
São Paulo, Brazil
KPMG South America Annual Partners Meeting
Clock Drift — Leading Transformation in the Age of AI Participation

What this talk is about
At KPMG South America’s Annual Partners Meeting, Neil Redding addressed the leadership question underneath every firm’s AI transformation: how do you lead as AI shifts from tool to participant, while the room quietly fills with AI agents? Drawing on thirty years of technology leadership, Neil showed why the pace of execution has outrun the pace of decision, what he calls Clock Drift, and what leadership must become in response: the Orchestration of Participant Dynamics. Partners left with shared language for the shift, and a practical first move toward 2030.
Take the playbook with you.
The talk was the start of a conversation, not the end of one. This is how you carry it into the week — three principles, three practices, and one shift in how you lead when AI becomes a participant rather than a tool.
For the partners at KPMG South America: the most credible way to help clients redesign how they work is to have done it inside your own firm first.
Here are the AI-generated notes.
These notes were generated from the keynote transcript (São Paulo, November 17, 2026).
One Sentence Takeaway
When AI can execute faster than a firm can decide, the answer isn’t tighter control. Redesign who decides what, treat the firm’s knowledge as its most defensible asset, and lead the whole system of people and AI agents as one orchestra.
Summary
Neil Redding opened KPMG South America’s Annual Partners Meeting by naming what broke: not strategy, but speed. Inside most firms, work still moves through specs, reviews and committees, while on everyone’s laptop AI-enabled work moves at a different tempo. Your best people can build a working tool in an afternoon, then wait weeks for approval. He calls that gap Clock Drift, and it turns a firm’s operating model into an attack surface: if you can’t decide at the speed of execution, someone else will execute your strategy first.
The shift underneath is from tool to participant. With tools we do tasks; with participants we co-create outcomes, and the difference is reasoning. Neil showed how he asked Claude in PowerPoint how his talk would land with a room of KPMG partners, and it warned him that careful review is the product they sell. Participants govern the potential of the system, and the nature of the interactions determines the nature of the outcomes.
Context is the only moat as AI commoditizes everything that isn’t proprietary. Neil asked what the firm would forget if a senior partner retired tomorrow, showed the context layer he built for his own business with Claude Code, and described the work of improving every relationship between people and AI agents as relationship engineering.
Those relationships come down to decisions. Professionals are moving from Doers to Deciders, and Neil’s first move is Delegate First. He closed with orchestration, the leadership model for a system of humans and AI, demonstrated a track, evaluate, reallocate, retune cycle, and said orchestration turns Clock Drift into Clock Drive.
Action Items
Find one approval that takes weeks for work AI can do in an afternoon. Time it, and ask who could decide it sooner.
Find out where shadow execution already happens in your practice: which teams use AI tools nobody approved, and what they built. Then give them rules, an owner and a review point.
Rewrite one recurring prompt from a task into a participant brief: what you’re trying to do, what can’t break and what to watch for.
Capture what a senior partner knows: why a client really stays, how each regulator reads things, what was decided in the hallway. Start your context layer there.
Before starting a piece of work you’ve done a thousand times, ask how you could delegate it to an agent or a team of agents. Start with low-stakes work, the way you would with a new hire.
Redesign one role around the new capabilities, starting with who reviews agents’ work and how.
Name someone on your leadership team to track how AI capabilities are changing and report what that makes possible.
Key Ideas
Clock Drift is about speed, not strategy
AI can now execute faster than organizations can decide. Samsung banned ChatGPT about 20 days after approving it, after three leaks of proprietary information. This February, two CNBC reporters with no technical background built a working replacement for Monday.com on air in about an hour for about $15, and the software sector fell about 30% in the following weeks. The problem isn’t that AI moves too fast. It’s that organizations fail to move with it.
Shadow execution is the natural response
Mark Pike, associate general counsel at Anthropic and a lawyer rather than an engineer, spent an afternoon describing to Claude what a good legal review looks like. The plugin he built now handles marketing copy review, contract redlining and privacy impact assessments for his team, cutting turnaround from days to hours. His version was approved. Studies put the share of employees using unapproved AI tools at about half, often because waiting for approval takes longer than doing the work.
Tools do tasks; participants co-create outcomes
Asking AI to summarize Q2 sales data into a slide deck is a task. Asking for the strongest case for doubling down on the mid-market, and what would undermine it, invites a reasoning partner. Framed as a participant, AI is told what you’re trying to do, what can’t break and what to watch for, and the work moves from command to collaboration and from one-off prompts to shared context.
AI rises to our level
Anthropic reported this year that how people prompt is how Claude responds, down to education level. The ceiling on the value of AI is no longer the model’s capability but the quality of the context and framing people bring. That is why participants govern the potential of the system.
Interactions become dynamics
In a traditional executive decision, a CEO requests analysis, reports and committees follow, and signals are missed for weeks. In an evolved pattern, the CEO briefs intent and constraints, AI analyzes continuously and people decide while the options are still live. The people, strategy and AI are the same; the interaction pattern changed, and so did the outcome. As these patterns compound, systems show new dynamics: emergence, reciprocity and adaptation, with AI now helping build the next generation of AI and handling much longer autonomous work.
