Tax loss harvesting is one of those high-value services that separates proactive advisers from order-takers. Done well, it can save a client thousands in taxes. Done manually, it eats hours of portfolio review time that most firms can’t afford to give every client every quarter.
The typical pattern looks like this: an adviser opens a portfolio in the weeks before year-end, scans for positions sitting at a loss, checks wash-sale rules, estimates the tax benefit, drafts a recommendation, and emails the client. For a book of 80 clients, that’s a sprint that consumes December. For clients who need attention mid-year after a market correction, it often doesn’t happen at all because the firm is underwater with other work.
The opportunity cost is real. A client with a $2 million taxable portfolio and $40,000 in unrealized losses might save $10,000 in federal tax if you act before year-end. Miss the window, and that value evaporates. Multiply that across a book, and you’re looking at hundreds of thousands in client tax savings left on the table because the firm can’t scale the monitoring work.
This is exactly the kind of repetitive, rules-based analysis that AI handles well. An agent can monitor every taxable account every day, flag opportunities as they emerge, draft client-specific recommendations with projected savings, and route them to the adviser for approval. The adviser reviews, adjusts if needed, and sends. Five minutes instead of an hour per client.
What manual tax loss harvesting monitoring actually costs
Most advisory firms don’t track the hours that go into proactive tax planning. It’s bundled into “client service” or “portfolio management” and treated as overhead. But when you pull the thread, the numbers are stark.
A typical adviser with 80 households might spend 30 to 50 hours in Q4 reviewing taxable accounts for harvesting opportunities. That’s a week of billable time redirected to a task that, while valuable, doesn’t generate new revenue. For firms that charge on AUM, it’s table stakes. For firms moving toward planning fees, it’s a cost center that scales poorly.
Paraplanners often carry the load. A mid-level paraplanner earning $75,000 might spend 15 hours a week in November and December pulling reports, checking cost basis, running tax estimates, and drafting recommendations. That’s $3,000 to $4,000 in labor cost for a single month’s push, not counting the opportunity cost of advice documents that get delayed because the team is buried in year-end tax work.
The bigger cost is the clients you don’t reach. If you have 120 taxable accounts and can only review 60 before the deadline, the other 60 get a generic email reminding them to “talk to their accountant.” That’s not advice. It’s a missed chance to demonstrate value and a missed chance to save the client real money.
Firms in the $5 million to $15 million revenue band typically see $70,000 to $200,000 in annual leakage from work that should happen but doesn’t scale. Tax loss harvesting is a textbook example. The knowledge is there. The intent is there. The capacity isn’t.
How an AI agent monitors portfolios for harvesting opportunities
An agent built for tax loss harvesting does one thing continuously: it watches. Every taxable account in the firm’s custody system gets scanned daily for positions that have dropped below cost basis by a threshold the firm sets, typically 5% or $5,000, whichever is larger.
The agent pulls current market prices, compares them to the cost basis on file, checks the holding period to avoid short-term wash sales, and flags any position that meets the criteria. It doesn’t wait for quarter-end. It doesn’t wait for the adviser to remember to run a report. It just runs.
When it finds an opportunity, it calculates the projected tax benefit. If a client holds 500 shares of a tech stock purchased at $120, now trading at $95, the unrealized loss is $12,500. Assuming a 37% federal marginal rate and 3.8% net investment income tax, the potential federal savings is roughly $5,100. The agent includes that number in the alert.
It also checks wash-sale rules. If the client sold the same security in the past 30 days, or if they hold it in another account (IRA, 401(k), spouse’s account), the agent flags the conflict. This is where manual reviews break down. An adviser looking at one account in isolation might miss that the client’s spouse bought the same stock two weeks ago in a joint account. The agent sees the whole picture because it’s connected to the firm’s data.
The output is a draft recommendation: “We’ve identified an opportunity to harvest a $12,500 loss in your XYZ Tech position. This would generate approximately $5,100 in federal tax savings this year. We recommend selling the position and reinvesting in a similar but not substantially identical security to maintain your equity exposure. Please reply by Friday if you’d like us to proceed.”
The adviser reviews it, adjusts the language if needed, adds any client-specific context, and sends. Total time: five minutes. The client gets a proactive, personalized recommendation with a clear dollar benefit. The firm demonstrates value. The tax savings compounds over time.
