R. (2013) 'Positioning and presenting design science research for maximum impact', MIS Quarterly, 37(2), pp. 337-355.
Hevner, A. R., March, S. T., Park, J. and Ram, S. (2004) 'Design science in information systems research', MIS Quarterly, 28(1), pp. 75-105.
Kerr, M. S. (1991) 'Leveraging human assets in law firms: Human resource management and just-in-time training', Law Practice Management, 17(4), pp. 26-31.
Kevill, A. (2014) How are dynamic capabilities enacted in micro-enterprises and what role do different parties and managerial time allocation play in this enactment? White Rose eTheses Online.
Koch, R. (1998) The 80/20 principle: The secret of achieving more with less. London: Nicholas Brealey Publishing.
Kotter, J. P. (1996) Leading change. Boston, MA: Harvard Business Press.
Mitra, A. (2025) A Guide to Choosing the Right LLM for Your Enterprise. Analytics India Magazine.
Noy, S. and Zhang, W. (2023) 'Experimental evidence on the productivity effects of generative artificial intelligence', Science, 381(6654), pp. 187-192.
Peffers, K., Tuunanen, T., Rothenberger, M. A. and Chatterjee, S. (2007) 'A design science research methodology for information systems research', Journal of Management Information Systems, 24(3), pp. 45-77.
Rogers, E. M. (2003) Diffusion of Innovations (5th ed.). New York: Free Press.
Simon, H. A. (1972) 'Theories of Bounded Rationality', in McGuire, C. B. and Radner, R. (eds.) Decision and Organization. Amsterdam: North-Holland Publishing Company, pp. 161-176.
Teece, D. J. (2007) 'Explicating dynamic capabilities: the nature and microfoundations of (sustainable) enterprise performance', Strategic Management Journal, 28(13), pp. 1319-1350.
Teece, D. J. (2014) 'The foundations of enterprise performance: Dynamic and ordinary capabilities in an (economic) theory of firms', Academy of Management Perspectives, 28(4), pp. 328-352.
Teece, D. J., Pisano, G. and Shuen, A. (1997) 'Dynamic capabilities and strategic management', Strategic Management Journal, 18(7), pp. 509-533.
Tidestrom, A. (2022) 'Mental health in small business: A review and research agenda', International Small Business Journal, 40(4), pp. 497-513.
Ucbasaran, D., Westhead, P. and Wright, M. (2010) 'The Nature of Entrepreneurial Experience, Business Failure and Comparative Optimism', Journal of Business Venturing, 25(6), pp. 541-555.
Venkatesh, V. and Davis, F. D. (2000) 'A theoretical extension of the technology acceptance model: Four longitudinal field studies', Management Science, 46(2), pp. 186-204.
Venkatesh, V., Morris, M. G., Davis, G. B. and Davis, F. D. (2003) 'User Acceptance of Information Technology: Toward a Unified View', MIS Quarterly, 27(3), pp. 425-478.
Walsh, J. P. and Ungson, G. R. (1991) 'Organizational memory', Academy of Management Review, 16(1), pp. 57-91.
Weick, K. E. (1984) 'Small wins: Redefining the scale of social problems', American Psychologist, 39(1), pp. 40-49.
Welsh, J. A. and White, J. F. (1981) 'A small business is not a little big business', Harvard Business Review, 59(4), pp. 18-32.

Industry Reports & Grey Literature

British Chambers of Commerce (2024). Most SMEs still struggling to embrace AI. Available at: https://www.britishchambers.org.uk (Accessed: 24 July 2024).
Department for Digital, Culture, Media & Sport (2022). AI activity in UK businesses: Executive summary. London: GOV.UK. Available at: https://www.gov.uk/government/publications/ai-activity-in-uk-businesses/ai-activity-in-uk-businesses-executive-summary (Accessed: 11 January 2022).
Hu, K. (2023) ChatGPT sets record for fastest-growing user base. Reuters. Available at: https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/ (Accessed: 18 October 2025).
McKinsey & Company (2023) The economic potential of generative AI: The next productivity frontier. McKinsey Global Institute.
Money.co.uk (2025). UK business statistics 2025 - business facts and stats report. Available at: https://www.money.co.uk (Accessed: 8 July 2025).
Moneypenny (2025). The state of AI adoption in UK businesses | 2025 trends & insights. Available at: https://www.moneypenny.com (Accessed: 1 September 2025).
Ofcom (2024) Connected Nations 2024. Office of Communications. Available at: https://www.ofcom.org.uk/research-and-data/multi-sector-research/infrastructure-research/connected-nations-2024 (Accessed: 18 October 2025).
University of Brighton (2025) Developing absorptive capacity in SMEs. Available at: https://research.brighton.ac.uk/files/208088/Getting_the_Tail_to_Wag.pdf (Accessed: 18 October 2025).

