Client onboarding automation for professional services firms: intake to engagement without the chasing

A new client signs an engagement letter, and then the work of actually starting stalls for two weeks. Someone emails a document list. Half of it comes back incomplete. A staff member re-types the same name, address and tax ID into the practice management system, the CRM and the billing tool, because none of them share a record. The partner who should be doing billable work is instead chasing a signature. This is the onboarding gap, and at a 20 to 500 person accounting, law, consulting, agency or recruitment firm, it repeats every week.

What does onboarding actually involve at a professional services firm?

Client onboarding is everything between a prospect agreeing to work with you and that client being ready for a team to start billable work. The steps are recognizable across accounting, law, consulting, agencies and recruitment firms, even though the documents differ by practice.

Not every step applies to every engagement. A one off contract review has a shorter list than a new audit client. Map the list before building anything, so it is clear which steps are fixed and which vary by client type.

Where does client onboarding actually stall?

None of the usual delays are exotic. They repeat across firms of every size and every practice area.

• Chasing missing documents. The questionnaire asks for a prior year return or a bank statement, half of clients do not attach it, and someone has to ask again.
• Re-keying the same details. A name, address, tax ID or matter number gets typed once for the engagement letter, again for practice management, and again for billing, with three chances to enter it differently.
• Incomplete questionnaires. Questions get skipped, and nobody notices until the assigned associate opens the file expecting to start work.
• Unclear ownership. No one is assigned to notice a stalled file, so it sits until the client calls asking why nothing has happened.

None of these are technology problems on their own. They are coordination problems that a connected process removes, once someone has written down what a complete file looks like for each client type.

Which systems carry a new client through the firm?

A new client typically passes through five kinds of systems, often with no shared record between them.

Practice management platforms already automate parts of this. Karbon, for instance, lets a firm standardize onboarding with templates that trigger once a proposal is accepted, send clients reminders for missing information, and keep a shared record of every email, note and task tied to that client (Karbon). Clio Grow does the same for law firms, feeding one intake form into the engagement letter and the client record, and Canopy adds a client portal for uploading documents directly. None of that removes the need for a person to decide whether to take the client. It removes the need to type the same address three times.

The mechanics of reading and checking the documents that arrive are the same ones covered in AI document processing for invoices, contracts and forms; onboarding just applies them to engagement letters, prior returns and entity documents instead of invoices.

What should stay with your professionals, not a workflow?

Three kinds of decisions in onboarding are judgment calls, not process steps, and automating past them creates the exposure the process was meant to reduce.

1. Acceptance and conflict decisions. Whether the firm can take a client, and whether a conflict exists, is a judgment made by a partner or a designated reviewer, informed by a search that software can run but should not resolve.
2. Risk assessment. How much oversight a new client needs, and on what terms, reflects the firm's own risk appetite.
3. Identity and due diligence judgment. Deciding what level of verification a client needs is a professional call, not a form field.

Engagement letters do more work than most firms give them credit for. The AICPA describes one as the document that "establishes boundaries of the working relationship, manages expectations, and is often helpful in resolving disputes," paired with terms covering conflicts of interest, records management and how the relationship ends (AICPA). The same article notes that tax services accounted for 77 percent of the claims asserted against CPA firms under AICPA's own Professional Liability Insurance Program in 2024, a reminder of how much rides on getting scope and terms right early.

Formal identity verification rules are narrower than many firms assume. FinCEN's customer due diligence rule requires banks, credit unions, brokers or dealers in securities, mutual funds and futures commission merchants to identify and verify the beneficial owners of legal entity customers (FinCEN); it does not extend to law or accounting firms as a class. Many firms still run their own identity and entity checks as policy, particularly for new entities or higher risk engagements, which is sound practice even though it is not the same as being a regulated financial institution. Deciding what counts as sufficient verification for a given client is a professional judgment, not a field a form can tick on its own, and it stays with the appropriate professionals.

What can automation and AI actually do in onboarding?

The realistic scope covers everything that leads up to a decision, not the decision itself: one intake form instead of three, a checklist that already knows what a new corporate tax client needs versus a new individual client, and a system that reads what has been uploaded and says clearly what is missing.

Picture the same new client moving through a connected version of this. The engagement letter goes out through the firm's e-signature tool, and signing it triggers one intake form, not three, with a checklist already reflecting the service line and client type. As the client uploads a prior return, a formation document or a bank statement, the system reads it, checks the fields against the questionnaire and firm records, and marks the checklist complete or flags what is missing, in a message it drafts and a person approves before it sends: the kind of human in the loop control that keeps client facing communication accountable to someone. Once the file is complete, a conflict and acceptance summary is ready for the partner to sign off, and the client record is created once and pushed into practice management, the CRM, the document store and billing, instead of typed four times. None of the professional decisions moved. What moved was the delay between them.

Illustrative example: Consider a 60 person accounting firm taking on a new corporate tax client. A single intake form asks for entity type, prior preparer and services needed, and immediately generates the right checklist: three years of returns, current bank statements, a fixed asset schedule and an entity formation document. As documents arrive, the system reads them, matches the entity name and tax ID against the engagement letter, and flags a state where the client is registered but did not mention it. That flag, and the completed file, go to the engagement partner, who resolves it and signs off, and the client record is created once in the firm's practice management, billing and document systems. The associate assigned to the file opens it complete on day one instead of assembling it.

How do you measure whether onboarding is actually working?

Three measures cover most of what matters, and none need a new system, only a baseline recorded before anything changes.

• Time from signed engagement to ready for work. The gap between signature and a file the assigned team can actually start on.
• Touches per client. How many separate emails, calls or reminders it takes to get a client's file complete.
• Incomplete submission rate. The share of client files that arrive missing something on the first pass.

Record these for a few weeks before changing anything, the same way you would before automating any process. Which business processes to automate first covers how to choose the client type or workflow to start with.

How should a firm start?

1. Pick one client type or service line with enough volume to be worth the effort, rather than trying to standardize every engagement type at once.
2. Write down the current checklist for that client type with the people who actually do the intake work, including the exceptions they handle by hand today.
3. Record a baseline for the three measures above over a few weeks.
4. Decide, before building anything, which steps are rule based and which need a person, using a table like the one above as the starting template.
5. Connect the systems you already use rather than replacing them, starting with the practice management and document tools your team knows.

How Kastling approaches client onboarding for professional services firms

Document collection and client onboarding sit inside our AI Integration & Automation service: collecting agreed documents, extracting relevant information, flagging what is missing or inconsistent, and preparing a record for a responsible reviewer. An engagement usually starts with a free discovery call about the client types, systems and documents involved, and a separately scoped paid audit is common for this kind of work, though not required for every project.

Before anything is built, we agree which steps run automatically and which need a partner's sign off, and we connect the practice management, CRM, billing and document systems already in use rather than replacing them. Acceptance decisions and conflict judgments stay with the appropriate professionals, and we agree up front how success will be measured. This sits alongside the other professional services workflows we look at, including document collection and review routing.