Short answer: a financial advisory firm should use a website chatbot for exactly two things, explaining how the firm works and screening who should get a discovery call. It answers fee model, minimums, services, credentials, custodian and process questions, then captures enough to know whether this person fits. It gives no financial advice of any kind, ever, and it should say that plainly rather than dancing around it.
Why advisory intake is a screening problem
Most advisory firms are not short of enquiries, they are short of good ones. The inbound splits roughly into people well below your minimum, people looking for a free second opinion on a decision they have already made, people who want a product rather than a relationship, and, occasionally, exactly the client you want.
The last group has a shape you can describe. Investable assets in a certain range. A life event driving the enquiry, retirement, a liquidity event, an inheritance, a divorce, a business sale. Complexity you are equipped for, equity compensation, multiple entities, cross-border. And a willingness to work on an ongoing basis rather than a one-off question.
Screening for that in a first conversation is uncomfortable to do by phone and easy to do in chat, because a widget asking about asset range is a form field, not an awkward moment.
What to capture before booking a discovery call
Keep it to what genuinely determines fit:
Reason for the enquiry. Retirement planning, a windfall, a business exit, wanting a second opinion, dissatisfaction with a current advisor. This single answer tells you more than anything else.
Approximate investable assets. A band, not a number. Under a threshold, between two thresholds, above. People answer bands.
Whether they currently work with an advisor. A switching conversation is different from a first-time one.
Timeline. Now, this quarter, exploring for later.
Service interest. Ongoing wealth management, a one-off plan, retirement income, estate coordination, business owner planning.
Location. Registration and licensing constrain who you can serve.
Contact details and preferred time.
Clerkzo's in-chat form supports custom fields, so this becomes a short structured intake rather than a chatty back-and-forth, which is what you want when the fields are identical for every prospect. See lead capture.
If someone falls below your minimum, the honest and useful outcome is a polite explanation and, where you have one, a referral path. That is a better experience than a discovery call that ends in a decline, and it protects your calendar.
The absolute boundary
This is the part that matters most in this sector, so be unambiguous.
A chatbot on an advisory website must not answer questions about what to invest in, whether to sell something, how to allocate a portfolio, whether an annuity is right for someone, when to take a pension, how to reduce a tax bill, or whether the market is a good place to be right now. It must not comment on specific securities. It must not produce anything that could be read as a recommendation.
It can say what the firm does. It can explain the difference between the firm's service tiers. It can describe the planning process, step by step. It can say the firm is fee-only or fee-based, what the fee schedule looks like, what the minimum is, which custodian holds assets, what credentials the team holds, and how often clients are reviewed. All of that is information about the firm, not advice to a person.
The correct behaviour when a visitor asks an advice question is a clean, confident redirect: that is a question for one of our advisors because the answer depends on your full picture, let me get you a time with them. Clerkzo's guardrails already keep the assistant restricted to your own content and make it refuse and hand off rather than improvise, and you can instruct it explicitly to decline anything advisory.
Say it in the greeting too. A visible line stating that the assistant provides information about the firm and does not give financial advice sets expectations before the first question.
The cost of a slow reply in advisory work
The trigger events that create advisory clients are short-lived. Somebody who has just been told their redundancy terms, just sold a business, or just inherited is in a decision window of days. They will talk to whoever responds while the question is still live. A week later they have either acted or gone quiet, and the second option is worse because they usually acted badly.
There is also a trust component. In a field where the product is judgement and attentiveness, response speed is the only evidence a prospect has before they meet you. A firm that answers a Sunday evening enquiry at nine on Sunday evening has said something about itself that no brochure can. The broader case is in how fast should you respond to leads.
The chatbot is not the response. It is the bridge that holds the enquiry, answers the process questions, screens for fit and gets a time in the diary before attention moves on.
Setting it up
Clerkzo trains by crawling your site, so your service pages, team bios, process page and FAQ carry most of the load. Add a knowledge file for what is not published: current minimum, fee bands, capacity, which states or regions you serve, your referral partners for enquiries you decline.
Then write the guardrail instructions in your own words, test them by trying to make the widget give advice, and keep testing. Ask it what to do with a pension pot. Ask it if the market is overvalued. If it engages at all, tighten the instruction and use the corrections loop to fix the specific answer permanently.
Notifications go out by email, webhook or Telegram, and leads land in an inbox with a New to Contacted to Qualified to Won or Lost pipeline so nothing sits unclaimed. Pricing is flat from $19/mo, and there is a free plan to build and test with before you publish. See pricing.
Frequently asked questions
Can it give investment advice?
No, and it must be configured never to. It answers questions about the firm, its fees, its process and its availability, and routes everything advisory to a qualified advisor.
Can it screen out prospects below our minimum?
Yes. An asset-range field is a standard qualifier and one of the highest-value things you can ask.
Will it book meetings?
It captures booking requests with preferred times and routes them to whoever owns the calendar. See appointment booking.
Is client data handled appropriately?
Clerkzo is SOC 2 Type II examined, GDPR compliant and HIPAA assessed. Even so, collect only what screening requires and keep account-level detail out of a chat window.
How quickly can it go live?
One script tag, about ten minutes, with done-for-you setup available.