Guides

From Personal Rolodex to Shared Company Brain: A Migration Guide

A practical, step-by-step guide to moving client knowledge out of consultants’ heads, inboxes and notebooks into one shared system, before someone hands in their notice, not after.

TB Team Boilr

· July 4, 2026 · 14 min read

Abstract dark liquid-metal texture showing many streams converging into one channel, representing scattered client knowledge merging into a shared system

TL;DR

Around 90% of an organisation's working knowledge is tacit, meaning it lives in people, not in any system [3]. In a recruitment agency, that is client history, objection-handling notes and the real decision-maker map, scattered across one consultant's CRM free-text, inbox and memory. Migrating that into a shared "Company Brain" is not a one-weekend CRM import. It is a structured project: audit what exists, triage by revenue risk, design a schema that is searchable rather than just archived, migrate the history that matters, and then wire capture into daily work so the brain keeps filling itself. Get the structure wrong and you inherit the same problem in a shinier interface: 50-55% of CRM initiatives already fail to deliver their intended value, mostly from poor adoption rather than bad technology [4]. This guide is the concrete "how", from data audit to passive capture habit, using boilr's Company Brain as the worked example of what "done" looks like.

Why This Is a Migration Project, Not a Filing Exercise

Most agencies already know, in the abstract, that client knowledge sitting in one person's head is risky. Fewer treat fixing it as an actual project with a plan, a timeline and an owner. That gap is why the knowledge stays where it is.

The scale is bigger than it looks: roughly 90% of organisational knowledge is tacit, undocumented and held by individuals rather than captured in any searchable system [3]. A CRM full of client records does not change that ratio if the useful part of each record is a free-text note only the account owner understands.
Admin load hides the real cost: UK recruiters lose an average of 17.7 hours per vacancy to manual admin, including chasing and updating notes, worth roughly £17,000 per recruiter per year in lost productivity [7]. Migrating knowledge properly, once, is cheaper than every consultant re-doing that admin forever.
Tool switching is where knowledge falls through the cracks: checking several windows, importing candidate data by hand and hunting for the right contact detail are named as some of the biggest quiet time drains in recruitment workflows [7]. Every extra system a consultant has to check is another place client context can go missing.
"We have a CRM" is not the same as "it's captured": a proper migration collects notes from every corner of the old system, not just the standard export, because standard exports routinely omit attachments and scattered notes [1]. If nobody asks that question, the migration looks complete and is not.
Agencies are already prioritising this: 50% of agencies say CRM and database optimisation is a top focus for the year, and 50% are prioritising admin automation [8], which tells you this is a live, budgeted problem, not a hypothetical one.

Treat this the way you would treat any system migration: an audit phase, a structuring phase, a migration phase and a "make it stick" phase. Skipping straight to "let's just move everything into one place" is how agencies end up with a bigger, still-useless pile of notes.

The Four Things You Are Actually Migrating

"Client knowledge" is vague enough that migrations stall before they start. Narrow it to four concrete categories, each with a different capture method, and the project becomes tractable.

1. Client History

Past briefs, placements that worked, roles that got pulled and why, and how that client actually likes to be updated. Source: CRM free-text fields, email threads, the consultant's own memory (interview them, do not just export their inbox).

2. Objection and Fee Patterns

The specific pushback a client gives on fee, exclusivity or timelines, and the response that actually worked last time. Source: almost never written down. This has to be extracted through a structured interview, not a data export, since it was never in a field to begin with.

3. The Decision-Maker Map

Who actually signs off, who just fields the initial call, who was burned by a previous agency. Source: LinkedIn org structure plus the consultant's working knowledge; rarely fully present in a CRM contact record.

4. Live Pipeline and Signal Context

What is currently in motion for that account, and what triggered the last conversation (a funding round, an exec hire, a fresh job-posting run). Source: this is the one category that should not be migrated as a static snapshot, because it goes stale immediately. It needs to be captured continuously going forward, which is the difference between a migration and an ongoing system.

Ad Hoc Handover vs a Structured Migration

The difference is not whether you have a CRM. It is whether moving knowledge into it is treated as a real project with a method, or as "export, import, done."

