A proposal from The Training Bureau · for Kroll Iberia · July 2026
This proposal is only a temporary waypoint in the journey we hope to make together.
Kroll Iberia today
Where the journey starts. Kroll Iberia as it works now; dedicated managers, skilled analysts, world-class output. But inefficiencies and opportunities to improve abound.
Kroll + ITB Beta where we are now
ITB wraps the recruiting, training and process workflows Kroll Iberia already runs. A data-free collaboration — no sensitive Subject or client data enters our systems. We work entirely alongside you.
Kroll + ITB the full collaboration
When we are greenlighted to handle sensitive Subject and client data, ITB wraps the desk itself. Every click, every subject saved, every workflow becomes part of one system — value driven by a repeatable, intelligent setup that adapts to every case. Every piece of work finished to the highest standard, by definition.
Nothing replaced.
Everything remembered.
ITB — The Training Bureau — will wrap the recruiting, training and process that Kroll Iberia already runs with a recording layer.
Every test, session and report feeds one employee profile.
The profile helps managers determine ability, and lets each analyst/intern become self-correcting.
Every section below opens as the workflow runs today. Flip each one to see it with ITB, phase by phase. Or press T.
Five phases
Kroll Iberia recruits, onboards, trains, works and reviews. The workflow works. This proposal does not replace a single step of it.
One leak
Each phase produces evidence of how an analyst thinks and writes. Today that evidence is used once, then discarded.
One fix
ITB wraps each phase as it stands and compounds the evidence into a profile that reviews reports before the manager does.
The “today” views below come from what we have seen and heard of the Madrid workflow. Correct us where we're wrong — we'd welcome it.
The idea
One profile, compounding.
Run today, every part of the workflow produces evidence — test results, exercises, corrections — and every part discards it. The manager's memory is the database.
Run with ITB, every part of the workflow feeds one employee profile. By the time a report lands on a desk, its reviewer already knows the analyst — because the analyst has been building their own profile since the day they applied.
The phases below show what that looks like, step by step.
Phase I · Recruitment
Who are you actually hiring?
Two gates decide the hire: a written test and an interview. Both are read by a manager. Neither is protected, and neither measures how the candidate investigates.
The friction
- The take-home test has no anti-cheat and no AI detection — in 2026, an unprotected test no longer proves authorship.
- It measures summarising ability and English output. Investigative capacity is never tested.
- The manager reads raw output. Nothing beyond the finished text can be measured or compared.
The same two gates, wrapped. A benchmark measures how the candidate investigates, the written test keeps its job and gains protection — and the manager receives a profile, not raw output.
What changes
- A benchmark test measures what the written test can't. A timed, open-web investigation exercise — find, verify, write up — scored against a rubric, and over time against Kroll's own analyst pool.
- The Kroll written test, ITB-wrapped. Same evaluation of summarising and English output, now completed in a monitored environment that records the process (timing, drafts, paste events), so authorship is evidenced, not guessed. No black-box “AI detector.”
- ITB triages both into a candidate profile and presents it to the manager. The interview starts from evidence, not from scratch.
The candidate profile becomes page one of the employee profile.
Phase II · The first week
The best training window is week one.
Today it opens with waiting.
Between offer and start date, the hire waits. Week one is compliance courses. Real training starts whenever a desk frees up.
The friction
- The first week is when a hire absorbs the most — and it is spent on compliance alone.
- Candidates routinely ask for preparatory tasks before day one. There is nothing to hand them.
- The days between acceptance and start date are the only training window that costs zero desk time — and they are spent on nothing.
ITB's automated training is ready the day the offer is signed. Strictly optional before day one, then running through week one alongside compliance.
What changes
- Optional early start. Hires who ask for prep finally get it, from acceptance day. Never required — employment begins on the start date.
- A narrative foundation course (2–3 h) — boolean search, visual identification methods, working techniques — then daily scored exercises through week one.
- Tradecraft foundations: signal vs. noise, investigative common sense, facts vs. truth.
- Initial storytelling and writing training, before the first real case lands.
Every exercise is scored. Performance data reaches the profile from day one — or earlier.
