ArkoliteArkolite
Our POV · POV 01 · AI strategy

Which job should AI do in your business?

Eugene Yap, Managing Partner3 min read
The short answer

Match the role of AI to the stability of the decision. Two questions decide it: how stable is the context, the information the decision depends on, and how stable are the rules, the process that turns that information into a decision? The answers point to one of four roles: Builder, Agent, Researcher or Sounding Board.

88%

of organisations use AI in at least one function

<40%

see a measurable effect on enterprise earnings

Source: McKinsey, The State of AI global survey, 2025.

The gap between adoption and value is a deployment question. The technology works, and a small group of high performers is already extracting real value. What sets them apart is the match between the tool and the task. They choose the role of AI by the shape of the decision in front of them.

The AI Fit Matrix

Find your quadrant.

Pick a decision your team makes every week. Answer the two questions.

The information it depends on
The rules for deciding
Context the information →
Stable rules · Volatile context

AI as Agent

Let AI respond adaptively within clear bounds, where the stakes are low and a stable system can take over.

People stay in the loop for high-stakes calls.

Volatile rules · Volatile context

AI as Sounding Board

Use AI to test assumptions, explore alternatives and surface blind spots, widening the range of thinking in the room.

People decide. AI expands the thinking.

Stable rules · Stable context

AI as Builder

Use AI to write the workflow: code-based processes that run the same way every time and remove variance.

People design the system and audit the output.

Volatile rules · Stable context

AI as Researcher

Let AI gather and synthesise the most current information, doing the legwork for a decision still taking shape.

People apply judgement and creativity.

Rules the decision process →

Most of today’s attention sits with Builder and Agent, where AI drives productivity. The larger returns, in innovation and risk, come from Researcher and Sounding Board.

Treat AI as a precision instrument, matched to the job in front of it.
What to do next
  1. 1

    Map it.

    List your ten largest AI use cases and place each one on the matrix.

  2. 2

    Harden the routine.

    Move stable, repeatable work from agents into code that runs the same way every time.

  3. 3

    Aim higher.

    Put AI to work as researcher and sounding board on the decisions that shape innovation and risk.

Questions leaders ask
What is the AI Fit Matrix?

A two-by-two that matches the role of AI to a decision, based on how stable its context and its rules are. It gives four roles: Builder, Agent, Researcher and Sounding Board.

When should a business use AI agents?

When the rules are well defined and the context keeps moving, the stakes are low, and a stable system can take over if needed. People stay in the loop for high-stakes calls.

Where does AI create the most value?

Productivity gains come from Builder and Agent work. The larger returns, in innovation and risk management, come from AI as a researcher and a sounding board for leadership decisions.

Sources
  1. McKinsey & Company (2025). The State of AI: global survey.
Put it to work

Strategy for the AI-first leader.

Our Advisory partners map your decisions to the right role for AI, then build the team and tools to deliver it.

Next brief
POV 02Future of work3 min

When should your team question the AI?

People follow AI answers about 80% of the time, even when they are wrong. Rewards and feedback double how often errors get caught.

Read the brief