ChamberOne

Onboarding a new client matter: an AI-assisted step-by-step checklist

By the ChamberOne team 6 minute read

Every malpractice insurer will tell you that the most expensive mistakes in a law firm happen at the beginning of a matter, not the end. A conflict missed before the first advice, an engagement letter that never got signed, fee terms agreed verbally but not recorded, a limitation date seeded too late — these are the errors that lead to write-offs, claims and disciplinary referrals. AI tools can accelerate several steps of the onboarding process, but they do not remove the need for a lawyer to take each gate consciously. This checklist walks through the full sequence, marking where AI earns its keep and where human sign-off is non-negotiable.

Key points

  • Run the conflict check before any substantive advice — document who ran it, what it found and who cleared it.
  • Client identification and AML checks are a broader obligation than a single electronic verification; a qualified person must review and document the full risk assessment.
  • Engagement letters must set out scope, fees and billing terms precisely — vague terms are a leading cause of write-offs.
  • Seed all key dates and assign named tasks with clear owners on the day the matter opens, not later.
  • AI can draft, extract and suggest at each step, but a lawyer must review and accept each output before the matter advances.

Why onboarding errors cost more than they appear

Onboarding failures tend to be invisible in a firm's financial reporting. Work carried out before a matter record exists never appears in write-off reports — it simply disappears. When commercial terms are left vague or only clarified after work has started, billing disputes follow and fees are conceded. A structured onboarding routine, applied consistently to every new matter, closes both the revenue and the compliance gap.

On the compliance side, skipped or superficial AML risk assessments are a recognised process failure — steps treated as optional that are in fact required. Files that lack a documented client risk assessment, or that hold an assessment that does not genuinely evaluate money-laundering risk, are not the product of complex judgment errors; they are the product of a process that was not followed.

AI tools can help you move faster through parts of the onboarding routine. They do not remove the obligation to complete it properly.

Step 1: Run the conflict check before anything else

The conflict check comes first — before you give substantive advice, before you send an engagement letter, before you do anything that could be construed as acting for the client. You need the client's full name, any trading names or aliases, the name of any company involved, and the names of all adverse parties and related entities. Collect those at the initial enquiry stage so the check can run without delay.

Check the names against your firm's full record of current and former clients, all parties named in past matters, and any related contacts. Good practice is to include fuzzy matching for name variations — a system that only catches exact matches will miss common aliases and spelling differences. Where a potential conflict surfaces, stop and decide what the firm is permitted to do before proceeding. Document who ran the check, when, what it found, and who cleared it.

AI can help by running name matching across a larger dataset than a manual search would reach, and by flagging potential matches that share identifying characteristics even where the spelling differs. What AI cannot do is decide whether a potential conflict disqualifies the firm from acting. That judgment stays with a qualified lawyer.

Do
  • Collect full names of client, company, all adverse parties and related entities at first contact
  • Run the check before giving any substantive advice or sending an engagement letter
  • Use fuzzy name matching to catch spelling variants and aliases
  • Record who ran the check, what it found and who cleared it
Don't
  • Start substantive work before the conflict check is cleared
  • Rely on exact-match only searches — they miss common variants
  • Let AI decide whether a flagged match disqualifies the firm from acting
  • Skip documentation because the check came back clear
Conflict check: what to do and what to avoid

Step 2: Verify who the client is

Client identification is a separate step from the conflict check. For an individual, you need to establish and verify identity from a reliable, independent source. For a company, you also need to identify who has authority to give instructions and, in regulated work, who the beneficial owners are.

In jurisdictions where AML regulations apply to legal work — which includes England and Wales, Australia, Singapore, India, the United States and Sri Lanka, though the precise scope and obligations differ in each — firms must identify and verify the client, understand the purpose and intended nature of the retainer, and screen against sanctions and politically exposed person lists where required. The specific obligations vary materially by jurisdiction and matter type. If your firm operates across more than one market, check the applicable regime for each. Take advice from a lawyer qualified in the relevant jurisdiction about what your specific matter requires.

Using a purpose-built tool for identity verification, rather than a general AI assistant, is advisable: the tool needs to produce a timestamped, exportable evidence file that would satisfy an inspection. AI can flag potential issues and populate a checklist, but a qualified person must review the output and sign off on the risk assessment.

Whatever verification method you use, the obligation is broader than producing a positive result from a single check. Fee-earners must consider the full verification picture, document their assessment, and record why they were satisfied that the customer due diligence requirement was met.

Step 3: Agree scope and send the engagement letter

An unsigned or poorly scoped engagement letter is one of the most common sources of billing disputes and fee write-offs. When scope is vague, work expands, billing becomes contested, and write-offs follow.

The engagement letter must set out what work is included, what is not, who is responsible for what, the fee basis, the billing frequency, any retainer or payment on account required, how the client can terminate and what happens to fees if they do, and any jurisdiction-specific disclosures required by your regulatory regime. For matters that involve client money, the letter should also describe how funds will be held.

AI can draft an engagement letter from your firm's own template, pulling in the client name, matter type, acting team and fee terms from the matter record. That is a genuine time-saving. The draft still needs a lawyer to read it in full before it goes out. Common errors in automated drafts include the wrong billing rate, missing jurisdiction-specific disclosures, and scope language copied from a previous matter that does not fit the current one. Treat the AI draft as a first pass, not a finished document.

Once the letter is ready, send it for signature and do not carry out substantive billable work until the signed copy is returned.

