Skip to content
AI Implementation

Find · Workflow Discovery

Know what's worth building before you build it.

Workflow Discovery takes one workflow apart, measures it, tests what AI can actually do on your real cases and gives you a clear recommendation: go, go with conditions, or not yet.

Scope and fee are agreed in writing before work begins.

01 / Is it a fit?

When Discovery makes sense.

Good fit

  • The workflow repeats at least weekly.
  • More than one person or system touches it.
  • You can say what better would mean: faster, fewer errors or more capacity.
  • Someone on your team can own it after launch.

Not a fit

  • You want AI to make licensed, binding or financial decisions without review.
  • The process changes every time it runs.
  • No one can own the workflow once it's live.

Not sure which workflow? Discovery can begin by choosing one with you.

02 / What you receive

What Discovery is designed to deliver.

  1. 01

    Workflow map:

    every step, who does it, which system it touches and where it waits.
  2. 02

    Dated baseline:

    the few numbers that matter for this workflow, measured from your records and samples.
  3. 03

    A treatment for every step:

    remove, software, AI assist or person, with the reasoning.
  4. 04

    Tools check:

    what your current systems already do, including AI features you may already pay for, and what's missing.
  5. 05

    Feasibility test:

    where AI is proposed, we try it on a sample of your real cases, in your environment or with sensitive details removed, and report what it handled, what it missed and what needed a person.
  6. 06

    Data and risk review:

    what data is involved, where it goes, which vendors process it and what stays with people. We flag regulatory and coverage questions; your counsel, compliance lead or insurer decides them.
  7. 07

    Options and estimate:

    configure what you have, buy a specialized product, or build, compared with cost ranges and assumptions, plus a written quote for any work FindMeNow would do.
  8. 08

    Recommendation and readout:

    go, go with conditions, or not yet, presented in a working session with your team.

These eight deliverables are the standard scope. Your written proposal confirms what's included, and any exceptions or changes, before work begins. Deliverables are written so they can be used with us, with your own team or with another provider; ownership and use terms are set in your written agreement.

03 / How it runs

How it runs.

  1. 01

    Understand:

    interviews with the people who do the work, sample cases gathered, systems walked through.
  2. 02

    Map and test:

    the step map, baseline measures, tools check and feasibility test.
  3. 03

    Decide:

    treatments, data review, options and a readout with your team.

What we need from you

  • a workflow owner who can make introductions
  • a few interviews with the people involved
  • sample cases, reviewed in your environment or with sensitive details removed
  • read-only access or screen-shared walkthroughs where needed

04 / Sample pages

What the work looks like.

03 TreatmentsExample firm

Treatment table

StepTreatment
  1. 01Inquiry lands in a shared inbox

    Software

    Why: Rules handle this reliably. No judgment is involved.

  2. 02Work out which partner it's for

    Remove

    Why: Asking the right question up front removes the step.

  3. 03Email back for missing details

    AI assist

    Why: Inquiries vary. A tailored draft saves time, and a person checks it before it goes.

  4. 04Copy details into the CRM and a spreadsheet

    Software

    Why: This is a data-connection problem, not an AI problem.

  5. 05Research the prospect

    AI assist

    Why: Summarizing public information is a strength of AI, and linked sources make it quick to check.

  6. 06Check for conflicts and fit

    Person

    Why: Conflict and fit decisions carry professional responsibility.

  7. 07Decide whether to take the meeting

    Person

    Why: A judgment about the firm's business stays with the partner.

  8. 08Schedule the meeting

    Software

    Why: Scheduling is solved by ordinary software.

  9. 09Follow up if they go quiet

    Software

    Why: The message is approved once as a template, then sent by a rule.

08 RecommendationExample firm

Recommendation

  • Go
  • Go with conditions
  • Not yet

Of 9 steps: AI assists 2, software handles 4, 1 disappears and 2 stay with people.

Who approves what

  • Runs within rules4
  • Drafted, then approved2
  • Decided by a person2
  • Remove1

Stays with people: Conflict checks and the decision to take on a client.

What we'd measure: Time from inquiry to first meeting · Inquiries that never get a reply · Staff time per inquiry

Sample pages · example data

05 / After Discovery

Your plan, your choice.

  • 01

    Build it with us

    We scope the implementation in writing, then build it, test it with your team and measure the result.

  • 02

    Build it another way

    The plan is detailed enough for your team or another provider to use.

  • 03

    Wait

    If the recommendation is "not yet," you'll know why and what would change it.

06 / Questions

Questions leaders ask

It's specific enough to measure and fix properly, and a good first Discovery shows how to approach the next ones.

Have a workflow in mind?

Tell us what happens today and what better would look like. JJ will reply directly.