mcpbeat

Grad Ambidexterity

asgard-ai-platform/grad-ambidexterity

Apply organizational ambidexterity theory to balance exploration and exploitation activities. Use this skill when the user needs to diagnose whether an organization is over-exploiting or over-exploring, design structures that support both innovation and efficiency, or evaluate the tension between short-term performance and long-term renewal.

10k tokens
context cost
the whole folder, loaded on every use
4
files
instructions only
0
copies elsewhere
how many repositories repackaged it
223
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/asgard-ai-platform/skills --skill grad-ambidexterity

What comes with it

32 306 bytes besides the instruction
examples/sample_scenario.md
references/march-1991-model.md
references/structural-ambidexterity-design.md

The instruction itself

12 sections, as written by the author

Organizational Ambidexterity: Exploration vs Exploitation

Overview

Organizational ambidexterity refers to a firm's ability to simultaneously pursue exploration (innovation, experimentation, new opportunities) and exploitation (efficiency, refinement, execution of existing capabilities). March (1991) demonstrated that firms favoring one over the other face suboptimal outcomes: over-exploitation leads to competency traps, while over-exploration leads to failure traps.

When to Use

Trigger conditions:

  • User asks how to innovate without sacrificing current business performance
  • User is restructuring an organization to support both R&D and operations
  • User describes symptoms of a competency trap (good at the wrong things) or failure trap (too many experiments, no results)
  • User mentions "explore vs exploit", "innovation vs efficiency", or "ambidextrous organization"

When NOT to use:

  • For analyzing disruption from external entrants -> use grad-disruptive-innovation
  • For strategic alliances to access innovation -> use grad-coopetition
  • For internationalization decisions -> use grad-oli or grad-uppsala

Assumptions

IRON LAW: Over-Exploiting Kills Long-Term Innovation;
          Over-Exploring Kills Short-Term Revenue

Exploitation WITHOUT exploration leads to a COMPETENCY TRAP: the firm
becomes excellent at yesterday's business and is blindsided by change.

Exploration WITHOUT exploitation leads to a FAILURE TRAP: the firm
burns resources on experiments that never reach market scale.

There is no stable equilibrium — the balance must be actively managed.
  • Exploration and exploitation compete for scarce resources (attention, talent, budget)
  • The optimal balance shifts with industry dynamism and firm lifecycle stage
  • Senior leadership must actively manage the tension — it does not self-organize

Methodology

Step 1: Diagnose the Current Balance

Assess the organization's exploration-exploitation ratio:

| Indicator | Exploitation-Heavy | Balanced | Exploration-Heavy |

|-----------|-------------------|----------|-------------------|

| R&D spend (% revenue) | < 3% | 5-15% | > 20% |

| New product revenue (% total) | < 10% | 20-40% | > 50% |

| Time horizon of projects | < 1 year | Mixed | > 3 years |

| Tolerance for failure | Very low | Moderate | Very high |

| Process formalization | Rigid | Adaptive | Chaotic |

Step 2: Identify the Ambidexterity Mode

Choose the structural approach:

  • Structural ambidexterity (Tushman & O'Reilly): Separate exploration units from exploitation units with different cultures, processes, and metrics. Senior leadership integrates at the top.
  • Contextual ambidexterity (Gibson & Birkinshaw): Individual employees switch between exploration and exploitation based on context. Requires supportive culture (discipline + stretch + trust + support).
  • Sequential ambidexterity: Alternate between periods of exploration and exploitation (less common, suits smaller firms).

Step 3: Design the Integration Mechanism

For structural ambidexterity, define:

  • Separate unit boundaries (physical, cultural, reporting)
  • Integration points (shared senior team, knowledge transfer rituals)
  • Resource allocation rules (fixed exploration budget vs dynamic)

For contextual ambidexterity, define:

  • Behavioral expectations (% time on exploration vs exploitation)
  • Cultural enablers (psychological safety for experimentation)
  • Metrics that reward both (balanced scorecard approach)

Step 4: Monitor and Rebalance

Establish review cycles (quarterly pipeline health, annual market trends) to detect drift toward either trap. Define trigger conditions for rebalancing.

Output Format

# Ambidexterity Assessment: {Organization}

## Current State Diagnosis
- Balance: Exploitation-heavy / Balanced / Exploration-heavy
- Evidence: {key indicators}
- Risk: Competency trap / Failure trap / None

## Recommended Ambidexterity Mode
- Mode: Structural / Contextual / Sequential
- Rationale: {why this mode fits}

## Design Recommendations
- Exploration unit: {scope, budget, metrics, reporting}
- Exploitation unit: {scope, budget, metrics, reporting}
- Integration mechanism: {how they connect}

## Rebalancing Triggers
- {Condition 1}: shift toward more exploration
- {Condition 2}: shift toward more exploitation

Gotchas

  • Structural separation without integration is just a spin-off: If the exploration unit has no connection to the core business, you lose synergies. The senior team MUST integrate.
  • "Innovation theater" is not exploration: Hackathons and labs that never ship products waste resources. Exploration must have a path to market.
  • Context matters for mode selection: Structural ambidexterity suits large firms with resources to maintain separate units. Contextual suits smaller firms where everyone wears multiple hats.
  • The balance point shifts: A startup should be exploration-heavy. A mature firm in a stable industry can be exploitation-heavy. There is no universal ratio.
  • Metrics misalignment is the #1 killer: If exploration units are judged by exploitation metrics (quarterly revenue), they will be shut down before they can deliver.

References

  • For March (1991) formal model of adaptive systems, see references/march-1991-model.md
  • For Tushman & O'Reilly structural design templates, see references/structural-ambidexterity-design.md

How to use it

Copy the folder

Take asgard-ai-platform/grad-ambidexterity from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.