AI Reality Check

Let's skip the hype and talk about what's actually happening.

AI isn't coming for your job in some distant future. It's already reshaping entire industries right now. Some companies are using it to do more with less. Others are watching their competitors pull ahead while they debate whether to "wait and see."

This guide is a reality check. We'll look at where automation is genuinely headed, which roles and industries face the most disruption, and how to assess whether your business is positioned to benefit - or get left behind.

Fair warning: some of this will be uncomfortable. But uncomfortable truths are more useful than comfortable lies.

Here's what's real today, not speculation:

Routine knowledge work is already being automated. Not perfectly, not completely, but enough to matter. Tasks that used to require a junior analyst, a support agent, or an admin assistant can now be handled - or at least drafted - by AI tools that cost £20-50/month.

Agentic systems are emerging. We've moved beyond "AI that writes emails" to "AI that handles entire workflows" - booking meetings, processing invoices, triaging support tickets, generating reports. These agents are clunky and require oversight, but they're improving monthly.

The cost curve is collapsing. What cost £10 in API calls two years ago now costs £0.50. What required a specialist prompt engineer can now be done with a natural language conversation. The barriers to adoption are falling faster than most companies realise.

The talent gap is real. Companies that started experimenting two years ago now have internal expertise, documented playbooks, and proven ROI. Companies that haven't started are increasingly behind - and the gap is widening, not narrowing.

Predictions are risky, but directionally, expect:

Agents that work overnight. Submit a task at 5pm, review results at 9am. AI that can plan, execute, hit roadblocks, try alternatives, and deliver completed work - not just drafts.

Industry-specific automation. Generic AI tools are being wrapped into vertical solutions for legal, finance, HR, healthcare, logistics. These won't replace humans outright, but they'll dramatically change staffing ratios.

Integration everywhere. AI won't be a separate tool you log into. It'll be embedded in your email, your CRM, your ERP, your everything. The question won't be "should we use AI?" but "are we using it effectively?"

Consolidation of advantage. Early adopters will have compounding benefits: better data, refined processes, trained staff, proven ROI. Late adopters will face a steeper climb to catch up.

Not all roles and industries face equal disruption. Here's an honest assessment.

These roles have significant overlap with current AI capabilities. This doesn't mean they'll disappear - but the number of people needed will likely decrease substantially.

Administrative and coordination roles

• ➢Executive assistants, office managers, scheduling coordinators
• ➢AI can already handle calendar management, travel booking, meeting coordination, document preparation
• ➢The shift: Fewer admins, with remaining roles focused on judgement calls and relationship management

Entry-level analysis and research

• ➢Junior analysts, research assistants, data entry specialists
• ➢AI handles data gathering, synthesis, first-pass analysis, report generation
• ➢The shift: Senior people do more directly; junior roles become AI-augmented or disappear

Routine customer interaction

• ➢Tier-1 support agents, basic customer service, FAQ handlers
• ➢AI already handles 40-60% of routine queries at leading companies
• ➢The shift: Humans handle escalations, complex issues, and relationship-critical interactions

Content production at scale

• ➢Social media managers, basic copywriters, content mill writers
• ➢AI generates acceptable first drafts for most routine content
• ➢The shift: Fewer producers, more editors; premium human-written content becomes a differentiator

Transaction processing

• ➢Bookkeeping, invoice processing, claims processing, data reconciliation
• ➢AI matches, categorises, and flags exceptions with increasing accuracy
• ➢The shift: Accountants become reviewers; processing staff numbers decline significantly

These roles have partial overlap with AI capabilities. They'll change significantly but probably won't see mass displacement in the near term.

Professional services (law, accounting, consulting)

• ➢AI handles research, document review, first drafts, analysis
• ➢Human judgement still required for advice, strategy, client relationships
• ➢The shift: Leverage ratios change; fewer juniors supporting each partner

Software development

• ➢AI accelerates coding by 30-50% for routine work
• ➢Complex architecture, novel problems, and maintenance still require humans
• ➢The shift: Smaller teams ship more; "AI-native" developers outcompete traditional ones

Sales (transactional)

• ➢AI handles lead qualification, follow-up sequences, data entry
• ➢Relationship selling and complex deals still human-dependent
• ➢The shift: SDR/BDR roles shrink; closers become more productive

Middle management (operational)

• ➢AI handles reporting, status tracking, routine coordination
• ➢Leadership, motivation, and difficult decisions remain human
• ➢The shift: Flatter organisations; fewer coordinators, more doers

These roles have characteristics that make automation harder - but don't assume permanent safety.