Context is the only moat
What differentiates a firm is its proprietary context: client relationships, internal decisions and deep knowledge of its services and clients’ needs. Most of it lives in people’s heads. Neil showed the context layer he built with Claude Code, drawing in documents, CRM records, meeting transcripts, chat history and voice notes. A simpler start: ask your chatbot how its context window is looking, then save the context of the conversation to a file.
Relationship engineering
The industry calls it context engineering; Neil calls it relationship engineering, because the unit of improvement is no longer the prompt but the relationship among all participants, from people to AI agents to robots. Like any relationship it takes work: context coherence, clear trust and delegation, guardrails and governance, and human review of the right things at the right moments, not everything.
From Doers to Deciders
Professionals are moving from mostly doing to mostly deciding, at every level of a firm, not just at the top. That changes what a firm hires for, what it trains for and what a career path looks like. A Deloitte study found 84% of surveyed companies had not redesigned a single job around these capabilities. For an advisory firm that is both a market opportunity and a question about its own practice, since the most credible way to help clients redesign how they work is to have done it yourself.
Delegate First
Before starting a piece of work, especially one you’ve done a thousand times, ask how you could delegate it to an agent or a team of agents. OpenAI announced this year that it was going agent first across all of its work. The idea is to make smaller mistakes, faster, with access granted the way you would for a new hire.
Leadership is the orchestration of participant dynamics
A conductor doesn’t command each musician. Everyone plays from the same score, the shared context, and the conductor keeps tuning as instruments change mid-performance. For a firm working across several markets, each moving at its own speed, the conductor doesn’t make every instrument play the same part. In practice, orchestration is a continuous cycle: track evolving capabilities, evaluate how outcomes change, reallocate responsibility between people and AI, and retune the interaction patterns.
Closing Message: Orchestration turns Clock Drift into Clock Drive
The real risk isn’t moving too slowly. It’s optimizing a business model the future no longer needs, and the real opportunity is exploration. Neil closed with mycelium, the forest’s original context layer, connecting slow trees and fast fungi and reallocating resources across the ecosystem as a model for a business made of people, AI agents and machines, all sensing, deciding and acting together. The decision clock and the execution clock will keep running at different speeds. Leaders can harness them so decisions stay deliberate while execution runs at AI speed: “Leadership is orchestration. And orchestration turns Clock Drift into Clock Drive.”
These notes were generated from the keynote transcript (São Paulo, November 17, 2026).
One Sentence Takeaway
When AI can execute faster than a firm can decide, the answer isn’t tighter control. Redesign who decides what, treat the firm’s knowledge as its most defensible asset, and lead the whole system of people and AI agents as one orchestra.
Summary
Neil Redding opened KPMG South America’s Annual Partners Meeting by naming what broke: not strategy, but speed. Inside most firms, work still moves through specs, reviews and committees, while on everyone’s laptop AI-enabled work moves at a different tempo. Your best people can build a working tool in an afternoon, then wait weeks for approval. He calls that gap Clock Drift, and it turns a firm’s operating model into an attack surface: if you can’t decide at the speed of execution, someone else will execute your strategy first.
The shift underneath is from tool to participant. With tools we do tasks; with participants we co-create outcomes, and the difference is reasoning. Neil showed how he asked Claude in PowerPoint how his talk would land with a room of KPMG partners, and it warned him that careful review is the product they sell. Participants govern the potential of the system, and the nature of the interactions determines the nature of the outcomes.
Context is the only moat as AI commoditizes everything that isn’t proprietary. Neil asked what the firm would forget if a senior partner retired tomorrow, showed the context layer he built for his own business with Claude Code, and described the work of improving every relationship between people and AI agents as relationship engineering.
Those relationships come down to decisions. Professionals are moving from Doers to Deciders, and Neil’s first move is Delegate First. He closed with orchestration, the leadership model for a system of humans and AI, demonstrated a track, evaluate, reallocate, retune cycle, and said orchestration turns Clock Drift into Clock Drive.
Action Items
Find one approval that takes weeks for work AI can do in an afternoon. Time it, and ask who could decide it sooner.
Find out where shadow execution already happens in your practice: which teams use AI tools nobody approved, and what they built. Then give them rules, an owner and a review point.
Rewrite one recurring prompt from a task into a participant brief: what you’re trying to do, what can’t break and what to watch for.
Capture what a senior partner knows: why a client really stays, how each regulator reads things, what was decided in the hallway. Start your context layer there.
Before starting a piece of work you’ve done a thousand times, ask how you could delegate it to an agent or a team of agents. Start with low-stakes work, the way you would with a new hire.
Redesign one role around the new capabilities, starting with who reviews agents’ work and how.
Name someone on your leadership team to track how AI capabilities are changing and report what that makes possible.
Key Ideas
Clock Drift is about speed, not strategy
AI can now execute faster than organizations can decide. Samsung banned ChatGPT about 20 days after approving it, after three leaks of proprietary information. This February, two CNBC reporters with no technical background built a working replacement for Monday.com on air in about an hour for about $15, and the software sector fell about 30% in the following weeks. The problem isn’t that AI moves too fast. It’s that organizations fail to move with it.
