Yes: HubSpot can run prospect and annual-review pipelines, log client emails and hold meeting references for a small firm, from the free CRM upwards. The limits sit around AI and data. Breeze Assistant reads your CRM data, AI training is on by default until a Super Admin switches it off, and protected-data fields are an Enterprise feature.
So the sensible set-up is plain: keep identifiers out of HubSpot, keep the official record in your compliance system, and let Breeze draft only what a person will rewrite and sign off. For a three-adviser firm that means Starter at $60 a month at list price, two pipelines, five custom properties and a written rule for which client emails Breeze may draft.
Two pipelines an advice firm needs: prospects and annual reviews
An advice firm has two cycles that look nothing alike. Winning a client is a short run with a hard gate at the end. Keeping one is a loop that comes round every year and must not be missed. Put them in separate deal pipelines. HubSpot's pipeline help page gives free tools no custom pipelines and Starter up to 15 (shared across record types), so the second pipeline is the practical reason an advice firm moves from Free to Starter. On Free you keep prospects in the default pipeline and run reviews from a date property and a saved list, described below.
For prospects, the useful column is the gate: what must be on file before a deal may move on. HubSpot's conditional stage properties can make a property required when a deal moves into a stage, and its pipeline help page lists no plan requirement for them, so look for the Conditional logic rules column on your Pipelines tab. Where it appears, require the File reference property at the gated stages. Where it does not, the gate is a written rule the team follows and the weekly check confirms.
| Prospect stage | Gate before moving on | Next action |
|---|---|---|
| Enquiry | Source recorded | Reply and offer an introductory call |
| Introductory call | Call notes saved, no advice given | Send the fact-find |
| Fact-find sent | Sent with a return date | Task: chase on the return date |
| Fact-find returned | Stored in the firm's own system, reference noted on the deal | Book the proposal meeting |
| Proposal meeting held | File note stored, reference noted | Send the agreement |
| Agreement signed | Signed agreement filed and identity checks complete | Hand to onboarding |
| Closed won / Closed lost | Lost reason chosen if lost | Start the review cycle |
Notice that three gates say "stored in the firm's own system, reference noted". HubSpot holds the pointer, not the content. A deal that reads "File note ref 2027-114, saved to the compliance system" tells the next person where to look without putting the sensitive document into the CRM.
The review pipeline is a loop, so create one deal per client per year, named for the year ("2027 review, [client name]"). Move it through the stages below, and close it as won when the follow-up is filed.
| Review stage | Exit rule | Typical timing |
|---|---|---|
| Review due | Deal created 60 days before the review month | T minus 60 days |
| Meeting booked | Date in the calendar and the client told | T minus 45 days |
| Preparation done | Valuation and changes since last time reviewed | T minus 5 days |
| Meeting held | Notes stored in the firm's system, reference on the deal | T |
| Follow-up sent and filed | Written summary sent and filed | T plus 7 days |
| Closed won / Closed lost | Review complete, or client declined or unreachable | T plus 14 days |
A wider look at how advisers use AI around reviews, including what to prepare, is in How Advisers Use AI to Prepare for Annual Client Reviews. This tutorial stays with the CRM mechanics.
The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowProperties that make the review loop work
The review loop needs a handful of custom properties. HubSpot's Starter page lists 10 custom properties on free tools and 1,000 on Starter, so the five below fit even on Free, with a little room to spare. They are enough to build a view that tells each adviser which reviews fall due next.
- Next review date (date, on the client's company or contact record): the one date everything else hangs from.
- Service tier (dropdown): the level of service the client pays for, which decides how many reviews they get each year.
- Review owner (HubSpot user): the adviser responsible, so views can be filtered per person.
- Last review completed (date): lets you spot a client whose date has slipped.
- File reference (single-line text on the deal): the pointer to your compliance system, never the content.