Physical work in unstructured environments

• ➢Trades, construction, maintenance, healthcare delivery
• ➢Robotics isn't keeping pace with AI; physical world is messier than digital
• ➢The caveat: Scheduling, admin, and coordination aspects will still automate

High-stakes judgement

• ➢Senior leadership, crisis management, strategy, negotiation
• ➢Stakes too high and situations too novel for current AI
• ➢The caveat: AI will inform these decisions even if not making them

Creative vision and taste

• ➢Art direction, brand strategy, product vision, design leadership
• ➢AI generates options; humans choose and refine
• ➢The caveat: Execution roles shrink even as vision roles remain

Trust-dependent relationships

• ➢Therapists, senior advisors, executive coaches, key account managers
• ➢Human connection is the product, not a means to an end
• ➢The caveat: Admin aspects of these roles will still automate

Some industries face more immediate disruption than others.

Professional services (legal, accounting, consulting)

The entire business model relies on billing human hours for knowledge work. AI directly attacks this foundation. Firms that adapt will thrive with fewer people; firms that don't will face pricing pressure from AI-augmented competitors.

Signs of disruption already visible: Major firms quietly reducing graduate intakes; clients demanding AI-assisted delivery; alternative legal service providers growing rapidly.

Financial services (banking, insurance, asset management)

Huge volumes of document processing, analysis, and routine decision-making. Back-office transformation is already well underway at leading institutions.

Signs of disruption: Major banks announcing "efficiency programmes" with 20%+ headcount targets; insurers automating claims processing; robo-advisors capturing market share.

Media and content

Content that was differentiated by production capability is now commoditised. The economics of content mills, local news, and routine B2B content are broken.

Signs of disruption: Layoffs across digital media; AI-generated content in major publications; agencies restructuring around smaller teams.

Customer service / BPO

The entire value proposition of offshore contact centres - cheaper labour - is undermined when AI can handle routine queries at near-zero marginal cost.

Signs of disruption: BPO companies investing heavily in AI; clients insourcing with AI tools; chatbot resolution rates climbing.

Healthcare (admin, not clinical)

Clinical work is protected by regulation and stakes; administrative work is wide open. Scheduling, billing, prior authorisation, documentation - all automatable.

Manufacturing and logistics

Physical operations still need humans; planning, scheduling, and optimisation increasingly don't. The office attached to the factory is more exposed than the factory floor.

Education

Teaching will remain human; grading, administration, content creation, and tutoring support are automating rapidly.

Construction and trades

Physical complexity and site variability make automation hard. But project management, estimating, and back-office functions are fair game.

Agriculture

Some automation in controlled environments; open-field farming remains complex. Supply chain and admin sides more exposed.

Hospitality and retail (physical)

In-person service experiences are the product. Back-office operations, scheduling, and inventory management are automating.

Now that you understand the landscape, let's assess where your business stands.

Answer honestly:

Role composition

• ➢What percentage of your workforce does routine knowledge work (analysis, processing, coordination, content creation)?
• ➢What would happen if that work became 50% faster or 30% cheaper at competitors?

Competitive dynamics

• ➢Are your competitors visibly investing in AI and automation?
• ➢Are you seeing pricing pressure that might indicate AI-enabled efficiency?
• ➢Are new entrants emerging with different cost structures?

Business model vulnerability

• ➢Do you bill for human time? (If yes, AI attacks your core revenue model)
• ➢Is your differentiation based on scale of human effort? (AI commoditises scale)
• ➢Could a smaller competitor with AI tools match your output quality?

Score each from 1 (not at all) to 5 (completely):

Scoring:

• ➢32-40: You're ahead. Focus on scaling what works.
• ➢24-31: You've started. Focus on deepening and measuring.
• ➢16-23: You're dabbling. Need to accelerate and systematise.
• ➢8-15: You're behind. This is urgent.