Shadow execution is the natural response
Mark Pike, associate general counsel at Anthropic and a lawyer rather than an engineer, spent an afternoon describing to Claude what a good legal review looks like. The plugin he built now handles marketing copy review, contract redlining and privacy impact assessments for his team, cutting turnaround from days to hours. His version was approved. Studies put the share of employees using unapproved AI tools at about half, often because waiting for approval takes longer than doing the work.
Tools do tasks; participants co-create outcomes
Asking AI to summarize Q2 sales data into a slide deck is a task. Asking for the strongest case for doubling down on the mid-market, and what would undermine it, invites a reasoning partner. Framed as a participant, AI is told what you’re trying to do, what can’t break and what to watch for, and the work moves from command to collaboration and from one-off prompts to shared context.
AI rises to our level
Anthropic reported this year that how people prompt is how Claude responds, down to education level. The ceiling on the value of AI is no longer the model’s capability but the quality of the context and framing people bring. That is why participants govern the potential of the system.
Interactions become dynamics
In a traditional executive decision, a CEO requests analysis, reports and committees follow, and signals are missed for weeks. In an evolved pattern, the CEO briefs intent and constraints, AI analyzes continuously and people decide while the options are still live. The people, strategy and AI are the same; the interaction pattern changed, and so did the outcome. As these patterns compound, systems show new dynamics: emergence, reciprocity and adaptation, with AI now helping build the next generation of AI and handling much longer autonomous work.
Context is the only moat
What differentiates a firm is its proprietary context: client relationships, internal decisions and deep knowledge of its services and clients’ needs. Most of it lives in people’s heads. Neil showed the context layer he built with Claude Code, drawing in documents, CRM records, meeting transcripts, chat history and voice notes. A simpler start: ask your chatbot how its context window is looking, then save the context of the conversation to a file.
Relationship engineering
The industry calls it context engineering; Neil calls it relationship engineering, because the unit of improvement is no longer the prompt but the relationship among all participants, from people to AI agents to robots. Like any relationship it takes work: context coherence, clear trust and delegation, guardrails and governance, and human review of the right things at the right moments, not everything.
From Doers to Deciders
Professionals are moving from mostly doing to mostly deciding, at every level of a firm, not just at the top. That changes what a firm hires for, what it trains for and what a career path looks like. A Deloitte study found 84% of surveyed companies had not redesigned a single job around these capabilities. For an advisory firm that is both a market opportunity and a question about its own practice, since the most credible way to help clients redesign how they work is to have done it yourself.
Delegate First
Before starting a piece of work, especially one you’ve done a thousand times, ask how you could delegate it to an agent or a team of agents. OpenAI announced this year that it was going agent first across all of its work. The idea is to make smaller mistakes, faster, with access granted the way you would for a new hire.
Leadership is the orchestration of participant dynamics
A conductor doesn’t command each musician. Everyone plays from the same score, the shared context, and the conductor keeps tuning as instruments change mid-performance. For a firm working across several markets, each moving at its own speed, the conductor doesn’t make every instrument play the same part. In practice, orchestration is a continuous cycle: track evolving capabilities, evaluate how outcomes change, reallocate responsibility between people and AI, and retune the interaction patterns.
Closing Message: Orchestration turns Clock Drift into Clock Drive
The real risk isn’t moving too slowly. It’s optimizing a business model the future no longer needs, and the real opportunity is exploration. Neil closed with mycelium, the forest’s original context layer, connecting slow trees and fast fungi and reallocating resources across the ecosystem as a model for a business made of people, AI agents and machines, all sensing, deciding and acting together. The decision clock and the execution clock will keep running at different speeds. Leaders can harness them so decisions stay deliberate while execution runs at AI speed: “Leadership is orchestration. And orchestration turns Clock Drift into Clock Drive.”
Here’s how to bring this thinking into your organization.
Clock Drift
Clock Drift is the gap between how fast AI can execute and how fast an organization can decide. AI now acts in seconds while many companies still decide in weeks. The wider the drift, the more a business exposes itself to shadow systems, missed openings, and strategic misalignment.
AI as a participant, not a tool
Treating AI as a tool limits you to what you already know to ask for. Treating it as a participant engages its reasoning, its context, and what emerges between you. The ceiling on the outcome is no longer the model’s capability. It’s the quality of your intent, your context, and your framing.
Orchestration
Orchestration is the leader’s new core work: setting intent, tracking evolving capabilities, evaluating outcomes, reallocating responsibility, and retuning the patterns as conditions change. The shift is from managing people alone to conducting a system of human and AI participants together.
Questions leaders ask about this
What is Clock Drift?
Why does Clock Drift matter for professional services firms?
What does it mean to treat AI as a participant instead of a tool?
What is shadow execution, and how should a firm respond?
What happens to junior staff when AI does the first draft?
What is Delegate First?
What does orchestration look like in practice?
How do professionals deal with AI doing work they spent years mastering?

