Executive FAQ

1. What is MBT?

MBT (Much Before Transformation) is a pre-transformation research workspace for HR. It is used before an organisation issues an RFP, selects a vendor or commits capital. It structures the problem, gathers dated public evidence, traces symptoms back toward possible root causes, and states whether technology, operating-model change or wider executive action is likely to be needed.

2. Can modern AI models do this on their own?

Yes. Modern AI models can follow these research rules when they are given clear instructions. MBT does not claim otherwise. What MBT adds is a fixed workflow so those rules do not depend on someone writing and re-writing the right prompt each time:

• The same eight-step sequence is applied to every study.
• The study scope is saved once and attached to every later step.
• Source origins are counted by code, so syndicated copies of one document count as one origin.
• AI responses must match a fixed structure before they are saved; non-conforming responses are rejected.
• Studies, evidence and outputs are saved in one place instead of across individual chat histories.
• Decks and dossiers are built from the saved study record, not re-written by AI at export time.

3. Why not have an analyst use several AI models across different tabs?

An analyst can do this. They can query several models in separate conversations and combine the results. The difference is that the analyst then becomes the link between every step:

• Each handoff is a manual copy-and-paste, so what is carried forward depends on the person and the prompt they write that day.
• Checks such as counting source origins or confirming the output structure have to be done by hand each time, and are easy to skip under time pressure.
• Confidential documents can be pasted into a public research tool by mistake.
• Two analysts working on the same question will usually produce different results, with no shared study record.

MBT runs the same steps automatically: each task goes to its assigned engine, workspace documents are not sent to the web research providers, code checks the structure and origin counts, and the study is saved as one dated record. It removes manual effort and variation; it does not do anything a capable analyst could never do.

4. How is MBT different from a consulting engagement?

MBT does not sell implementation services and does not name software vendors in its research output. It recommends capability categories only. It does not replace interviews, judgement or advisory work; it gives those conversations a structured, cited starting point.

5. How does MBT help protect transformation capital?

• Technology contribution ratings of None or Low are treated as valid answers, not failures.
• Problems are prioritised on Impact × Effort, both scored 1–10. Impact is business and people value; Effort is cost, complexity, dependency and change load.
• Where a root cause sits outside HR, MBT says so before money is committed.

MBT supports a decision; it does not make it, and it does not guarantee a transformation outcome.

6. Does MBT tell us what our problems are?

No. MBT never asserts what an organisation's problems are. Possible scenarios are written as relatable, non-judgemental questions — possibilities to confirm or reject. Your team decides:

• Each scenario is shown as a possibility, never as a finding about your organisation.
• Only scenarios your team confirms as real move forward into research; rejected ones are dropped.
• Problems your team types in their own words are kept word for word and analysed in full.
• Because the organisation confirms its own reality, the diagnosis is owned internally rather than imposed from outside.

7. What is the HR-alone check?

Where a root cause sits in business model, structure, governance, funding, portfolio, leadership mandate or operating model, MBT states that HR-led change alone will fail and names the wider transformation needed first.

8. How does MBT separate symptoms from causes?

Every item is placed on one level of a seven-level chain:

Root cause → Problem → Challenge → Complexity → Scenario → Symptom → Business impact

This chain is a structured way to organise inquiry, drawn from systems thinking. It is a working model, not a proven law of organisations.

9. How does MBT address workforce sustainability and governance?

MBT includes an optional Workforce Sustainability & Governance lens. It looks at the human and accountability risks of a transformation:

• Workforce sustainability: where automated workflows, AI systems or operating-model changes are proposed, it asks which teams are affected, what re-skilling may be needed, and whether monitoring or automation could affect psychological safety.
• Governance: every automated scenario must name a human decision checkpoint, and the HR-alone check flags when a workforce issue sits in board-level governance, wage structures or contractor models that HR cannot resolve alone.
• Boundaries: it covers workforce sustainability and governance only. It does not do environmental engineering or carbon accounting.

Outputs are diagnostic hypotheses to help leaders prepare. They are not statutory filings, compliance certificates or audited disclosures.

10. How are roles and data separated?

• Evidence discovery, research synthesis and reasoning run as separate steps with separate providers.
• The reasoning step does not browse the web.
• Uploaded workspace documents stay in the MBT backend and are not sent to the web research providers.
• Each organisation's data is separated by access rules in the database.
• Research and reasoning calls are made at temperature 0. This reduces variation but does not guarantee word-for-word identical output.

For details, see the Data & Security page.