Use HubSpot's default Deal type property too. Its default options are New Business and Existing Business, which you can edit, so prospect deals can be New Business and review deals Existing Business, and one report can split revenue between the two.
One default trips up renewal-style deals. HubSpot sets a new deal's Close date to the last day of the month if you leave it blank, and moves it to today's date when the deal is closed. So do not use Close date as your review date: a completed review would overwrite it. Keep the real date in Next review date, or switch off the close-date automation in your pipeline settings, which HubSpot says is possible.
On Free, where there is no second pipeline, the loop runs on the property alone. Filter the contact list on Next review date falling in the next 60 days, filter by Review owner, and save the filter as a view if your plan allows it. Each adviser opens that view at the start of the week. It is less tidy than a pipeline but it stops a review being forgotten.
A three-adviser firm, worked through
This example is illustrative. Three advisers share 240 clients, so each looks after about 80. Each client has one review a year, spread evenly, which means 20 review deals a month across the firm. The 60-day window therefore holds about 40 reviews at any time, and each adviser sees roughly 13 of them.
- Weekly view per adviser: reviews due in 60 days, about 13, sorted by Next review date.
- Admin load: if each review needs 40 minutes of admin (booking, pulling documents, filing the summary), 20 a month is 800 minutes, or 13.3 hours of firm time.
- Pipeline health check: any review deal still in Review due with fewer than 45 days to go means nobody has booked the meeting.
- Prospect side: ten new prospects a month, each taking about eight weeks to reach a decision, means roughly 20 live deals at any moment, which is a board an adviser can scan in a minute.
If the Breeze test below shows it can shave even 10 minutes off preparation, that is 200 minutes a month across 20 reviews, or 3.3 hours. That is worth having, and it is also a reminder that the saving is modest, so the compliance cost of getting it wrong deserves as much attention as the time gain.
Email logging and records: what HubSpot keeps and what it does not
HubSpot logs email on every plan, either through a connected inbox or by copying its BCC address into messages you send and forwarding received ones to a forwarding address. Both addresses are under Settings, Objects, Activities, on the Email Log and Track tab. The address is shared by everyone in the account, a logged message is capped at 40 MB including attachments, and a recipient who is BCC'd does not get a record or the email attached. A logged email attaches to the contact, the contact's primary company and the contact's five most recent open deals, so an open review deal picks up the correspondence without anyone filing it.
The decisions that matter for an advice firm are on the same tab.
- Contact email logging rules. "Log replies only" logs an incoming message only if it belongs to a thread already in the CRM. "Log all emails (from known contacts)" logs mail to and from people who already exist as contacts. A third option also creates new contacts from incoming mail. For a firm that wants a clean client history without capturing everything, replies only is the cautious starting point.
- The never-log list. Add personal and family addresses, and any domain that should never reach a client record, so private mail cannot land in the CRM.
- Redacted email bodies (beta). A Super Admin can opt in and tick "Redact email body content". HubSpot then logs the sender, recipients, subject and attachments but replaces the body with a placeholder. It applies account-wide to every logged email after the switch, it cannot be turned on per inbox, and redacted content is permanently removed. Note that attachments still log, so this does not keep client documents out of HubSpot.
What does a logged email give you? A timeline on the contact that shows who wrote what and when, which saves hunting through inboxes. What it does not obviously give you is the guarantees a regulator typically expects of a record: that it cannot be altered, that it is kept for a set period, and that you can produce it in a form that is complete. I could not confirm any of those from the pages I read, so treat the timeline as a working copy, and keep the official client file in the system your compliance lead has approved.
Ask HubSpot support or your account manager five questions, and write the answers into your own procedures: who on your plan can edit or delete a logged email or note; whether an audit trail records the change; how long deleted data is kept; how you export a complete timeline for one client; and what happens to the data if you cancel. Then check the answers against your retention rules before the first client is added. Can AI Write File Notes From Adviser Meetings? Draft, Check, Sign Off covers the drafting and sign-off side of meeting records.