For each major workflow in your business, score:

Scoring criteria:

Repetitive (5 = daily, identical pattern; 1 = occasional, varied)

High-frequency, consistent tasks are ideal for automation.

Text-based (5 = all text/data; 1 = physical/relationship-dependent)

AI excels at text; struggles with physical world and nuanced relationships.

Reviewable (5 = easy to check output; 1 = hard to verify until too late)

You need to be able to catch AI errors before they cause damage.

Low failure cost (5 = mistakes easily fixed; 1 = errors are catastrophic)

Start where the downside is limited.

Clear owner (5 = one person accountable; 1 = shared/unclear ownership)

Automation needs someone responsible for monitoring and improving.

Top scorers (20+): Prime candidates for immediate AI pilots.

Mid-range (15-19): Worth exploring with appropriate safeguards.

Lower scores (<15): Address the limiting factors first, or deprioritise.

Let's be direct: some businesses can afford to wait. Others can't.

Your competitors are already moving. If you're seeing AI mentioned in competitor marketing, or noticing they're doing more with apparently less, the clock is ticking. First-mover advantages compound.

Your margins are thin. A 20% productivity improvement at a competitor translates to either price cuts you can't match or profit margins that fund further investment.

Your business depends on routine knowledge work. If you're selling human hours for tasks AI can partially automate, your pricing power is eroding.

Your industry is consolidating. In consolidating markets, efficiency advantages accelerate winner-take-more dynamics.

You're already having trouble hiring. AI tools can partially address talent shortages - but only if you adopt them.

Your differentiation is genuinely human. If clients pay for your judgement, relationships, or creativity specifically - not your output volume - you have more runway.

Your industry is heavily regulated. Regulated industries move slower. But don't confuse "slower" with "never" - healthcare, finance, and legal are all seeing significant AI adoption despite regulation.

You're already the efficiency leader. If you've already optimised your operations heavily, the marginal gains from AI might be smaller than for inefficient competitors.

Your competition is equally unprepared. If your whole industry is moving slowly, the penalty for waiting is lower. But industries can tip quickly once leaders emerge.

Consider the compound effects:

Year 1: Competitors who start now build basic capabilities and train their teams.

Year 2: Those competitors have refined their processes, proven ROI, and are scaling what works.

Year 3: The gap has widened. You're now trying to hire AI talent in a market where experienced people go to companies with established practices.

Year 4+: You're playing catch-up against competitors who've been compounding advantages for years.

This isn't fear-mongering. It's the pattern we've seen with every major technology shift - cloud, mobile, digital marketing. The companies that adopted early didn't just survive; they often displaced the cautious incumbents.

If this assessment has created some urgency, here's the smallest meaningful response:

1. Run the self-assessment above - Know your exposure and readiness score
2. Identify your top-scoring workflow - One place to start
3. Assign an owner - Someone accountable for exploring AI for that workflow

1. Run a real pilot - Not a demo, an actual workflow running with AI assistance
2. Measure the results - Time saved, quality maintained, issues encountered
3. Document what you learn - Create the foundation for scaling

1. Scale what works - Take successful pilots to full adoption
2. Build internal capability - At least one person who deeply understands your AI tools
3. Update your strategy - Factor AI into hiring, pricing, and competitive positioning

• ➢Budget: £1-5k for tools and experiments (to start)
• ➢Time: 4-8 hours/week from whoever owns this
• ➢Executive attention: Regular check-ins, not just annual strategy reviews
• ➢Willingness to iterate: The first approach rarely works perfectly

This assessment isn't meant to create panic. It's meant to create clarity.

Some businesses are genuinely at risk if they don't start adapting now. The combination of rapid AI improvement, falling costs, and competitor adoption creates a window that's closing.

Other businesses have more runway - but that doesn't mean infinite runway. The question isn't whether AI will affect your industry, but when and how much.

The companies that come out ahead won't be the ones with the biggest AI budgets or the fanciest tools. They'll be the ones that started early, learned fast, and built AI into how they actually work.

That process starts with an honest assessment of where you stand. You've just completed one. The question now is: what will you do with that information?

If you're unsure about your assessment or want help prioritising, we're happy to talk through it. No pitch, just perspective. Sometimes an outside view helps clarify what's actually urgent versus what can wait.