Phase III · Ongoing training
From watching to doing.
Training is a senior analyst presenting to a new hire who watches. The content is good. The format can't tell whether any of it landed.
The friction
- Significant senior time invested, with uncertain impact on the hire.
- Sessions are visual at best — never participatory.
- There is no way to evaluate what the hire understood.
The same sessions — Kroll databases, storytelling, techniques, case types — run inside ITB courses. The hire participates instead of observing, and every answer is captured.
What changes
- Kroll's trainers keep the content — they select, modify or create exercises from the ITB course library.
- Training becomes interactive and social. Hires do the work, seniors give feedback on the spot.
- In-built courses capture participation data automatically. Understanding is measured, not assumed.
ITB triages every session into the employee profile the manager can consult.
Phase IV · Working Gators Desk
A case is a thousand clicks.
The daily craft is hand-built. Every objective breaks into manual tasks — cycle through a normal day, chart by chart.
typed click · mouse click — each box below the analyst is one click of hand-work, case after case
The friction
- Every micro-task costs minutes, dozens of times a day — and a case is made of hundreds of them.
- The craft lives in each analyst's head. Two analysts, two different searches, two different charts.
- What the work reveals — who is fast at what, which workflows pay — is lost when the case closes.
The same objectives, run through Gators Desk. Each fan of tasks collapses into one click that generates the queries.
one mouse click clears the whole row — the clicks travel up a notch
What changes
- Nothing is saved in the program. No Subject or client data is ever stored. Results are processed and collected in-Kroll.
- LLMs touch search strings and other non-sensitive inputs, nothing else. Most bundles are deterministic.
- The value is in the clicks: faster work, fewer misses, the same craft with the friction removed.
- ITB reads the behaviour, not the case — what works well and what doesn't, who works well at what — and designs the next click-saver from it.
Every bundled click teaches the profile how this analyst works best.
Phase V · Report submission
The review loop, closed.
Every draft climbs to the manager, who works as a senior analyst: spot the mistakes, send it back, read it again. The feedback lives in exactly one place — the manager's memory.
The friction
- The manager spends senior hours, round after round, perfecting each report until it is fit to submit.
- Feedback is retained mentally — or in notes, if someone is diligent. So the same mistakes come back.
- The loop repeats identically for every analyst. Nothing learned from one review transfers to the next.
The profile built across phases I–IV now does the senior-analyst pass. The analyst writes in MS Copilot, guided by Gator — and the manager receives a report that has already been reviewed once, against this analyst's own known habits.
What changes
- The senior-analyst pass, automated. Gator Senior Analyst reads the profile and the case type, then loads a checklist for this analyst, this report, into the MS Copilot Kroll already runs. No new tool on the desktop. No report text leaves Kroll's environment.
- The analyst drafts and checks the report with Copilot at hand, guided by Gator's checklist. The review becomes their own responsibility.
- The manager still signs off every report — and their feedback flows back into the profile, so the next checklist already knows about it.
- Recurring mistakes stop recurring. Once corrected, a pattern lives in the profile and in every next checklist.
The profile is now doing the work it was collecting for — and the manager's feedback keeps it sharp.
In one look
The same workflow, side by side.
The profile can't be backfilled. It will never fully know the analysts who start before it does.
Built for scrutiny
The profile is performance data about employees, so it is designed for exactly that conversation. Kroll's data, held in Kroll's environment, visible to the analyst it describes, scoped to training and review — shaped to clear GDPR and works-council review before the pilot starts.
Nothing new on the desktop
The review checklist arrives inside the MS Copilot Kroll already runs, and report text stays where it lives today. Analysts open ITB for training, and nothing else changes.
A pilot that can say no
Fixed fee, ends on a date. If the profile hasn't earned its place, you keep the benchmark results — the first hard numbers on what review actually costs — and we leave quietly.
The ask
Start with one wrap.
A four-week pilot wraps a single phase of your next real intake. We recommend the written test — nothing to install, nothing to migrate, and the profile gets its first page. The wrap proves itself in a month. The compounding shows up with every hire after.
Pick the week →Darío García Giner · The Training Bureau · dario@thetrainingbureau.com