Where Breeze stops: the settings, the data and the review rule
HubSpot's Breeze Assistant page says the assistant pulls from your CRM data, customer records and HubSpot's knowledge base, drafts emails and proposals, helps prepare for and follow up on meetings, and can analyse a pipeline. It says the assistant is included with every subscription at no additional cost. For an advice firm the question is not what it can do but what it can read, and four settings answer that.
A Super Admin finds them under Settings, Account management, AI, on the Access tab. The table gives each default as the Knowledge Base lists it, with the position I would start from in an advice firm. Your compliance lead may reasonably choose differently.
| Setting | Default | Where I would start | Why |
|---|---|---|---|
| Give users access to generative AI tools and features | On | Off until your policy is approved, then on | The master switch; Breeze Assistant needs it on |
| Data access: CRM data | On | On only if the CRM holds no identifiers | Covers CRM records and default properties |
| Data access: customer conversation data | On | Off until the logging policy is set | Covers call transcripts and emails, which is where client detail lives |
| Data access: files data | Off | Leave off | Keeps stored files out of the assistant's reach |
| Give users access to Breeze Assistant | On | On after approval, with a written rule on who uses it and for what | Its own switch, separate from the master one |
| AI Model Training | On | Off, unless compliance decides otherwise | Controls whether HubSpot may use your account data to train its AI models; you can still use AI features |
The data-access toggles and the training setting are account-level, so they apply to the whole firm, and changes to the training setting are written to the audit log. Access to Breeze Assistant itself is narrower to control. HubSpot's AI set-up guide says you can remove it for specific users or teams, or make it available only to admins, through user permissions, and it recommends limiting agents to specific users because they can change processes and use credits. That is what makes the one-adviser pilot below possible: one person gets the assistant while the rest of the firm waits.
Know where the assistant stops by design, too. HubSpot's pipeline help page says Breeze Assistant cannot create or update pipeline rules, tags or automations, and cannot change contact or company lifecycle stages. That is reassuring for a regulated firm, because the structure of your pipelines stays in human hands, and it means any request to "move this client to the next stage" is a job for a person.
The second limit is the data feature itself. HubSpot's Sensitive Data feature, which stores protected information in specially permissioned properties, appears only on Enterprise subscriptions; the Knowledge Base lists no Free, Starter or Professional option. It says you can store only certain types of sensitive data and points to separate terms. HubSpot also states that accounts with Sensitive Data turned on are opted out of AI model training and cannot opt back in, that Breeze Assistant works with those accounts but with restrictions, and that people should avoid sharing sensitive information in the inputs they give Breeze, or avoid using Breeze if they do not want it to process their Sensitive Data.
The practical rule for a Free or Starter firm follows. Keep account numbers, national ID numbers, health details and detailed financial figures out of properties, notes and logged email bodies altogether, and keep a reference to the record in your own system. A hypothetical mistake shows why. An adviser pastes a client's pension account number into a free-text note for convenience. With the CRM data toggle on, it is now data the assistant may be asked to draw on, and it has to be found and removed if you later change your policy. The cheap fix is not to paste it in the first place.
HubSpot's AI trust page says third-party AI providers process your data only to deliver a feature, temporarily, under contract terms that bar them from training their models on HubSpot customer data, and that admins control what data the AI can reach. It does not address restrictions for regulated use, so those answers have to come from the settings above and from your own policy. If you are comparing this with an adviser-specific tool, AI CRM for Financial Advisers: What Is Built In and What Is Bolted On sets out the differences.