This is what Omni Ops is built to do. The Meeting Prep Agent already pulls portfolio data and goal progress into pre-meeting briefs. Extending that same logic to tax loss monitoring is a natural fit. The agent doesn’t replace the adviser’s judgment. It removes the manual scan work and surfaces the opportunities that matter.
Building the workflow end to end
The mechanics are straightforward. The agent connects to the firm’s portfolio management system, typically via API or a nightly data feed. It pulls positions, cost basis, and market values for every taxable account. It runs the loss calculation, checks wash-sale rules, and generates a queue of flagged opportunities.
That queue goes to a dashboard the adviser or paraplanner reviews each morning. The dashboard shows the client name, the position, the unrealized loss, the projected tax savings, and any wash-sale conflicts. The adviser clicks into a flagged opportunity, reviews the details, and either approves the draft recommendation or edits it.
Once approved, the agent sends the email to the client. It can also log the recommendation in the firm’s CRM, attach the supporting calculation, and set a follow-up task if the client doesn’t respond within three days. All of that happens without the adviser touching the CRM or the email separately.
For firms that want tighter control, the agent can route the queue to a paraplanner for first-pass review before it reaches the adviser. The paraplanner checks the math, confirms the wash-sale logic, and flags any edge cases (like a client who mentioned in a recent meeting that they’re planning to gift shares to a charity). The adviser sees a clean queue of vetted opportunities, not raw data.
The workflow scales. Whether the firm has 50 taxable accounts or 500, the agent runs the same process every day. The adviser’s time stays flat. The client experience improves because recommendations arrive when the opportunity is fresh, not weeks later after a manual review cycle.
One advisory firm in our network describes the shift this way: “We used to do a year-end scramble and hope we caught the big ones. Now we get alerts in real time. A client’s position drops 10% after earnings, and we’re in their inbox the next day with a recommendation. It feels like we’re paying attention, because we are.”
What this looks like in a real advisory practice
Let’s walk through a scenario. It’s mid-October. The market has a rough week. A client’s taxable account holds a position in a healthcare stock that’s down 12% from cost basis, an unrealized loss of $18,000. The client is in the 35% federal bracket. The potential tax savings is around $7,200.
The agent flags the opportunity on Thursday morning. The adviser sees it in the dashboard during their daily review, confirms the client has no wash-sale conflicts, and approves the draft recommendation. The email goes out at 10 a.m.
The client replies by Friday afternoon: “Yes, go ahead.” The adviser places the trade, reinvests the proceeds in a similar sector ETF to maintain exposure, and logs the transaction in the CRM. Total adviser time: eight minutes across two days.
Without the agent, this opportunity might not surface until December, when the adviser does their annual tax review. By then, the stock could have recovered, the loss could be smaller, or the client could have sold the position for another reason. The $7,200 in tax savings would be gone.
Multiply that scenario across a book of 80 clients and three or four market corrections a year, and you’re looking at dozens of opportunities that never get actioned because the firm doesn’t have the capacity to monitor continuously. The agent changes the equation. It watches. It calculates. It drafts. The adviser reviews and approves. The client saves money. The firm demonstrates value.
This is the kind of workflow we map in the AI audit for financial advisory firms. We look at where your team spends time on repetitive portfolio analysis, where opportunities slip through because the manual process can’t scale, and where an agent can take over the monitoring work without replacing the adviser’s judgment.
The compliance and accuracy question
Tax rules are specific. Wash-sale periods, cost basis adjustments, and state tax implications vary by client. An agent that gets the math wrong or misses a conflict creates liability, not value.
This is why the agent doesn’t act autonomously. It flags, calculates, and drafts. The adviser or paraplanner reviews every recommendation before it goes to the client. The agent’s job is to surface the opportunity and do the math. The adviser’s job is to confirm it fits the client’s situation and approve the language.
The agent also learns from the firm’s historical decisions. If an adviser consistently skips harvesting opportunities below $3,000 because the client prefers simplicity, the agent adjusts its threshold for that client. If a paraplanner always adds a note about state tax implications for California residents, the agent includes that language in future drafts for those clients.
The compliance documentation is automatic. Every recommendation the agent generates gets logged in the CRM with the supporting calculation, the date it was sent, and the client’s response. If the firm needs to produce an audit trail six months later, it’s already there. No one is hunting through email threads or trying to reconstruct what was said.
For firms subject to SEC or state oversight, this kind of documentation is table stakes. The agent doesn’t create new compliance risk. It reduces it by making sure every proactive recommendation is logged, calculated, and traceable.