AI Platform & Tool Documentation

Alibaba Cloud (2025) Qwen. Available at: https://qwen.ai/ (Accessed: 18 October 2025).
Apple (2025) Use ChatGPT with Apple Intelligence. Available at: https://support.apple.com/guide/mac-help/use-chatgpt-with-apple-intelligence-mchlfc5cf131/mac (Accessed: 18 October 2025).
LangChain (2025) Local LLMs. Available at: https://python.langchain.com/docs/how_to/local_llms/ (Accessed: 18 October 2025).
Manus (2025) Manus: Hands On AI. Available at: https://manus.im/home (Accessed: 18 October 2025).
Meta AI (2025) Meta AI. Available at: https://en.wikipedia.org/wiki/Meta_AI (Accessed: 18 October 2025).
Mistral AI (2025) Le Chat enterprise AI assistant. Available at: https://mistral.ai/products/le-chat (Accessed: 18 October 2025).
Moonshot AI (2025) Kimi (chatbot). Available at: https://en.wikipedia.org/wiki/Kimi_(chatbot) (Accessed: 18 October 2025).
OpenAI (2025) How your data is used to improve model performance. Available at: https://help.openai.com/en/articles/5722486-how-your-data-is-used-to-improve-model-performance (Accessed: 18 October 2025).
xAI (2025) Grok. Available at: https://x.ai/grok (Accessed: 18 October 2025).

Technical Blogs & Industry Analysis

Arsturn (2025) Claude vs. Local LLMs: Choosing the Right AI for Your Business. Available at: https://www.arsturn.com/blog/claude-vs-local-llms-choosing-the-right-ai-for-your-business (Accessed: 18 October 2025).
Binadox (2025) Best Local LLMs for Cost-Effective AI Development in 2025. Available at: https://www.binadox.com/blog/best-local-llms-for-cost-effective-ai-development-in-2025/ (Accessed: 18 October 2025).
Built In (2025) 5 Aspects of Data Privacy to Consider in AI Adoption. Available at: https://builtin.com/articles/ai-adoption-data-privacy (Accessed: 18 October 2025).
CamoCopy (2025) AI Assistants Privacy Report. Available at: https://www.camocopy.com/ai-assistants-privacy/ (Accessed: 18 October 2025).
CreativeTechs (2025) Keep Sensitive Data Private by Disabling AI Training Options. Available at: https://www.creativetechs.com/2025/09/05/keep-sensitive-data-private-by-disabling-ai-training-options/ (Accessed: 18 October 2025).
ema.co (2025) Best Open-Source LLMs for Enterprises in 2025. Available at: https://www.ema.co/additional-blogs/addition-blogs/best-open-source-llms (Accessed: 18 October 2025).
F11Photo (2025) Keep Sensitive Data Private by Disabling AI Training Options. Available at: https://f11photo.com/2025/09/05/keep-sensitive-data-private-by-disabling-ai-training-options/ (Accessed: 18 October 2025).
Genspark (2025) Genspark: Super AI Agent. Available at: https://play.google.com/store/apps/details?id=ai.mainfunc.genspark (Accessed: 18 October 2025).
Ignesa (2025) The Truth About Local LLMs: When You Actually Need Them. Available at: https://ignesa.com/insights/the-truth-about-local-llms-when-you-actually-need-them/ (Accessed: 18 October 2025).
Medium (2025) The Great AI Privacy Divide: Claude, ChatGPT, Gemini, and Copilot. Available at: https://medium.com/@michael_79773/the-great-ai-privacy-divide-one-year-later-two-worlds-apart-e74ce6187f1f (Accessed: 18 October 2025).
n8n (2025) How to run local LLMs for private, secure, and cost-effective AI automation. Available at: https://blog.n8n.io/local-llm/ (Accessed: 18 October 2025).
NetFriends (2025) AI Privacy Policy Evaluation: ChatGPT vs Gemini vs Claude. Available at: https://www.netfriends.com/blog-posts/ai-privacy-policy-evaluation-chatgpt-vs-gemini-vs-claude (Accessed: 18 October 2025).
SBTDC (2025) Protecting Your Data: Responsible AI Practices for Small Businesses. Available at: https://sbtdc.org/blog/protecting-your-data-responsible-ai-practices-for-small-businesses (Accessed: 18 October 2025).
Sentisight (2025) Open Source LLMs You Can Actually Deploy in 2025. Available at: https://www.sentisight.ai/open-source-llms-you-can-actually-deploy/ (Accessed: 18 October 2025).
Signity Solutions (2025) On Premise vs Cloud Based LLM: Which Is Right for Your Industry? Available at: https://www.signitysolutions.com/blog/on-premise-vs-cloud-based-llm (Accessed: 18 October 2025).
Skywork AI (2025) How to Keep ChatGPT Data Private: A 2025 Guide to Enterprise AI Security. Available at: https://skywork.ai/skypage/en/How-to-Keep-ChatGPT-Data-Private:-A-2025-Guide-to-Enterprise-AI-Security/1976537200450400256 (Accessed: 18 October 2025).

Your Role: The Expert Evaluation

Click to expand

A Message from Bear

Feedback on the ALeM

A short video explaining what feedback is most helpful and how it will be used.

Candid feedback is sought on this artifact across four key areas:

Theoretical Coherence

Is the link between ALeM and Dynamic Capabilities Theory sound?

Practical Utility

Is this framework genuinely useful and simple to implement?

Completeness

What critical gaps, assumptions, or blind spots have I missed?

Contextual Appropriateness

Does this reflect the operational reality of UK micro-businesses?

Thank You for Your Time

I genuinely appreciate you taking the time to review this artifact. Your expertise and honest feedback are what will make this framework truly useful for the micro-business owners it's designed to help. Every observation — whether positive or critical — strengthens this research.

Ready to Share Your Feedback?

Send Feedback via Email (Pre-Filled with 4 Key Questions) Book a 15-Minute Call

Choose to send a pre-filled email or book a guided walkthrough call. Thank you again — Bear.