A drafting rule for AI-written client emails
Settings decide what Breeze can read. A drafting rule decides what it is allowed to write for you. Sort every email an adviser might ask it for into three bands, write the bands into your procedures, and have the team use them for a month before anyone argues about edge cases. An AI draft that goes out is still your firm's communication, so the same approval rules apply to it as to anything written by hand.
| Band | Examples | Rule |
|---|---|---|
| Green | Meeting logistics, thank-you notes, chasing a signed form, a reminder to send a statement | Draft with Breeze, read it, send |
| Amber | Review agendas, a summary of what was discussed, a description of how the firm works | Draft with Breeze, adviser rewrites, a second person checks against the file note |
| Red | Recommendations, reasons for a recommendation, projections, performance statements, anything with account or identity numbers | Adviser writes it, using the firm's own template and approval route |
Here is a before-and-after. The first draft is an illustration of the sort of text an assistant produces when asked for a follow-up after a review, not captured Breeze output.
Hi [first name], following our review I recommend increasing your monthly pension contribution, which should improve your long-term outcome and keep you on track for your goals.
It is polite and it is a problem. It states a recommendation that may not be in the file note or the written advice, it makes an outcome claim, and it commits the firm to a position in an email that nobody approved. Here is what the adviser sends instead.
Hi [first name], thank you for your time yesterday. As agreed, I will send our written summary and recommendations by Friday. In the meantime, please send your latest pension statement so it is ready for your file. [your name]
The fix moves the email from red to green: it confirms logistics, promises only what the firm will do, and leaves the advice to the document that carries the firm's sign-off. A good habit is to put the band in the request itself: "Draft a green-band email confirming Thursday's meeting and asking for the latest statement. Do not mention products, performance or recommendations."
The same thinking applies to a broader question some advisers ask first, whether to use a general assistant at all. Can Financial Advisers Use ChatGPT? What Compliance Allows covers that, and AI Agents for Financial Advisers: What They Can Do Without Advising covers the agent side, where the same admin-not-advice line applies.
One week of testing, scored in a log
Pick one adviser as the pilot, set the AI toggles the way your compliance lead agrees, and use fictional or low-risk records for the first few days. Time five ordinary tasks with and without Breeze, and note what you had to change. This log is an illustration of how the entries might look.
| Task | Minutes without | Minutes with (including checking) | Edits and flags |
|---|---|---|---|
| Review preparation summary | 20 | 12 | Fixed two dates; nothing invented |
| Meeting follow-up email (green) | 10 | 6 | Removed a product mention; rewritten |
| Reply to a prospect enquiry | 8 | 5 | Tone softened |
| Review agenda (amber) | 12 | 9 | Second person checked against the file note |
| Which reviews are due next month with no meeting booked? | 6 | No usable answer | Use the saved view instead |
| Total, first four tasks | 50 | 32 | 18 minutes saved, 36% |
The decision rule is the same as for any tool: keep it if it saves at least three minutes on most of the tasks and nothing wrong survives the check. In the illustration the saving is real and modest, and the fifth task shows the sort of thing to expect: a question the pipeline could answer with a filter is better answered by the filter. Run the test with the adviser who is most sceptical as well as the one who is keenest. Their edits tell you more than the minutes do.
What a three-adviser firm pays, and when to move up
Free allows two users, so three advisers cannot stay on it. Starter at the $20 list price is $60 a month, or $720 a year, and includes 500 HubSpot Credits a month; unused credits do not roll over. HubSpot's pricing page was showing a discount on Starter for new customers when I checked, so the first year may cost less, but budget on the list price.
Professional is a different price level. HubSpot's Sales Hub and Service Hub Professional seats were last listed at $90 a month billed annually or $100 billed monthly, plus a $1,500 onboarding fee; confirm that on the pricing page, because Professional prices change. For three advisers that is $3,240 for the first year on annual billing or $3,600 on monthly billing, plus the $1,500 fee: $4,740 or $5,100. What it buys, according to HubSpot's pricing and help pages, is fully customisable workflows, up to 300 on Sales Hub Professional against Starter's 50 with restricted triggers and actions, including workflows that create tasks automatically. It does not buy Sensitive Data, which is listed for Enterprise only. For a firm whose pipeline is simple and whose rules live in compliance procedures, Starter plus discipline is usually enough, and the Professional question can wait until the weekly review shows what you are doing by hand that a workflow could do.
