MMarketing Against The Grain
← All frameworks
ProductivityBarbara Jovanovic

Weekly AI Friction Audit

Turn recurring workflow friction into a redesigned process each week

Difficulty
Moderate
Time to result
~weeks to results
Steps
7
Confidence
97%

The Weekly AI Friction Audit treats excessive time or repeated annoyance as evidence that a process should be reconsidered. The operator selects a task that felt slow, narrates the current workflow to AI, and requests an analysis of what can be simplified, delegated, outsourced, or connected through an integration such as Zapier. AI proposes a revised workflow and can also generate the prompts required to execute it. The operator tests the change within a limited scope, preserves human control over consequential decisions, and observes whether it genuinely reduces effort without harming quality. Repeating the audit regularly replaces the older habit of following a process until it breaks with continuous, lightweight operational improvement.

Origin

Barbara Jovanovic described using AI to reassess time-consuming work every week instead of waiting until an established process fails under pressure. Extracted from Marketing Against The Grain.

Core principles

  • 01Repeated friction is a signal to redesign the workflow
  • 02Processes should be revisited before they visibly break
  • 03AI can diagnose delegation and integration opportunities
  • 04The operator must explain both the current process and desired outcome
  • 05Small frequent reviews prevent inefficient habits from becoming permanent

How to run it

  1. 1

    Spot the friction

    Review the week and identify a recurring task that consumed disproportionate time or attention.

    Pro tip Track repeated interruptions and handoffs, not only total task duration.

    Watch out Do not redesign a process solely because one unusual case was difficult.

  2. 2

    Explain the current workflow

    Narrate the actual inputs, actions, tools, decisions, and outputs involved.

    Pro tip Include unofficial workarounds because they often reveal the real bottleneck.

    Watch out An idealized process description will produce irrelevant recommendations.

  3. 3

    Request an efficiency diagnosis

    Ask AI which steps can be removed, simplified, delegated, or automated.

    Pro tip Require the model to preserve named quality and control requirements.

    Watch out Faster is not better if the redesign hides errors or weakens accountability.

  4. 4

    Explore integrations

    Ask which existing tools or connectors could remove manual transfers and repetitive actions.

    Pro tip Prefer tools already in the stack before adding new subscriptions.

    Watch out Automation can silently propagate bad data across systems.

  5. 5

    Design the revised workflow

    Have AI map the new sequence and generate any reusable prompts or handoff templates.

    Pro tip Keep explicit checkpoints for approval, source verification, and exceptions.

    Watch out Do not leave ownership ambiguous between the human and automated steps.

  6. 6

    Run a bounded test

    Apply the redesign to a limited, representative task and compare time, quality, and error rates.

    Pro tip Record the baseline so improvements can be measured.

    Watch out Do not deploy an unverified workflow across every client or project.

  7. 7

    Review and standardize

    Keep the redesign if it works, revise it if results are mixed, or revert if quality declines.

    Pro tip Schedule another review after enough real cases have accumulated.

    Watch out A process may need reassessment as models, tools, or business requirements change.

In the wild

Automating a slow weekly handoff

A solo agency owner notices that transferring meeting outputs into content tasks takes too long. She narrates the current steps to AI, asks what can be delegated or connected through Zapier, and tests the proposed workflow with one client before wider use.

The operator replaces repeated manual transfers with a tested, documented workflow.

Common mistakes

Waiting until failure

A process followed unchanged until capacity collapses creates urgent, risky redesign work.

Automating before mapping

Connecting tools without understanding decisions and exceptions can accelerate the wrong process.

Adding tools by default

The audit should deepen or connect existing tools when possible rather than create a larger, harder-to-maintain stack.

Is it for you?

Best for

Solo operators and lean teams whose workflows evolve rapidly as AI tools and business needs change.

Not ideal for

Stable, regulated processes where frequent unvalidated changes would increase operational risk.

From the transcript

It kind of makes you reassess your processes on a very regular basis because you have these like prompts that you're using.

Barbara Jovanovic · 34:00

Whereas now I feel like I can just revisit that every week.

Barbara Jovanovic · 34:30

Tell me where I can be more efficient. Tell me where what I can outsource, tell me what integration I can set up with like…

Barbara Jovanovic · 34:30

From the episode

I Run a 6-Figure Agency with ZERO Employees (AI Only)

Barbara Jovanovic