Why firms wait and what it costs them
The most common objection we hear is, “We’ll get to this once we’ve cleaned up our data.” The portfolio feeds are messy. Cost basis is missing for some accounts. The CRM doesn’t talk to the portfolio system. The firm wants to fix all of that before they think about automation.
The problem with that logic is it never happens. Data cleanup is a project that stretches across quarters. Meanwhile, the firm is still doing manual tax reviews in December, still missing mid-year opportunities, and still leaking value because the process doesn’t scale.
The better path is to start with the accounts where the data is clean. If 60% of your taxable accounts have accurate cost basis and live in a modern custody system, the agent can monitor those 60% today. The other 40% stay on the manual process until the data catches up. You get immediate value from the subset that’s ready, and you build momentum for cleaning up the rest.
The other objection is cost. Firms assume that building an agent like this requires a six-figure software project and a year of implementation. That’s the old model. With Omni, we’re talking about a 60-minute audit that maps the workflow, identifies the data sources, and scopes the agent build. Most firms go live with a first-pass agent in 30 to 60 days, not 12 months.
The cost of waiting is the tax savings your clients don’t capture and the hours your team spends on manual reviews that could be automated. For a firm doing $8 million in revenue, that’s $70,000 to $150,000 a year in leakage. Waiting another year to “get the data right” costs more than building the agent today with the data you have.
What the Omni Audit uncovers
The audit we run for advisory firms is not a sales pitch. It’s a 60-minute working session where we map one high-cost workflow, identify where an agent can take over the repetitive work, and scope what the build looks like. You walk out with three things: a process map, a draft agent spec, and a cost estimate.
For tax loss harvesting, the audit typically uncovers three bottlenecks. First, the firm is running manual reports at fixed intervals (quarterly or year-end) instead of monitoring continuously. Second, the adviser or paraplanner is doing the math by hand, copying data between systems, and drafting emails from scratch. Third, the recommendations aren’t logged consistently, so the firm has no clean record of what was proposed and when.
The agent we spec addresses all three. It monitors daily, calculates automatically, drafts client-specific recommendations, and logs everything in the CRM. The adviser’s role shifts from doing the analysis to reviewing the queue and approving the recommendations. The client experience improves because opportunities surface in real time, not weeks later.
We also map the integration points. Where does the portfolio data live? How does the agent pull cost basis? Where do the recommendations get logged? If the firm uses Salesforce, Redtail, or Wealthbox, we show how the agent writes back to those systems. If the custody platform has an API, we use it. If not, we work with nightly data exports.
The audit is not theoretical. We’re looking at your actual systems, your actual workflows, and your actual data. By the end of the hour, you know what’s possible, what it costs, and how long it takes to go live. Book a 60-min Omni Audit and we’ll map it together.
The broader pattern
Tax loss harvesting is one example of a larger pattern: proactive client service that requires continuous monitoring and can’t scale manually. The same logic applies to rebalancing alerts, required minimum distribution reminders, and beneficiary review prompts. These are all high-value touchpoints that demonstrate the firm is paying attention, but they’re labor-intensive to execute consistently.
An agent doesn’t replace the adviser’s judgment. It removes the manual scan work, surfaces the opportunities that matter, and drafts the client communication. The adviser reviews, adjusts, and approves. The client gets a proactive, personalized recommendation. The firm scales service without scaling headcount.
This is what we mean by Omni Advisory. The platform is built to handle the repetitive monitoring and drafting work that consumes paraplanner and adviser time. The Meeting Prep Agent pulls portfolio data and goal progress into pre-meeting briefs. The Advice Document Agent drafts SOAs and ROAs from meeting transcripts. The Client Onboarding Agent runs guided fact-finds and collects KYC documents.
Tax loss harvesting fits the same pattern. The agent monitors. The agent calculates. The agent drafts. The adviser reviews and approves. The client saves money. The firm demonstrates value. The process scales.
If your firm is doing $3 million to $15 million in revenue and you’re still running manual tax reviews in December, you’re leaving money on the table. Not just in tax savings for clients, but in the hours your team spends on work that could be automated. See Omni for financial advisory firms and let’s map what an agent doing this work looks like in your practice.
The firms that move first on this won’t just save time. They’ll deliver a level of proactive service that most competitors can’t match because they’re still stuck in the manual review cycle. That’s the edge. That’s what clients pay for. And that’s what AI makes possible when you build it into the workflow instead of bolting it onto the side.