One-Sentence Takeaway

Use AI as a B-minus assistant and performance coach to compress back-office work, while restaurant leaders keep responsibility for hospitality, problem solving, decisions, and imagining what comes next.

Summary

AI is becoming part of the normal fabric of work, much as PCs, spreadsheets, the internet, mobile devices, cloud tools, data science, and remote work did before it. The AutoCAD story established the keynote's central distinction: technology can create extraordinary leverage as an assistant without becoming the expert, decision-maker, or accountable professional.

Generative AI predicts likely output from patterns in existing information. That makes it broadly useful, but also probabilistic, consensus-driven, and vulnerable to weak source material. Dan described AI as a B-minus assistant that can raise a person's baseline across unfamiliar tasks. Restaurant leaders still provide the A-plus expertise, operating context, memories, hopes, judgment, and standards for hospitality.

For King's Seafood leaders, the practical opportunity is to reduce back-office drag, strengthen coaching, improve recurring work, and make useful knowledge easier to reach across locations. The goal is not to transfer guest, employee, staffing, financial, or operational decisions to AI. It is to create more capacity for leaders to stay present on the floor, solve hard problems, make thoughtful decisions, develop their teams, and imagine better guest experiences.

Action Items

[ ] Identify one recurring administrative task that takes time away from guests, floor leadership, coaching, or difficult operating decisions.

[ ] Test how an approved AI assistant could summarize, compare, organize, prepare, or critique that work without placing recipes, financials, employee information, or other confidential material in an unapproved tool.

[ ] Complete one important message, plan, schedule, appraisal, coaching conversation, or guest response yourself, then ask AI to compare it with prior strong examples and explain how the next version could be better.

[ ] Set a recurring prompt that gives you one new AI-supported idea for your specific role, hospitality outcomes, and labor, margin, time, consistency, and governance constraints.

[ ] Before acting on consequential AI-supported work, inspect its assumptions, source quality, missing context, and uncertainty. Keep the final decision human, track whether the workflow creates real capacity, and reinvest that time in guests and team leadership.

Key Ideas

Assistant, not automator

AutoCAD helped an architect create precise work faster without becoming an architect. AI should be framed the same way: people supply intent, expertise, review, judgment, and accountability.

AI is a B-minus assistant

AI can raise a person's baseline in research, communication, planning, facilitation, and other supporting work. It should extend the expertise a leader owns, not replace it.

Probability is not certainty

AI chooses the most likely answer even when its confidence may be lower than the situation requires. A polished response does not reveal the uncertainty, source quality, or missing context underneath it.

The internet is part of the training data

The glue-in-the-pizza example showed how jokes, sarcasm, trolls, novices, and experts can become mixed source material. A popular or plausible answer is not automatically a trustworthy answer.

Treat output as a recommendation

Autocomplete provides a familiar model: accept useful suggestions, reject weak ones, and continue with the work you intended to do. More consequential work requires the same control with stronger review.

Memories and hopes shape judgment

People make decisions with experience, context, relationships, and a future they want to create. Those inputs differ across leaders and cannot be reduced to probability alone.

Use AI as a performance coach

Do not only ask AI to create work. Ask it to compare new work with prior strong work, rank it against clear criteria, identify gaps, and coach the next iteration.

AI is more like an intern than a copilot

Give it instructions, review the work, improve the instructions, and learn its capability ceiling. The leader remains responsible for the quality and consequences of everything the assistant touches.

Protect the top of the work pyramid

AI is strongest at helping with communication, process, and investigation. The capacity created should be reinvested in solving novel problems, making judgment-rich decisions, and imagining better outcomes.

Create more floor presence

Back-office efficiency matters when it creates more space for guests, coaching, service recovery, team leadership, and the operating details that protect hospitality.

King’s Seafood Opportunities

From back-office drag to floor presence

• Predict shift demand with AI across approved operating signals.

• Draft schedules around AI-predicted staffing gaps.

• Turn P&Ls into AI-generated action briefs before shifts.

From manager instinct to guided leadership

• Generate coaching options for cross-generational conversations.

• Surface performance patterns to improve appraisals.

• Rehearse role-aware coaching before difficult conversations.

From scattered prompting to governed improvement

• Ground AI answers in approved restaurant standards.

• Route sensitive prompts through privacy review guardrails.

• Compare manager work against AI-built performance baselines.

These are candidates for governed exploration. King's Seafood leaders remain responsible for approved tools and sources, confidentiality, employee decisions, guest experience, financial judgment, operating standards, and the consequences of action.

