JEDDORE / OSAsk AI

03 · AI & Automation

AI where it changes a decision. Automation where it removes a handoff.

I use AI in four ways, each tied to a system that is running or a demo you can use. No slideware.

Four ways I use AI

01 · PredictReported: Jeddore's own professional record, shown with context. Not live company data.

Predictive scoring

Models that learn from won and lost history which signals actually come before revenue. In production, a logistic-regression inquiry score captures 66% of eventual wins at 2.4× lift.

02 · AnswerVerified: Public and checkable: a live product, a public page or a shipped app.

Grounded assistants

Language models that answer only from an approved source, quote it, and say so when the source doesn't cover the question. ReadyRef's rulebook assistant works this way, and so does Ask Jeddore AI.

03 · AutomateReported: Jeddore's own professional record, shown with context. Not live company data.

Workflow automation

Replacing re-keying, manual triage and reminder emails with event-driven flows: capture once, route by rule, escalate by exception.

04 · AdoptReported: Jeddore's own professional record, shown with context. Not live company data.

AI enablement & tooling

A company-wide AI program: literacy training, a sales prompt library and a pre-send standard for checking AI-assisted output. Plus an MCP server that lets AI assistants drive ReadyRef's Play Maker.

In practice

One lead, four systems

How the pieces fit together in a revenue engine.

  1. 01Capture

    A form passes complete source data into the CRM.

  2. 02Score

    A model ranks the lead by likelihood to close and explains why.

  3. 03Route

    Automation assigns it, creates the task and sets the follow-up.

  4. 04Assist

    A grounded assistant answers product and policy questions with sources.

  5. 05Learn

    Outcomes flow back into reporting and retraining.

Illustrative: A demonstration on synthetic data. Not a real system or result. A pattern, not a specific client system. See it run in the Workflow Automation Explorer.

Operating principles

How I decide what to automate

  1. 01
    Start with the decision, not the model.

    Which decision gets better, by how much, and who acts on it? If there is no answer, there is no project.

  2. 02
    Use the data you already have.

    CRM history, process logs and documents usually hold the signal. New data collection comes later.

  3. 03
    Explain or don't ship.

    A score or answer that can't say why gets ignored. Explainability is a feature.

  4. 04
    Ground language models.

    Retrieval over approved sources, citations in every answer, and a clear refusal when the source is silent.

  5. 05
    Keep a human on exceptions.

    Automate the routine path; route the unusual cases to a person with context.

  6. 06
    Measure against the old way.

    Every automation gets a before-and-after metric: time, error rate, conversion or cost.

Try it

Don't take my word for it.

Score a lead, run a workflow, or interview the portfolio.

Perspective

What brings you here?

Pick a perspective and the site reorders itself around what you care about. The facts don't change, only the emphasis. Change it any time from the top bar.

Evidence

  • Loading…