About Neil
Neil Redding is The Near Futurist™, helping leaders turn emerging technology into practical business value. He founded Redding Futures after building Near Future practices at Gensler and Thoughtworks, and has keynoted SXSW, the Dubai Future Forum, and events for Meta and Microsoft.
Recent events

ITForum CIO Summit 2026
Neil addresses the Top CIOs in Brazil

SPIW 2026 highlights
Neil at São Paulo Innovation Week 2026

SXSW 2026 summary
Neil's extensive SXSW 2026 participation.

Agentic AI Keyshop for Business Leaders
Neil leads an Agentic AI keyshop to 40 business leaders across diverse industries.

RD Summit 2025 Summary
Neil delivers the keynote at the largest marketing and sales event in LATAM.

Saúde Business Forum 2025 Summary
Neil delivers the opening keynote at the leading C-suite gathering in the Brazilian Healthcare industry, with many video testimonials.

Rio Innovation Week 2025 summary
Neil delivers the closing keynote as well as additional panels at Rio Innovation Week 2025, one of the largest innovation gatherings in the world.

AI Agents Keyshop for the C-Suite
Neil leads an Agentic AI keyshop (highly interactive, actionable keynote) for 20 C-levels across leading Brazilian and global companies.

Universo TOTVS 2025 summary
Summary of Neil's participation at the annual gathering of TOTVS, the technology company that powers 25% of Brazilian GDP.

SXSW 2025 summary
Highlights of Neil's SXSW 2025 activity
Recent events

ITForum CIO Summit 2026
Neil addresses the Top CIOs in Brazil

SPIW 2026 highlights
Neil at São Paulo Innovation Week 2026

SXSW 2026 summary
Neil's extensive SXSW 2026 participation.

Agentic AI Keyshop for Business Leaders
Neil leads an Agentic AI keyshop to 40 business leaders across diverse industries.

RD Summit 2025 Summary
Neil delivers the keynote at the largest marketing and sales event in LATAM.

Saúde Business Forum 2025 Summary
Neil delivers the opening keynote at the leading C-suite gathering in the Brazilian Healthcare industry, with many video testimonials.

Rio Innovation Week 2025 summary
Neil delivers the closing keynote as well as additional panels at Rio Innovation Week 2025, one of the largest innovation gatherings in the world.

AI Agents Keyshop for the C-Suite
Neil leads an Agentic AI keyshop (highly interactive, actionable keynote) for 20 C-levels across leading Brazilian and global companies.

Universo TOTVS 2025 summary
Summary of Neil's participation at the annual gathering of TOTVS, the technology company that powers 25% of Brazilian GDP.

SXSW 2025 summary
Highlights of Neil's SXSW 2025 activity
Recent events

ITForum CIO Summit 2026
Neil addresses the Top CIOs in Brazil

SPIW 2026 highlights
Neil at São Paulo Innovation Week 2026

SXSW 2026 summary
Neil's extensive SXSW 2026 participation.

Agentic AI Keyshop for Business Leaders
Neil leads an Agentic AI keyshop to 40 business leaders across diverse industries.

RD Summit 2025 Summary
Neil delivers the keynote at the largest marketing and sales event in LATAM.

Saúde Business Forum 2025 Summary
Neil delivers the opening keynote at the leading C-suite gathering in the Brazilian Healthcare industry, with many video testimonials.

Rio Innovation Week 2025 summary
Neil delivers the closing keynote as well as additional panels at Rio Innovation Week 2025, one of the largest innovation gatherings in the world.

AI Agents Keyshop for the C-Suite
Neil leads an Agentic AI keyshop (highly interactive, actionable keynote) for 20 C-levels across leading Brazilian and global companies.

Universo TOTVS 2025 summary
Summary of Neil's participation at the annual gathering of TOTVS, the technology company that powers 25% of Brazilian GDP.

SXSW 2025 summary
Highlights of Neil's SXSW 2025 activity