Talk Flow

Technology assistance changes the baseline

• The AI-generated opening

  • The quality threshold for AI-supported work is moving quickly.

  • AI assistance can place information and coaching into the work as it happens.

• The AutoCAD copy-and-paste moment

  • Work that once took days could be completed in seconds with greater precision.

  • Once technology assistance becomes dramatically better, returning to the old way feels unreasonable.

• Assistant vs automator

  • AutoCAD increased an architect's leverage without becoming an architect.

  • AI should help experts create more value while people retain responsibility for the work.

• The technology staircase

  • PCs, spreadsheets, the internet, mobile devices, cloud tools, data science, and remote work each changed how work gets done.

  • Generative AI is the next copy-paste moment moving into the normal operating baseline.

Understand the assistant before trusting it

• Cracking open the AI piñata

  • Leaders need to separate useful assistance from confusion, risk, and novelty.

  • The goal is to find applications that create more value than friction

• Prediction, consensus, and the B-minus assistant

  • The `Once upon a time` exercise showed AI predicting what comes next.

  • Consensus-driven output tends toward the middle of the available answers.

  • AI can raise a person's baseline in many supporting tasks without replacing their chosen expertise.

• Confidence and hallucinations

  • AI selects the most likely answer even when its confidence may be lower than the situation requires.

  • Fluent output does not reveal the uncertainty underneath it.

  • Review should become stronger as the consequences increase.

• Glue in the pizza

  • Training material contains experts, novices, jokes, sarcasm, and deliberate nonsense.

  • A popular or plausible answer is not automatically a trustworthy one.

• Autocomplete and human judgment

  • Accept useful suggestions, reject weak ones, and continue with the work you intended to do.

  • Human decisions also depend on memories, hopes, context, relationships, and the future people want to create.

Apply AI to restaurant leadership

• The restaurant-leadership question

  • Ask how AI will change the work of restaurant GMs and operators, not only which tools are popular.

  • Include guest value, leadership capacity, consistency, labor, margin, time, and governance in the question.

• Three restaurant-leadership shifts

  • Move from back-office drag to floor presence.

  • Move from manager instinct to guided leadership.

  • Move from scattered prompting to governed improvement.

• A restaurant day in the life

  • Approved AI assistance can help prepare leaders, surface patterns, support coaching, and coordinate work across locations.

  • People remain responsible for hospitality, judgment, standards, and action.

• AI before the large technology rollout

  • Some capabilities depend on internal platforms and vendor roadmaps.

  • Leaders can begin with approved assistance for recurring writing, research, planning, workflow, and translation work.

• AI as a performance coach

  • Complete important work yourself, then ask AI to compare it with prior strong examples.

  • Use ranking and critique to improve the next attempt instead of delegating away the expertise.

Keep people at the top of the work pyramid

• Copilot vs intern

  • An intern receives instructions, attempts work, receives review, and improves through feedback.

  • Repeated review reveals the assistant's capability ceiling.

  • The leader remains accountable for everything the assistant touches.

• The lower work layers

  • Communication, process, and investigation form the broad base of the work pyramid.

  • AI can help organize information and compress parts of those layers.

• Solve, decide, and imagine

  • People remain responsible for novel problems that require local expertise.

  • Decisions require judgment, context, preferences, and responsibility.

  • Imagining requires choosing which possible future deserves commitment.

• Communication and searchable notes

  • AI-supported notes can preserve useful context and make it searchable.

  • Notes can help people learn from conversations they did not attend.

• Notes as a song

  • AI can repackage the same information into a form people are more likely to remember and revisit.

  • The song reinforced the solve-decide-imagine refrain.

Create more capacity for hospitality

• Translation and access

  • AI can reduce language barriers across restaurant teams and guest interactions.

  • People can communicate in the language most comfortable for them.

• Managing in a world of more

  • Information, expectations, staffing demands, training needs, and possible work keep growing.

  • The hours and resources available to leaders remain constrained.

• Keep discovering useful assistance

  • Ask for one new role-specific AI idea on a recurring cadence.

  • Test useful ideas within approved tools, information, and review boundaries.

• Reinvest the capacity

  • Compress lower-layer work without removing valuable human interaction.

  • Use the resulting time for guests, team development, problem solving, decision making, and imagining better hospitality.

Closing Message - Managing in the world of more

Information, systems, expectations, and possible work keep growing while available hours remain fixed. The closing challenge was to flip the work pyramid and reinvest saved capacity in solving, deciding, and imagining the future of design, make, and engineering.