Are You an AI Zoomer, Bloomer, Gloomer or Doomer?
Zoomers, Bloomers, Gloomers and Doomers are attitude archetypes associated with Reid Hoffman's Superagency and examined in McKinsey's US employee research. They are not generations or diagnoses. This practical IT Club guide explains what each viewpoint contributes, why sceptical people may still use AI, how shadow AI appears, and how management can turn disagreement into safer adoption.
Show the same AI demonstration to four people.
- “Why aren't we using this already?”
- “This could be brilliant — if we introduce it properly.”
- “What happens when it gets something important wrong?”
- “Why are we allowing this technology at all?”
Same technology. Four very different reactions. They have been described as Zoomers, Bloomers, Gloomers and Doomers.
Which one are you?
And perhaps more importantly: which ones work for you? These labels are a useful way to discuss attitudes towards AI. They are not psychological diagnoses, scientifically fixed personality types, demographic generations or an IT Club invention. Most people will not fit perfectly into one box.
What are the four attitudes towards AI?
Reid Hoffman used the four archetypes in his discussion of AI and the future in Superagency. McKinsey subsequently used the same labels in US employee research about generative-AI sentiment. The names are memorable; the value for a small business is practical. They give people a less personal way to explain what they want, what they fear and what controls they need.
⚡ ZOOMER — “Why isn't it live?”
A Zoomer is enthusiastic about AI and sees competitive opportunity in moving quickly.
- Likely to experiment early, adopt new tools and spot automation opportunities
- Comfortable tolerating uncertainty while a useful idea is being explored
- Strength: momentum and opportunity spotting
- Watch out for: shadow AI, uncontrolled automation, data exposure and moving faster than governance can support
Enthusiasm does not mean recklessness. A Zoomer may be the person who finds the valuable use case before anyone else has seen it.
🌱 BLOOMER — “Let's test it properly.”
A Bloomer is optimistic about AI but wants responsible development, measurement and sensible controls around it.
- Likely to run pilots, involve people, measure results and expand what works
- Interested in opportunity without treating every demonstration as a business case
- Strength: balanced adoption
- Watch out for: process becoming so cautious that a useful experiment never starts
Bloomers are often the bridge between a bold idea and a safe, evidence-based trial.
🌧️ GLOOMER — “Show me what can go wrong.”
A Gloomer is concerned about consequences and believes AI needs strong oversight. That does not make a Gloomer anti-AI.
- Likely to ask about jobs, privacy, accuracy, hallucinations, bias and cyber security
- Wants to know who checks the output and what happens when it fails
- Strength: risk identification and challenge
- Watch out for: fear or analysis paralysis preventing useful experimentation
Research shows that many people described as Gloomers still use, understand or feel comfortable using generative AI. Sceptical about AI does not necessarily mean unwilling to use AI.
☠️ DOOMER — “Should we be doing this?”
A Doomer is deeply concerned that AI's societal or organisational risks may outweigh its benefits.
- Likely to question human control, employment disruption, misinformation, security and autonomous systems
- May raise existential or systemic questions that a product demonstration avoids
- Strength: forces difficult assumptions and consequences to be considered
- Watch out for: blanket rejection that drives AI use underground rather than stopping it
Do not ridicule this position. A difficult question can be valuable even when the business ultimately decides to proceed.
What does McKinsey's US research say?
McKinsey's January 2025 report, Superagency in the workplace, used a US employee survey from October–November 2024 and asked respondents to self-identify with the four attitudes. It reported the following distribution:
| Attitude | Share of US employees | Broad description |
|---|---|---|
| Bloomer | 39% | AI optimist who wants to collaborate on responsible solutions |
| Gloomer | 37% | More sceptical and wants extensive top-down controls |
| Zoomer | 20% | Wants AI deployed quickly with few guardrails |
| Doomer | 4% | Has a fundamentally negative view of AI |
These are reported proportions from McKinsey's US employee research. They do not describe all UK workers, all businesses, global opinion or IT Club readers.
The really interesting finding is not simply that the workplace contains different opinions. McKinsey reported that 94% of Gloomers and 71% of Doomers said they had some familiarity with generative-AI tools. Approximately 80% of Gloomers and about half of Doomers said they were comfortable using generative AI at work.
In other words, the categories are not simply AI users versus AI non-users. Someone can use AI regularly and still be worried about where it is heading. Someone can think AI is transformative while also demanding controls.
Usage and attitude are not the same thing
McKinsey also reported a gap between employees and leaders. C-suite leaders estimated that only 4% of employees used generative AI for at least 30% of daily work, while the employee figure was three times greater. Looking ahead, 20% of leaders believed employees would use generative AI for more than 30% of daily tasks within a year, compared with 47% of employees who thought they would.
That is a US survey finding, not a forecast for your business. It does suggest a useful management question: is AI adoption already happening from the bottom up while leadership is still discussing whether to begin?
