The platform

Viki is AI. It is not an assistant.

A chatbot answers from what it was trained on. Viki answers from what you gave it — your documents, your configuration, your data — and shows you which file each answer came from.

The reading is where the weeks go

An assessment has always been the same shape: gather everything, read it, form a view, defend the view, write it up. The reading is the expensive part, and it is the part a person does worst — selectively, from memory, under deadline.

Viki reads all of it. It indexes what it finds, scores each dimension of the assessment against it, and holds a citation for every claim, so any figure in the final report traces back to its source. The consultants then do what people are actually good at: judgement, challenge, and the conversation with your board.

A live Viki engagement dashboard

It reads the whole estate.

Not a sample. Not an interview. However big the estate, coverage is total.

A brain that already knows the work.

It knows the method before it meets youThe delivery method, the accelerators and the deliverable templates already live in the brain. It does not start from a blank page.
It reads your evidence in fullDocuments, configuration, exports and data — read end to end, not sampled, and at a scale no review team works through by hand.
Every claim is tied to its sourceNothing is asserted without the file behind it, so any figure can be traced months later in front of whoever is asking.
It stops where the evidence stopsWhat cannot be proven is marked, not guessed — the difference between an assessment and a confident opinion.

What the platform does — not claims about what it did for someone else.

Four modules. One brain.

VISTA AEGIS ASTRA AURORA V ONE CORPUS STRATEGIC TACTICAL OPERATIONAL

Each module brings its own dimensions, accelerators and deliverables. The engine underneath — retrieval, scoring, provenance, the fabrication gate — is shared.

Which stack

The method does not assume a stack.

VISTA, AEGIS, ASTRA and AURORA are assessment methods, not tools bolted to one vendor. The dimensions, the scoring, the provenance discipline and the deliverable standard are the same whatever you run.

What is stack-specific is the corpus behind them — and today ours is deepest, by a distance, on SAP. We would rather tell you that here than have you discover it during your engagement.

  • 01ASTRA is stack-independent

    It assesses your operation, your data and your organisation — not your ERP. There is nothing to port.

  • 02AEGIS is method-neutral

    The seven lenses apply to any large platform programme. Our deepest evidence base behind them is SAP.

  • 03VISTA is deliberately SAP

    It reads the estate itself, object by object. That is only possible with real depth in one platform, and ours is SAP.

  • 04Another stack?

    Ask. If we cannot yet do it to the standard on this site, we will say so rather than learn on your engagement.

And the whole engagement runs here.

Not a folder of documents — a live workspace where evidence, analysis, tasks, risks and deliverables sit against each other.

Work breakdown, phases and accelerators inside a Viki engagement

Work breakdown, phases and status — live. Eighty accelerators, any of which Viki drafts.

The distinction that matters

Analysis first, then writing.

Viki forms and stores a scored analytical position before any document is generated. Deliverables are compiled from that stored position — so two deliverables from one engagement cannot contradict each other, and neither can drift from the evidence.

Most AI document tools do the opposite. They write, and the analysis is whatever the writing implied.

  • 01Evidence hub

    Every source file indexed and searchable, with what has been read and what has not.

  • 02Analysis

    Each dimension scored on a four-point scale, with confidence and the hypotheses behind it.

  • 03Deliverables

    Compiled from the stored analysis, exported to Word or PDF in your house format.

  • 04RAID and status

    Risks, assumptions, issues and dependencies tied to the findings that produced them.

Next step

See it against your own material.

The fastest way to judge this is to point it at something real. Bring a document set and we will show you what it reads — and what it refuses to answer.