11. What is the Agentic Audit check on sources?

MBT can run an automatic check of a cited public source against free public registries. Each check is recorded as an audit trail:

• OpenAlex: looks for a matching scholarly record and whether it has been retracted. No match means the source is not an indexed scholarly publication.
• SEC EDGAR and UK Companies House: where a company is in scope, looks for its official filings so workforce and restructuring statements can be compared against them.
• World Bank and ILO: returns national labour indicators for the scope countries as a reference baseline.
• Wayback Machine: looks for archived copies of the source page, which helps confirm it existed at the stated date.

Only the public source title and link, plus the company and country names in scope, are sent. Workspace documents are never sent. The verdicts follow fixed rules, not an AI call. A registry may be unavailable or not cover a company; the check then says so. A source passing these checks is not proof its claim is correct.

12. Does MBT recommend vendors?

No. Research output names capability categories only, so your architecture team evaluates vendors on its own terms.

13. How does MBT handle scope drift?

Industry, geography, context and research horizon (current, or the next 3, 5, 10 or 15 years) are locked at the start of a study. Later steps stay within that scope.

14. What happens when internal documents disagree with public sources?

The disagreement is disclosed, not hidden, in this form:

"Workspace document states X; public sources suggested Y; internal data used per policy."

15. How does MBT bridge the CHRO with CFO and CIO priorities?

Each leader often looks at a transformation through a different lens: the CHRO through workforce capability, the CFO through capital risk, and the CIO through systems architecture. MBT gives all three the same diagnostic record before contracts are signed:

• With the CFO: every claim carries a label (FACT, EVIDENCE, INFERENCE or HYPOTHESIS), so it is clear which parts of the case rest on cited evidence and which are assumptions.
• With the CIO: technology contribution is rated None or Low where an obstacle is structural, cultural or managerial, so software is not bought to solve a non-technical problem.
• Across all three: the HR-alone check states when a root cause sits in business model, funding, governance or operating model, and names the prerequisites to address before an HR initiative can succeed.

MBT does not make trade-off decisions or resolve disagreements between leaders. It means they are reviewing the same evidence, the same causal chain and the same list of prerequisites.

16. Are MBT's findings presented as fact?

No. Outputs are diagnostic hypotheses and a cited evidence record, not audited findings. Every claim carries a label:

• FACT — an established, checkable statement.
• EVIDENCE — supported by cited sources with title, publisher, date and link.
• INFERENCE — a reasoned conclusion where no direct source exists; no citation is invented.
• HYPOTHESIS — a possibility that your team needs to test.

17. How does MBT help institutionalise pre-transformation research?

When research is done through individual prompting or consulting decks, it tends to sit in personal chat histories, laptop folders and email attachments. When the person leaves, the reasoning and sources often leave with them. MBT keeps that work as a lasting organisational record:

• Each study is saved centrally with its locked scope, cited sources, origin counts, confirmed problems and causal chain.
• Every study uses the same seven-level chain, the same four claim labels and the same Impact × Effort scoring, so studies can be read side by side.
• Workspace documents stay in your organisation's private workspace and can inform later studies.
• Past studies remain available for review after the analyst who ran them has moved on.

MBT does not currently offer live co-editing or an automatic view that combines findings across many studies. Each study is a separate, dated record.

18. What does leadership receive?

A dated research dossier, presentation decks and workbooks, built from the saved study. For each confirmed problem they include:

• The causal chain and evidence: the seven-level chain from root cause to business impact, cited sources with origin counts, and a label on every claim (FACT, EVIDENCE, INFERENCE or HYPOTHESIS).
• Business and HR transformation acts: changes to process, operating model, organisation design, policy, governance, leadership, skills or culture, each with what it involves and how it works. Where a wider business transformation is needed, the dossier sets out its workstreams, sequencing with HR and executive sponsor, and states plainly when HR action alone will not be enough.
• Digital transformation acts: the technology-enabled changes that support the same outcome, described as capability categories with no vendor names, alongside a technology contribution rating and a note on what technology cannot resolve.
• Agentic AI scenarios: modelled automated workflows linked to a transformation act, each with its trigger, actions, value area and a named human decision checkpoint. They are modelled for review, not deployed.
• KPI scorecard: business transformation KPIs (did the operation change) and digital transformation KPIs (is the capability used), split into leading (1-3 months) and lagging (6-12 months), with target, baseline, formula, cadence and the act each one measures.
• Cost of inaction and a phased roadmap prioritised by impact and effort.
• Every output carries a run id, a valid-as-of date and a re-check-after date.

Where a run did not establish a baseline or KPI, the output says so rather than filling the gap. The final layout is produced by code from the saved record. AI is not used to reword the content at export, though the underlying research was produced with AI and should be reviewed by your team.