Step	Ad Hoc Handover	Structured Migration
Starting point	Whatever the standard CRM export happens to include	A full backup plus a direct interview with the account owner, since standard exports miss notes and attachments [1]
Prioritisation	Everyone gets equal, minimal attention, so nothing gets proper attention	Top-revenue, single-owner accounts triaged and migrated first
What gets captured	Whatever fits in a free-text note field, if anyone bothers	Structured fields for history, objections, decision-makers and signal context
Ongoing capture	Relies on the consultant remembering to write things down	Wired into work already happening: verified outreach, signal tracking, call notes
Validation	Nobody checks it worked until someone actually leaves	A dry-run handover tests it before you need it for real
Result 12 months later	Same problem, now spread across more tools	A living, searchable Company Brain that keeps improving with use

The Migration Playbook: Step by Step

This is the order that works. Skipping steps is the single biggest reason migrations stall halfway through and quietly get abandoned.

Step 1: Audit What Actually Exists

List every place client knowledge currently lives: CRM free-text, personal inboxes, spreadsheets, sticky notes, someone's memory.
Pull a full backup, not a standard export: standard exports routinely drop attachments and scattered notes, so ask the vendor directly how notes transfer and whether they stay searchable [1].
Run a data audit before anything moves: separate active relationships from dead records and identify duplicate company and contact entries [2].

Step 2: Triage by Revenue Risk

Rank clients by revenue and single-owner exposure: which accounts sit with exactly one consultant and no documented backup context.
Migrate top 20% first: do not try to migrate every historical account in one pass. Depth on your highest-revenue relationships beats shallow coverage of everything.
Flag anything with an active notice period: those accounts jump the queue, since the extraction window is closing now.

Step 3: Design a Schema That Is Searchable, Not Just Stored

Define fields, not free text: separate "objection logged" from "outcome" from "decision-maker" so each is filterable later, rather than buried in one paragraph.
Map old terminology to new: "Interviewed" in one system might mean "In Process" in another; misaligned field mapping is a common source of migrations that look complete but are not [2].
Never auto-delete duplicates during migration: migrate everything first and merge duplicates afterwards inside the new system, so you do not silently lose placement history [1].

Step 4: Migrate the History (Interview, Do Not Just Export)

Sit down with the account owner: a CRM export captures what was typed. It does not capture what was never written down, which is most of the objection-handling and decision-maker knowledge.
Use a structured template, not an open conversation: ask for the same four categories (history, objections, decision-makers, live context) for every account, so the output is consistent and comparable.
Time-box it: 30-45 minutes per top account is enough for a first pass. Perfection is not the goal; a usable starting point is.

Step 5: Test With a Sandbox Before You Go Live

Migrate a small subset first: ATS and CRM vendors typically offer sandbox environments to import limited data and check search, mapping and accuracy before the full migration [2].
Check it is actually searchable: a note that migrated but cannot be found by the next consultant has not really been migrated.
Do not run both systems in parallel indefinitely: updating old and new systems simultaneously creates two incomplete databases instead of one complete one [1].

Step 6: Wire In Passive, Ongoing Capture

Stop relying on consultants to remember to log things: the accounts with the most valuable undocumented knowledge usually belong to your busiest billers, who have the least spare time to write it down.
Attach capture to work that already happens: outreach sends, signal alerts and call notes are natural capture points if the system logs them automatically instead of asking for a separate report.
Review the "make it stick" phase like any change: multi-format training and sustained support after go-live are what separate migrations that hold from ones that quietly decay back into personal notebooks [6].

Step 7: Validate With a Dry-Run Handover

Pick a real account and simulate a handover: can a colleague answer who the decision-maker is, what the last objection was and what is live in the pipeline, using only what is in the system?
Treat "can they run it independently" as the bar: transfer is not complete until the successor can operate the account without asking the original owner [5].
Schedule a 30-day check-in: gaps in a migration usually surface after go-live, not during it [5].