If your main need is compliance-grade meeting notes, archiving and records in one place, an adviser-specific CRM may fit better than a general one, and the comparison in the further reads below is worth an hour before you commit. HubSpot does the pipeline and the prompts well. It is less clear that it should be where the official client record lives.
Questions advisers ask about HubSpot and AI
Can I store client account numbers in HubSpot?
HubSpot's Sensitive Data feature, built for protected information, is listed only on Enterprise subscriptions, and HubSpot's terms limit which sensitive data types may be stored. On Free or Starter, keep account numbers, national ID numbers and health details out of properties and notes, and store a reference to your own system instead.
Does Breeze Assistant use my client data to train AI?
HubSpot's AI settings include an AI Model Training toggle, which the Knowledge Base lists as on by default, and a Super Admin can switch it off. HubSpot's AI trust page also says its third-party AI providers are contractually barred from training their models on HubSpot customer data. Decide your position with your compliance lead, then set the toggle deliberately instead of inheriting the default.
Can an adviser let Breeze write client emails?
Let it draft routine messages such as meeting logistics, and rewrite anything before it is sent. Anything resembling a recommendation, a projection or a reason for advice should be written by the adviser and follow your firm's own approval process. An AI draft that goes out is still your firm's communication.
Is HubSpot's email log a compliant record?
Treat it as a working history, not an archive, unless your compliance lead has approved it as one. Ask HubSpot support who can edit or delete a logged email on your plan, whether there is an audit trail, how long deleted data is kept, and how you export the full timeline for a file request.
Further reads
- Wealthbox AI Notetaker Cost: Is It Cheaper Than Jump or Zocks? — An adviser-specific CRM's built-in notetaker, priced against standalone tools.
- Jump AI Set-Up for a Three-Adviser Firm: Templates, Consent, Review — Meeting-note AI set up with consent and templates.
- AI Automation for Financial Advisers: Onboarding Without Re-Keying — Onboarding without re-keying client details.
- Zapier for Financial Advisers: Five Automations for Onboarding — Five automations that connect your tools without code.
- AI Agents for Financial Advisers: What to Delegate Under Compliance — What to delegate to AI agents under compliance rules.
- ChatGPT Prompts for Financial Advisers: Follow-Ups and Review Prep — Prompts for follow-ups and review prep, with the checks.
- AI Receptionist for Financial Advisers: Cost, Scripts and Limits — Cost from published plans, a never-say rule sheet with replacement wording, and identity-check and escalation rules for an advice firm's phone.
- HubSpot for Accounting Firms: Lead Tracking and Onboarding Pipelines — A lead pipeline, an onboarding pipeline and an engagement letter task in HubSpot, plus a five-task test of whether Breeze earns its place.
- Ongoing Advice Review Evidence File: A Checklist for UK Adviser Firms — A grouped checklist for proving ongoing reviews happened: meeting record, review pack and delivery log, with a filled-in sample file.
- Wealthbox AI Agent: What It Automates and What to Review — What Wealthbox's Assistant, Agents, Notetaker and AI for Reports do, and the adviser review step each output still needs.
- AI for Asset Managers: Client Reports, Meeting Notes and Records — What a boutique asset manager can safely hand to AI, what stays human, and the record of AI-assisted work to keep for compliance.
- How Can AI Help Financial Advisors Find and Keep Clients Compliantly? — A solo adviser's month with AI: research prospects, draft content and nudge reviews, with an approval checkpoint before anything leaves the building.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: HubSpot Knowledge Base pages on AI settings, AI adoption, Sensitive Data, email logging and logging rules, pipelines, tasks, default deal properties and HubSpot Credits; HubSpot AI trust page; HubSpot pricing, Starter and Sales Hub pages (checked 4 November 2026).