Before you choose a tool, make the right call.

See what the Assessment contains, then take your score in 90 seconds.

Open the deliverable before you commit.

  1. A reasoned recommendation
  2. Sources and unanswered questions
  3. Alternatives, including no AI
  4. A test, an owner and stop conditions
Get the sample PDF

An 8-page example, so you see the exact shape of the deliverable before choosing. Yours follows the same structure, built on your own files.

Start with your score.

A few questions on how your firm works, with no account. The result shows at once. Loïc reads it and calls you if a 20-minute conversation makes sense for your firm.

Get my score in 90 secondsChoose your Assessment

Every paid Assessment includes a reasoned recommendation, the alternatives and an action plan, reviewed by a consultant before you receive them.

Skip to the front page

AI transformation partner for B2B service firms

We take on one new firm a week. Next opening: week of November 16.

AI Jungle
Menu
AI Jungle

Workflows and proof

Workflows and proof Working sampleEdition of November 13, 2026 · Proof 1 · Page S1

AI assessment example for an executive search firm.

Follow an example executive search firm from an initial brief to a clear first build. See which opportunities come first, which ideas we set aside and what the partners need to decide.

Get my score in 90 seconds

Free, with no account. Loïc Guyon, the founder, reads every result and calls the firms he can help.

Example Assessment · no client system connected

Back to the assessment

The executive search workflow brief

A three-partner executive search firm receives mandates through referrals. Partners still turn every mandate into a search brief, prepare candidate conversations and chase quiet client decisions. They want to protect judgement and reduce the work that waits between meetings.

Source material for the AI assessment

  • Two recent mandate briefs, with sensitive names removed.
  • One example of a partner-edited shortlist or client update.
  • The current source of truth for contacts, notes and status.
  • The approval owner and the point where external messages must stop.

Working assumption: the firm will share representative examples only after the access boundary and retention period are agreed.

The workflow map during the assessment

This diagram shows the first proposed workflow in this example. In your interview, the agent builds and revises the map from your answers, your tools and the people responsible.

Interview snapshot

Draft · 13 November 2026

Process

From mandate to approved brief

An example scenario: tools and access permissions are confirmed during the interview.

From mandate to approved briefAn example scenario: tools and access permissions are confirmed during the interview.Agreed accessOpen questionsAfter approvalProposedRedacted mandatebriefsOwner: PartnerTool unknownProposedDraft withsource evidenceOwner: AI agentTool unknownProposedCorrection andapprovalOwner: PartnerTool unknownProposedWorking searchcontextOwner: SearchteamTool unknownLinks follow only the relationships supplied.
  1. Proposed

    Redacted mandate briefs

    Owner: Partner
    Tool unknown

    Agreed access

  2. Proposed

    Draft with source evidence

    Owner: AI agent
    Tool unknown

    Open questions

  3. Proposed

    Correction and approval

    Owner: Partner
    Tool unknown

    After approval

  4. Proposed

    Working search context

    Owner: Search team
    Tool unknown

Scroll horizontally to read the full diagram.

LegendReportedProposedUnknown

Ranked AI workflow opportunities

  1. 01

    Mandate brief to partner-ready search brief

    Why: It repeats at the start of every mandate and keeps the partner in document assembly before the search judgement begins.

    First build:
    Extract the stated need, open questions, constraints and decision criteria into a draft brief that the partner can correct.
    Guard:
    The partner approves the brief before it becomes working context. Unsupported details stay marked as questions.
  2. 02

    Quiet client decision to follow-up queue

    Why: The next action is usually clear from the last approved exchange, but it is easy to lose while a partner is in interviews.

    First build:
    Prepare a small queue of follow-up drafts tied to the last approved note and the next decision needed.
    Guard:
    Nothing sends automatically. The relationship owner edits or approves each draft and can reject the suggestion.
  3. 03

    Meeting notes to partner preparation

    Why: Preparation is knowledge-heavy and recurring, while missing evidence is often discovered too late.

    First build:
    Turn an approved meeting record into a one-page brief with confirmed facts, open questions and a suggested agenda.
    Guard:
    Every fact links back to the approved record. The partner decides what belongs in the client conversation.

AI workflow options rejected at this stage

  • Autonomous outreach. The sample has no evidence that a safe send rule or a complete relationship record exists.
  • A general firm chatbot. It would answer broadly without proving one repeatable handover.
  • Automatic proposal or shortlist generation. Those outputs carry judgement and need a named reviewer first.

Unknowns that change the AI workflow scope

  • Which contact and notes systems are authoritative?
  • How long may approved examples remain in the working environment?
  • Who is the reviewer when the relationship owner is unavailable?
  • What baseline shows that preparation time fell without lowering quality?

The scope of the first AI workflow build

Start with mandate brief to partner-ready search brief. It is narrow enough to test and close enough to the firm’s judgement to expose the real transfer work.

  • Inputs: two redacted briefs and one partner-corrected example.
  • Output: one inspectable draft with evidence, gaps and open questions.
  • Approval: the partner edits and accepts the brief before it becomes working context.
  • Out of scope: CRM writes, candidate decisions, client sends and autonomous actions.

Success measure: After a short shadow run, the partner compares preparation time and correction quality against the agreed baseline.

The decision after the workflow assessment

If the shadow run shows a repeatable handover with acceptable corrections, build this workflow first. If the inputs are too inconsistent or the approval owner is unclear, stop and resolve that before adding another agent.

The full assessment tests these assumptions with your examples and leaves a decision brief with alternatives and a next-test plan, reviewed by a human consultant. Your score comes first; the format is agreed before any payment.

$0

The free Leverage Scorecard

About 90 seconds on how your firm works today and where senior time goes. It reads no client system and commits you to nothing.

Get my score in 90 seconds →

$999

The complete AI assessment with human review

Standard includes one consultant, two 45-minute sessions and a report within 24 to 48 hours of the last session. Department adds a two-week review of each department before the second session. Onsite adds on-site demonstration and training. The format and scope are agreed on Loïc’s call, before any payment; the Standard Assessment is deducted from month one.

See the assessment levels →

Before you decide, see where your firm stands.

Ninety seconds on how your firm runs today. You see where senior time goes, and when twenty minutes on one engagement would help, Loïc calls you.

A bundle of folded newspapers, tied with string, ready to leave.

AI Jungle LLC, run by Loïc Guyon.