5 KPIs That Tell You the Migration Actually Worked

A migration is not "done" when the import finishes. Track these for the following quarter to know whether the knowledge is actually usable, not just relocated.

Metric	What It Tells You	Target
% of top-20 accounts with structured (not free-text) history	Whether the highest-revenue relationships are actually protected	100% within 90 days
Time to answer a dry-run handover question	Whether the system is searchable in practice, not just populated	Under 5 minutes, no need to ask the account owner
New records created via structured fields vs free text	Whether the schema is actually being used going forward	80%+ structured
Weekly active usage rate (logins/edits per consultant)	Adoption, the single biggest cause of migrations failing to deliver value [4]	90%+ of the desk
Objection responses logged as reusable, not just noted	Whether tacit sales knowledge is becoming shared knowledge	80%+ of common objections

How boilr's Company Brain Keeps the Migration From Reverting

A one-off migration solves today's knowledge gap. It does not stop tomorrow's from forming, unless capture becomes part of how the work already happens. boilr is built as an AI sales employee, one per consultant, that researches companies, sources candidates, tracks buying and hiring signals, and drafts outreach for the consultant to verify and send. The Company Brain sits underneath that entire workflow as the shared memory layer everything else feeds.

What Gets Captured Automatically Once the Migration Is Wired In

Winning messages: outreach that actually converted, pooled by segment, with no manual tagging required.
ICP patterns: the traits of accounts that convert for your agency specifically, learned from real outcomes rather than a spreadsheet filled in once and never updated.
Top openers and sequences: the first lines and follow-up cadences that get replies, by account type and by signal.
Objection responses: logged automatically as consultants verify and send, rather than relying on someone remembering to write it down after the call.
Signal history: boilr's Signals module continuously tracks funding, exec moves, expansion and hiring velocity per account, so the account's story keeps building itself rather than needing a manual update.
Companies and Candidates data: enrichment on client companies and sourced candidate pools stays attached to the account, not to whoever last touched it.

The design goal is explicit: zero context lost when someone leaves, because the knowledge was never sitting only in their head to begin with. A new hire inherits the agency's proven patterns from day one, instead of the migration quietly decaying back into personal notebooks within a few months.

What Still Needs a Human

Migrating the knowledge does not migrate the judgment. Keep these firmly with your consultants:

The actual trust-building conversation with a long-standing contact.
Negotiating fee structures and contract terms.
Deciding which of the migrated objection responses actually fits this specific conversation.
Reading the room in a difficult call and deciding when to push versus hold back.
The verify-and-send step on every piece of outreach the system drafts.

6 Mistakes That Turn a Migration Into a Bigger Mess

Mistake #1: Exporting Instead of Interviewing

Why it fails: a standard CRM export only captures what was typed into a field. Most objection-handling and decision-maker knowledge was never typed anywhere [1].

Fix: pair every export with a structured interview of the account owner for the accounts that matter.

Mistake #2: Migrating Everything at Once

Why it fails: trying to migrate every historical account with equal depth spreads effort so thin that nothing gets properly captured.

Fix: triage by revenue and single-owner exposure first. Depth on your top 20% beats shallow coverage of everything.

Mistake #3: Free Text Instead of Structured Fields

Why it fails: a note migrated into a bigger free-text field is still not searchable or reusable. It is the same tribal knowledge with extra steps.

Fix: define fields for history, objections, decision-makers and signal context before migrating a single record.

Mistake #4: No Ongoing Capture Plan

Why it fails: a migration that only looks backward starts decaying the day it finishes, because new client history keeps accumulating in someone's head from that point on.

Fix: wire capture into work that is already happening, such as verified outreach and signal tracking, instead of asking for a separate log.

Mistake #5: Skipping the Sandbox Test

Why it fails: discovering that notes did not transfer, or are not searchable, after go-live means redoing work under pressure [2].

Fix: migrate a small subset first and check search, mapping and accuracy before the full cutover.

Mistake #6: Treating Go-Live as the Finish Line

Why it fails: adoption, not the technology, is the main reason 50-55% of CRM initiatives fail to deliver value [4]. A perfectly migrated system nobody actually uses is functionally the same as no migration at all.

