Thirteen agents. Each does one job well.

Thirteen agents, and how you meet each one

Nine have a screen you open and a result you save. Four work underneath them, and are listed here so the count means something rather than being a number on a slide.

Nine you open

• Track

  DataForge

  Statistics you can defend

  Thirty-one analyses, from basic statistics and Pareto to ANOVA, regression, capability and control charts. Twenty-seven of thirty are computed by a deterministic engine; the model only writes the explanation.

• Apply

  DocuForge

  From a short brief to a finished document

  Fourteen templates, from Charter and SIPOC to Control Plan and OCAP. Answer a short brief, edit what comes back, and export it branded. Eight of the fourteen draw their own diagram.

• Apply

  ProjectSensei

  A coach for every phase

  Guidance for the DMAIC phase you are in, a review of the artifact you wrote, and a recommendation of which tool fits the question. Suggestions can be added straight to the task board.

• Optimize

  ImpactMatrix

  Put a number on it

  ROI, cost of poor quality with a benchmark gauge, and a business case. Every output can be saved into the project as a deliverable, so the number and the work stay together.

• Learn

  SigmaTutor

  Ask the course anything

  A tutor that has read the course you are on. Answers are grounded in the actual module and section, and it shows you where the answer came from.

• Practice

  PathForge

  Your next step, worked out for you

  A readiness score from 0 to 100, the single next thing to do, and a prioritised list of what to work on. It changes as you do, and holds steady when nothing has.

• Practice

  CertiAI

  Practise before it counts

  Full-length and weak-area practice exams, graded on the server, with per-module accuracy and a readiness score so you know before you sit the real one.

• Learn

  FocusBot

  Keeps you moving

  Reads your progress and sends the nudge that fits: a streak about to break, an exam worth retrying, a module you are two sections from finishing.

• Track

  InsightStream

  Your learning, on a control chart

  The statistics the platform teaches, applied to your own scores: an individuals chart of quiz results, a Pareto of weak modules, and a projected certification date.

Four that work underneath

One retrieves the course passage an answer is grounded in. One reports on a cohort for whoever runs the programme. Two are finished endpoints whose controls are not surfaced in the interface yet, and each page says so.

• Learn

  SigmaHub

  Finds the passage the answer comes from

  The retrieval layer under the course. Every slide is embedded and searchable, filtered to your belt, so when SigmaTutor answers a question it answers from the material you are actually studying rather than from memory. You never open it; you see its work as the source shown beside an answer.

• Scale

  TeachPilot

  Reports on a whole cohort

  Cohort analytics for whoever runs the programme: who is enrolled, who was active this week, who has finished, and who is at risk, plus which sections take longest and where people stall. Written for an administrator rather than a learner, which is why it is not in the learner menu.

• Learn

  ClarityAI

  Says it again in plain language

  Takes a dense concept and returns a plain explanation, a real industry example, an everyday analogy and the key terms, pitched at your belt. The endpoint is finished and tested; the control that calls it from inside a lesson is not built yet, so today it runs behind the course rather than in front of it.

• Apply

  CollabForge

  Writes the retrospective from what happened

  Reads a project’s real activity, phase history, tollgate decisions and open comments, and drafts a retrospective or a standup digest from them rather than from what anyone remembers. The endpoint is finished and tested; the control that calls it from inside a project is not built yet.