How it works

How an engagement works

Seven stages, one accountable owner, and an email-first experience. The backbone is predictable; the intelligence allocated inside it is not.

Engagement stages

Step 1 of 7

Clarify the decision context

We convert a topic into a decision context: the alternatives, the owner, the deadline, and what being wrong would cost. Most requests become sharper here, and some become smaller.

If the decision cannot be stated as a choice between real alternatives, we say so before any money moves.

Select any stage directly, or move through them in order. Depth and duration depend on the decision under review.

Stage three, in detail

The constraints you approve before execution

Everything after stage three is optimised inside this envelope. Nothing material moves outside it without a written amendment you approve.

Constraints you approve up front
  • Consequence of being wrong

    What is genuinely at stake if the choice is wrong.

  • Desired thoroughness

    A quick sanity check, or a deep review of every alternative.

  • Deadline

    When the decision must be made, not when it would be nice to know.

  • Budget

    The envelope the process is optimised inside.

  • Confidentiality and providers

    What may be shared, abstracted, or excluded — and from which AI providers.

  • Independent human involvement

    Whether, and for which topics, external human experts may be commissioned.

  • Internal-team involvement

    Whether we may ask your people, who, and about what.

  • Output preferences

    The form the result should take to be usable.

IQ256 recommends the process. You approve the envelope — and any material change to scope, cost, timing, confidentiality boundaries, or effort asked of your team.

Your team

How much of your team's attention we may use

Some context exists only inside your company. You choose during onboarding whether we may ask for it, who may be asked, and about what.

Your team's attention is part of the intelligence budget. When their context is more valuable than another external opinion, IQ256 asks — within the boundaries you approved.

Email only, in the MVP: no calendars, no meetings, no task assignment inside your organisation. We do not replace your team; we use their attention where only their context can add value.

  • 01

    Minimal

    Decision owner only

    We contact the decision owner, and no one else unless it becomes genuinely necessary.

  • 02

    Targeted

    Named people, named topics

    You list who may be contacted and about which topics. Questions stay inside that list.

  • 03

    Collaborative

    Contribution welcome

    Listed people may contribute when their internal context can materially improve the review.

What you define during onboarding

  • Person, role, and email address
  • Topics they may be contacted about
  • Direct contact, or routed through the decision owner
  • Optional limits on their time or number of questions

Your team's attention is part of the intelligence budget. When their context is more valuable than another external opinion, IQ256 asks — within the boundaries you approved, by email, and in minutes rather than meetings.

Inside the stages

The backbone is stable. The allocation is not.

Stages four to six are not a fixed panel of experts answering in parallel. What happens inside them depends on what the evidence has just made important.

Examples of conditional paths inside a review
  • Product insight exposes infrastructure economics

    A new subproblem opens, and cloud architecture and FinOps expertise is justified.

  • Only internal context can answer it

    Ask one precise question of the decision owner, or of a team member you approved for that topic.

  • A planned research lane would repeat known information

    Cancel it before execution. Intelligence avoided is worth as much as intelligence added.

  • Disagreement is factual

    Verify it against evidence and close the subproblem with a resolved answer.

  • Disagreement is about risk tolerance

    Preserve both positions, with the conditions under which each holds.

  • The most useful output changes

    Redesign the deliverable — for example a short review plus one execution path per option.

Intelligence is allocated where it can add new information.

Client experience

Email in, report out

There is no portal to learn and no dashboard to check. The orchestration happens on our side of the boundary.

Illustrative client email

Fromceo@company.example

Toreview@iq256.com

SubjectAI pricing decision

“We are deciding whether to move from seat-based to usage-based pricing before Q1. Two of us disagree about what it does to enterprise renewals, and nobody here has priced a usage model before. The decision is mine and I need it settled in six weeks.”

Attachments
  • pricing-model.xlsx
  • customer-research.pdf
  • roadmap.pdf

Illustrative example. Company names and data are invented.

What comes back

  • Fit confirmed, or one question

    A short reply either confirms a review fits, or asks the single question needed to judge it.

  • NDA before sensitive disclosure

    Nothing confidential needs to move before the NDA and the sensitive-data map are agreed.

  • Scope, fixed price, boundaries, date

    Agreed in writing before any independent human expert is commissioned.

Clarifications during the review are batched, specific, and asked only when the answer would materially change the analysis. Your time cost is measured in minutes, not days.

Client-facing flow
  1. 01

    Client email: decision, deadline, sensitivity

  2. 02

    NDA and agreed scope

  3. 03

    IQ256 orchestration (Fusion AI, research, human experts)

  4. 04

    Batched clarifications, if any

  5. 05

    Report by email + closing invoice

If a live discussion is useful — usually to clarify the decision context or to walk through the report — we schedule one. Nothing else requires your calendar.

Independent input

Different experts, different questions

Each independent human expert is asked one purpose-specific question and given only the abstracted context that question requires. Two experts on the same review will not receive the same packet.

Illustrative packetPricing expertise · packet 2 of a review

Context provided

A European B2B SaaS company of roughly 100 people sells to mid-market operations teams. A new AI capability is planned. Cost behaviour per active account is described in normalised units below.

Your question

Which packaging structures remain defensible in procurement at this cost behaviour, and which single assumption would you test first?

Cost index

1.0 → 2.4

Adoption band

Weekly, 20–40%

Segment

Mid-market ops

Withheld from this packet

  • Client identity, customer names, and personal data
  • Original documents, contracts, and board material
  • The actual price list and absolute revenue figures
  • The questions and answers given to other experts

One independent human expert sees only the context and questions their contribution requires. Another expert on the same review receives a different packet.

Commercials

Payment and correction

Independent human experts are commissioned with real money before findings exist, so payment is split around that moment.

50% before expert commissioning. 50% after delivery.

One correction included

If the report contains a factual error or we made a material process mistake, we correct it. This is a correction commitment, not unlimited revisions.

No guaranteed outcome

We do not guarantee a particular conclusion, a commercial result, or that the recommendation will be comfortable. You remain the decision owner.

Edge cases

What happens when the process is tested

We isolate what the disagreement is actually about: different facts, different assumptions, or different risk tolerance. Factual disputes get verified. Risk-tolerance disputes are preserved in the report with the conditions under which each position holds.

The report shows the disagreement and what would resolve it, rather than presenting a false consensus. That is often the most decision-relevant part of the work.