Fix: plan sustained training and a 30-day check-in as part of the project, not an afterthought [5].

A 6-Week Migration Roadmap

Week 1: Audit and Map

List every place client knowledge lives. Pull full backups, not standard exports. Run a data audit to separate active relationships from dead records and flag duplicates [2].

Week 2: Triage and Prioritise

Rank clients by revenue and single-owner risk. Identify your top 20% and any accounts with an active notice period that need to jump the queue.

Week 3: Design the Schema

Define structured fields for history, objections, decision-makers and live context. Map old terminology to new so nothing gets lost in translation between systems.

Week 4: Migrate and Interview

Run the migration for your top-tier accounts. Pair every migrated record with a 30-45 minute structured interview of the account owner to capture what an export alone would miss.

Week 5: Sandbox Test and Wire In Ongoing Capture

Test a subset for searchability and accuracy before full go-live. Connect capture to outreach and signal tracking so new knowledge starts filling the system automatically from day one.

Week 6: Dry-Run Handover and Check-In

Simulate a real handover on a top account. Measure how fast a colleague can answer the key questions without asking the owner. Schedule a 30-day follow-up to catch what surfaces after go-live.

Ready to stop migrating knowledge by hand every time someone joins or leaves? See how boilr's Company Brain captures it automatically, from day one, without a separate migration project.

Frequently Asked Questions

How long does a client knowledge migration take for a recruitment agency?

For a focused migration covering your top-tier accounts, six weeks is realistic: one week to audit, one to triage, one to design the schema, one to migrate and interview account owners, one to sandbox-test and wire in ongoing capture, and one to validate with a dry-run handover. Full enterprise CRM migrations covering every historical record can run considerably longer, but most agencies do not need to migrate everything with equal depth on day one.

Should I migrate every historical client record, or just the active ones?

Triage first. Migrate your top-revenue, single-owner accounts with full depth (structured history, objections, decision-makers, live context), and treat dead or low-activity records as lower priority. Trying to migrate everything with equal effort is a common reason migrations stall before they finish.

Why can't I just export my CRM and import it into a new system?

A standard export usually only captures what was typed into a field, and it often omits attachments and notes scattered across the old system. Most of the valuable knowledge, like objection-handling patterns and the real decision-maker map, was never typed into any field to begin with. That has to be captured through a structured interview with the account owner, not a data export.

What is the biggest reason knowledge migrations fail?

Adoption, not technology. Roughly 50-55% of CRM initiatives fail to deliver their intended value, and poor user adoption is consistently identified as the primary cause rather than the software itself. A migration that is technically complete but that consultants do not actually use has not solved the underlying problem.

What should I capture first if I can only start with one thing?

Start with your top 20% of accounts by revenue, specifically the ones managed by a single consultant with no documented backup context. That is where the business risk is concentrated, and it gives you a manageable pilot to prove the process works before rolling it out agency-wide.

How is a "Company Brain" different from finishing a CRM migration?

A migration is a one-off project that moves existing knowledge into a system. A Company Brain is what keeps that system current afterwards, by capturing new knowledge passively as work happens (outreach sends, signal alerts, call outcomes) instead of relying on someone remembering to log it. Without that ongoing layer, a migrated system starts decaying back toward personal notebooks the day after go-live.

Do we need special software to do this, or can we do it in our existing CRM?

You can start structuring fields in most existing CRMs; the discipline matters more than the specific tool at first. The limitation most agencies hit is that a standard CRM stores what is typed into it but does not learn from outcomes automatically, so the ongoing capture step still needs to be wired in deliberately, whether that is through workflow rules in your current system or a layer like boilr's Company Brain built specifically for that job.

What does a successful migration actually look like six months later?

Structured, not free-text, records for your top accounts. A new hire who can answer basic account questions without asking a colleague. Objection responses that get reused rather than relearned. And, critically, the system still getting more useful over time rather than quietly going stale, because capture is wired into daily work rather than dependent on someone remembering.

Sources

Information sourced from public industry reports and research publications as of July 2026.

Back to all articles