Your business probably has all four
Imagine a management meeting about an AI tool that could summarise customer calls.
| Voice | What it says | What it contributes |
|---|---|---|
| Zoomer | “Automate it.” | Experimentation, speed and opportunity spotting |
| Bloomer | “Pilot it.” | A measured test, adoption and feedback |
| Gloomer | “Control it.” | Risk identification, governance and checking |
| Doomer | “Should we be doing it?” | Questions about consequences and assumptions |
The mistake is assuming one group needs to defeat the others. A good AI adoption programme can use the strengths of each.
Do not try to turn everyone into a Zoomer
AI adoption does not require everyone to become an enthusiast. A room full of Zoomers may move quickly but miss important risks. A room full of Gloomers may identify every possible problem but never deploy anything. A room containing different attitudes can be useful if disagreement becomes part of the governance process rather than organisational warfare.
You need people who ask “Can we?” and people who ask “Should we?”
The best meeting does not end with the loudest viewpoint winning. It ends with a clear use case, a proportionate test, a data boundary, a named owner, an agreed human check and a decision about what would make the business stop.
The Shadow AI problem
Management may be cautious about AI while employees are already using ChatGPT, Microsoft Copilot, Gemini, Claude or AI built into software the business already pays for. Ignoring AI does not necessarily mean no AI. It may mean unmanaged AI.
Likewise, banning AI without understanding why employees find it useful may push use underground. A safer response is to make it easy to declare what is being used, what information goes into it and what work it is helping with. That gives management something real to assess.
The absence of an AI strategy does not mean the absence of AI.
Read: Shadow AI — The Tools Your Team Isn't Telling You About →
Consequence changes attitude
The same person may have different attitudes for different tasks. They might be a Zoomer about AI meeting notes but a Gloomer about AI recruitment decisions. That nuance is rational: the more serious the consequence, the more caution becomes sensible.
| Use case | Reasonable starting posture |
|---|---|
| AI writes the first draft of an internal email | More room for experimentation, with a human reading before sending |
| AI summarises a meeting | Pilot with clear data boundaries and a way for people to correct the record |
| AI recommends who gets a job | Much stronger controls, documented human oversight and a careful assessment before use |
| AI can take an external action | Constrained permissions, approval points, logging and a safe fallback |
The more serious the consequence, the more caution becomes rational. This is why AI governance should be matched to the use case rather than imposed as one blanket reaction.
Read: The Unauthorised AI Agent Problem: When AI Takes Actions Nobody Approved →
Read: AI Accountability: Who Is Responsible When AI Makes a Decision? →
Which AI type are you? A lightweight self-assessment
Which statement sounds most like you?
- A. “If AI can do it, let's use it.” → Zoomer tendency
- B. “Let's test it, measure it and put sensible controls around it.” → Bloomer tendency
- C. “Before we deploy it, show me what can go wrong.” → Gloomer tendency
- D. “I remain unconvinced that the benefits justify the risks.” → Doomer tendency
Most people will not fit perfectly into one box. Your attitude may also change depending on the use case, the data involved and who bears the consequence if the system is wrong.
The IT Club AI attitude matrix
The following is an illustrative IT Club framework, not a scientific measurement. It separates attitude from actual usage. Someone can be cautious and actively use AI with controls; someone else can be enthusiastic but barely use it because the business has not made a safe route available.
| Cautious | Mixed | Enthusiastic | |
|---|---|---|---|
| Active use | Controlled Gloomer: uses AI but wants evidence and oversight | Practical Bloomer: tests, measures and improves | Fast Zoomer: experiments and spots opportunity |
| Passive / little use | Concerned Doomer: sees more risk than benefit | Waiting Bloomer: interested but needs a safe pilot | Blocked Zoomer: wants to move but lacks permission or support |
Attitude and usage are not the same thing. Do not infer a person's beliefs from a single tool, task or training session.
What should management do with each viewpoint?
| Viewpoint | Useful management response |
|---|---|
| Zoomers | Give controlled experimentation space. Define approved tools, data boundaries, approval points and a stop condition without crushing initiative. |
| Bloomers | Use them as potential AI champions to run pilots, collect feedback, measure results and help colleagues. |
| Gloomers | Invite the challenge. Ask what could fail and use the concerns to improve security, privacy, human oversight and fallback procedures. |
| Doomers | Listen to the underlying concern. Distinguish evidence-based risk from blanket rejection without dismissing the person as a technophobe. |
A business exercise for your next management meeting
Ask everyone to answer these questions individually before comparing the responses:
- 1Which attitude best describes you today?
- 2Where is AI already being used in the business?
- 3What are you excited about?
- 4What worries you?
- 5Which tasks should AI assist with?
- 6Which decisions should remain human?
- 7What controls would make you more comfortable?
The disagreement may tell you more about your AI readiness than another product demonstration.
Look for the gaps between what leaders think is happening and what staff say is happening. Then choose one small, bounded use case and turn the concerns into test criteria.
AI adoption needs an Operational Heartbeat too
AI attitudes and usage change because tools change, staff gain experience, risks change, policies change and new capabilities appear. A one-off AI policy or demonstration will not keep the picture current.
- What AI are we using?
- What AI do staff want to use?
- What is working?
- What is worrying people?
- What is happening outside policy?
- What should we test next?
Review what people are using, what is working, what worries them and where unmanaged AI is appearing. Give the review an owner and a date, then revisit it when the business changes.