Step 4: Open the matter file and configure it correctly

Opening the matter record is not a purely administrative act. The decisions made at this stage — matter type, billing rate, acting team, phases and responsible fee-earner — shape every downstream report, invoice and time entry for the life of the matter.

Set the matter type first, because it determines what phases, default tasks, custom fields and document templates will apply. Assign the responsible fee-earner and the full acting team, and check that each person's billing rate is correctly set for this client. If the client has agreed a rate that differs from the standard rate card, that override needs to be recorded now, not corrected when the first invoice is disputed.

AI tools that work inside a matter management system can suggest a task list based on the matter type, with proposed owners and rough deadlines organised by phase. That is a useful starting point, particularly for less common matter types. Review and adjust the AI-suggested list: it will not know about unusual features of this particular matter, client preferences, or anything agreed in the initial call that was not recorded.

Step 5: Seed key dates and assign initial tasks

Missing a limitation period or a court-imposed deadline is among the most serious errors a firm can make. The time to seed critical dates is the day the matter opens, not the week before the deadline.

Identify every date that matters: the limitation period if applicable, any court or tribunal deadlines, any contractual time limits in documents the client has provided, filing deadlines, and internal milestones such as the first progress report to the client. Enter each one in the matter diary with an owner and a reminder set far enough in advance that there is time to act.

AI tools can extract dates from documents uploaded to a matter — a court order, a lease, a contract — and propose deadline entries. That is useful for reducing manual extraction work. It is not a substitute for a qualified lawyer confirming that each date is correct and that the right rule applies. Deadline calculation depends on jurisdiction, court, matter type and sometimes the specific method of service. Every calculated date must be checked against the applicable rules before it is treated as confirmed.

Once dates are confirmed, assign the first tasks to named team members with clear completion criteria. A task that says 'review documents' is not actionable. A task that names the document, the required output and the deadline is.

  1. 1Conflict check: run, record and clear before any substantive advice
  2. 2Client identification and KYC: verify identity, screen sanctions and PEPs, document the assessment
  3. 3Engagement letter: draft from template, review in full, send for signature
  4. 4Matter file: open with correct type, team, rate and phases
  5. 5Dates and tasks: seed all key dates, assign named tasks with deadlines
  6. 6Billing setup: confirm rate card, retainer, billing frequency and trust ledger if needed
  7. Then back to step 1.
The matter onboarding sequence

Step 6: Confirm the billing setup and any client money

Before the first time entry is recorded, confirm that the billing configuration matches the engagement letter. Check the rate card or fixed-fee arrangement, the billing frequency, the invoice approval chain, and whether any payment on account has been received and correctly allocated.

If client money is to be held, open the trust or client ledger entry now and record the receipt with the correct matter reference. The rules on client money differ across jurisdictions — your firm's accounts team and the applicable professional rules are the reference points. What is consistent across all markets is that a client money entry recorded against the wrong matter, or held without a corresponding ledger entry, creates reconciliation problems that are time-consuming to resolve and can attract regulatory attention.

Confirm that any retainer agreed in the engagement letter has been received before substantive work begins. Many write-offs trace back to work carried out in advance of a retainer that was never collected.

Where human sign-off is non-negotiable

AI tools add genuine value at several points in the onboarding sequence: running name matching across a larger dataset, drafting the engagement letter from your template, populating task lists from the matter type, and extracting dates from uploaded documents. These are real time-savings.

What AI cannot do is make the judgment calls that onboarding requires. It cannot decide whether a conflict disqualifies the firm from acting. It cannot assess whether a client's explanation of the source of funds is plausible. It cannot determine whether scope language in an engagement letter fairly describes what was agreed. It cannot confirm that a limitation date was correctly calculated given the applicable jurisdiction and service rules.

The practical discipline is to treat AI output at every stage as a draft that a lawyer reviews and accepts — not as a completed step. The speed gain comes from having a well-structured draft to review rather than starting from blank. The risk reduction comes from the lawyer actually reading that draft, correcting it, and signing off before the matter moves forward.

Sources

  1. lawfirmambition.co.uk: Client onboarding that reduces risk

How ChamberOne handles this

  • At intake, an enquiry form captures the client and party names; a name-based conflict warning runs automatically before the matter is created.
  • Each matter type sets up phases, default tasks and an engagement letter template; the engagement letter is drafted with merge fields filled from the matter, the client and the parties.
  • The engagement letter can be sent to the client for signature on their phone; the firm receives a signed PDF with a certificate page recording the typed name, time, IP address and document fingerprint.
  • AI can produce a matter brief and suggested tasks on request, each saved as a draft until a team member accepts it.
  • Every AI call is recorded in the audit log with the model used and token counts.
  • Time can be recorded with a timer in the top bar; rate cards, fixed fees and milestones are applied at matter level; client money is kept on a trust ledger with three-way reconciliation.

What it does not do

  • The conflict check is a name-based warning; the decision to act stays with the lawyer.
  • AI outputs — including matter briefs and suggested tasks — can be wrong or incomplete and need checking; they remain drafts until accepted.
  • Click-to-sign is a simple electronic signature and does not meet the formalities required for deeds, wills or other documents where those formalities apply.
  • The client-money ledger has not been assessed against any professional accounts rules and does not by itself make a firm compliant with them.

This article is general information about the law, legal practice and technology, current when it was published. It is not legal advice. Laws and guidance change: check the sources linked above and take advice on your own circumstances from a lawyer qualified where you are.

Read next