The IT Club view
There probably is not one “correct” attitude towards AI. AI contains real opportunity and real risk. Businesses need enough enthusiasm to experiment and enough scepticism to avoid doing stupid things with powerful technology.
You do not need everyone to love AI.
You need everyone to understand how your business intends to use it. The healthiest AI strategy may need a Zoomer to ask “why not?”, a Gloomer to ask “what if?”, and a Bloomer to work out how to test it properly.
Frequently Asked Questions
What is an AI Zoomer?
A Zoomer is an AI-attitude archetype for someone who wants AI deployed quickly and sees opportunity in moving faster. It is not a generation, diagnosis or fixed personality type.
What is an AI Bloomer?
A Bloomer is optimistic about AI but wants to test it, measure results and introduce sensible controls. Bloomers can help turn enthusiasm into responsible adoption.
What is an AI Gloomer?
A Gloomer focuses on what could go wrong and asks for oversight around accuracy, privacy, security, jobs and bias. Gloomers are not necessarily anti-AI; many still use and feel comfortable with AI.
What is an AI Doomer?
A Doomer believes AI's societal or organisational risks may outweigh the benefits and asks whether some uses should happen at all. That challenge can surface assumptions that a purely enthusiastic conversation misses.
Where did Zoomers, Bloomers, Gloomers and Doomers come from?
The four labels are associated with Reid Hoffman's discussions of AI in Superagency. McKinsey used the archetypes in its January 2025 report on AI in the workplace and surveyed US employees about their attitudes.
Which AI attitude is most common?
In McKinsey's reported US employee research, Bloomers were the largest group at 39%, followed by Gloomers at 37%, Zoomers at 20% and Doomers at 4%. Those figures should not be generalised to UK workers, every business or IT Club readers.
Are Gloomers anti-AI?
No. Gloomers are more sceptical and want stronger oversight, but McKinsey reported that 94% of Gloomers in its US research had some familiarity with generative AI and approximately 80% were comfortable using it at work.
Do AI sceptics still use AI?
They can. Attitude and usage are not the same thing. Someone may use AI for low-consequence drafting while rejecting AI making recruitment or safety decisions.
Why are employees and managers different about AI?
Managers may see policy, risk and accountability; employees may see the daily task and the time it could save. McKinsey reported a gap between what US leaders believed employees would use and what employees expected to use, which makes direct conversation valuable.
What is Shadow AI?
Shadow AI is AI use inside a business that has not been approved, assessed or recorded. It may involve public chat tools, AI browser extensions or AI features inside existing SaaS products.
Should businesses ban unapproved AI?
A ban may be necessary for a particular high-risk use, but a blanket ban without understanding staff needs can push use underground. Start by making it safe to declare tools, then assess data, purpose, settings, supplier terms and consequence.
How should managers deal with employees who resist AI?
Ask what the concern is and what evidence or control would make the use acceptable. Do not treat a question about privacy, accuracy or job impact as a refusal to learn. A bounded pilot with a human fallback can answer a specific question better than pressure can.
Do businesses need everyone to embrace AI?
No. They need a shared understanding of where AI is permitted, what information may go into it, who checks the output and which decisions remain human. Different levels of enthusiasm can coexist with clear rules.
Can someone have different AI attitudes for different tasks?
Yes. Consequence matters. A person may be enthusiastic about meeting notes but cautious about recruitment, credit, medical or safety decisions. That is often a sign that the conversation is becoming more specific and useful.
How should SMEs introduce AI safely?
Choose a bounded use case, define the data boundary, name an owner, agree human checks, measure value, record failure modes and set a stop condition. Include both people who ask “can we?” and people who ask “should we?”
Sources and Further Reading
McKinsey — Superagency in the workplace: Empowering people to unlock AI's full potential →
McKinsey — From doom to zoom in AI →
Read: Shadow AI — The Tools Your Team Isn't Telling You About →
Read: Why AI Needs Guardrails →
Plain-English Takeaway
The healthiest AI strategy may need a Zoomer to ask “why not?”, a Gloomer to ask “what if?”, and a Bloomer to work out how to test it properly. You do not need everyone to love AI. You need everyone to understand how your business intends to use it.
Tap to follow The IT Club Briefing on WhatsApp.
Related Articles
Google Maps Is Becoming an AI Agent — Why Local Businesses Should Care
Google Maps is evolving from a place-search tool into a conversational AI assistant that can help users find businesses and, in supported situations, move towards ordering, booking or reserving. The shift matters because an AI agent may narrow the customer's choices before they ever see the map.
Read articleThe Race to Replace: Why AI Needs to Start Doing the Work
Generative AI started by answering questions. The bigger economic shift may come when AI agents take responsibility for pieces of real work — and jobs are bundles of tasks, not indivisible units.
Read articleAmerica Loves AI — So Why Does It Hate the Data Centres That Power It?
Seven in ten Americans oppose an AI data centre being built in their local area — while using AI services dozens of times a week. This article examines the growing tension between AI demand and the physical infrastructure required to run it, and why the question matters for businesses in the UK.
Read article