Lacerte AI: How AI Actually Fits Into a Lacerte Practice

Lacerte sits at the center of the operational stack for many CPA firms. The workflows that route through it are the workflows where AI investment shows up first on the P&L: PBC (the prepared-by-client document list) reconciliation, tax workflow routing, client-data ingestion, trial-balance reconciliation, 1040 review. A commissioned AI layer that integrates cleanly with Lacerte addresses those workflows without forcing the operator to migrate off the system of record.

The most common integration pattern is a read-and-suggest pattern. The AI layer reads structured records out of Lacerte, runs the workflow it was commissioned to run, and writes back a suggested action that a human reviewer approves inside Lacerte's native UI. The system of record stays Lacerte. The AI layer never bypasses the human-in-the-loop step for production-data writes.

For lighter-touch workflows we have shipped read-only layers that extract structured data out of Lacerte, hand it to a reasoning step, and emit a report. No writes back. The operator uses the report as input to their existing decision process. Time to ship is faster, integration risk is lower.

For heavier workflows where the audit trail is structured and the failure cost is bounded we have shipped fully bidirectional integrations that close the loop end-to-end with structured logging. These engagements take longer (six to seven weeks rather than four to five), require more diligence on the read/write permissions inside Lacerte, and ship with a runbook (the written operating instructions) for human review of edge cases.

Lacerte does not publish a public REST API (the connection one piece of software offers to another) and does not publish webhooks. Intuit ships a Lacerte Software Developer's Kit instead, and every integration a firm can legitimately run sits on top of it. The level you build at depends on what the workflow needs to read, what it needs to write, and whether the firm runs Lacerte locally or on a host.

The SDK, read path. The SDK ships an ODBC driver and a .NET library. A developer licence is requested through Intuit's developer portal, and Intuit states plainly that SDK support does not come through normal Lacerte support channels. Read access to the Lacerte database is where almost every workflow we scope begins, because reading is where the leverage is and writing is where the risk is.

The SDK, write path. Through the ODBC driver you can update fields on records that already exist. You cannot create a client that way. Intuit's own answer in its SDK forum is that client creation goes through the SDK API rather than the ODBC driver, and commercial tools that populate returns either use the SDK API or drive Lacerte's input screens directly.

The document layer. Where the SDK does not expose what a workflow needs, the honest fallback is the documents themselves: source PDFs, the firm's own checklists, and e-file acknowledgements. Slower to build against than an API would be, but it does not depend on an interface Intuit has not published, so it does not break when the SDK does not move.

One deployment note that decides schedules more often than code does. The SDK install is per workstation rather than per server, ships in 32-bit and 64-bit builds, wants the Visual C++ 2015 redistributable and administrator rights, and the Lacerte sign-in has to be re-entered when the session expires after twenty-four hours. On hosted Lacerte, whether that is Rightworks, Verito or a private cloud image, the host has to be in the room during setup.

Treating the integration as an afterthought. The AI work is the easy part. The integration is the hard part. Operators that under-invest in the integration boundary spend the entire build cycle fighting authentication, rate limits, and edge-case schema. The commission scopes the integration boundary in the first week.

Skipping the human-in-the-loop step too early. Closing the loop end-to-end on day one is a recipe for hidden errors. Every engagement starts with human review of every AI output. Only after the operator has seen the output quality hold for sixty to ninety days does the human-in-the-loop step relax to spot-check.

Underestimating the data-cleanup work. Lacerte contains data the operator has entered over years. Some of it is clean. Some of it is not. The AI layer's quality is bounded by the data it reads. Cleaning happens as part of the build, not as a prerequisite for it. If the data is unworkable we flag it in the audit call.

Building bespoke when a product would suffice. If Lacerte already has a productized AI feature that covers the workflow, the operator should evaluate it before commissioning a custom build. We will tell the operator honestly when that is the right answer.

Specific numbers are bound by NDA (a signed non-disclosure agreement) but the pattern is consistent across the engagement set: the operator runs the workflow faster, with fewer hands, and with a structured record of every AI-generated suggestion alongside the human approval.

A typical Lacerte commission scope: one or two specific workflows, read-and-suggest pattern, four-to-seven-week build cycle, fixed fee from $10K depending on integration depth and workflow complexity. The audit call identifies the workflow. The prototype demonstrates feasibility against the operator's real data inside seven to ten days. The production build ships inside the operator's own cloud tenant (a private cloud account) under NDA (a signed non-disclosure agreement).

The operator owns the Lacerte integration code, the AI prompts, the model selection, and the data pipeline at handoff. We do not retain a license, a recurring fee, or a vendor relationship that the operator depends on.

There is no public REST (a standard way for software to exchange data over the web) API and there are no webhooks. Intuit publishes a Lacerte Software Developer's Kit for third-party vendors and for firms that write their own code against the Lacerte database. The kit includes an ODBC driver and a .NET library, and the developer licence is requested through Intuit's developer portal.

The way to know whether this playbook applies to your own setup is the $499 AI-Ready Audit, which checks it against your actual stack rather than the general case.

This matters before you scope anything. An automation plan that assumes REST endpoints and event subscriptions is planning against an interface that does not exist, and that assumption usually surfaces after a firm has already committed a tax-season date.

Four workflows automate cleanly around Lacerte without touching the return itself: client document intake and classification, chasing missing items, e-file acknowledgement monitoring, and extension batching. Two automate only partially. Prep review can flag against the firm's own checklist but cannot sign off, and inbound K-1 handling extracts reliably while basis and at-risk treatment stay with the preparer.

None of the four require leaving Lacerte, which is the point. The system of record does not move.

Partly, and the distinction is the one that decides build cost. Through the SDK's ODBC driver you can update fields on records that already exist. Creating a new client is a different operation and Intuit's own SDK forum answer is that it goes through the SDK API, not ODBC.

A read-and-suggest build, where the layer proposes and a person commits, sits at the bottom of our fee band and ships in five to seven weeks. Anything writing back through the SDK API sits higher, because the testing burden against live client tax data is real and it is not optional.

The right shape of the buying motion has the operator-owner or operating partner in the room for the audit call. The constraint identification is too consequential to delegate to a department head. The implementation work that follows can and should be delegated; the decision on which constraint a commission addresses cannot.

Three questions per reference. First, what was the named constraint the commission addressed at this operator. Second, what was the measured result twelve months post-handoff, in dollars or hours. Third, does the reference operator still run the system. Vague references on any of those three are flags. ColabContent provides direct introductions to past commission operators for any prospect that asks; a fifteen-minute call to the operator is the most honest signal a prospect can get.

The fixed fee is set after the $499 AI-Ready Audit, after the integration depth is named, and after both sides have written the constraint in a sentence. Overages occur when the operator changes the scope mid-build (a different workflow, a different integration, an additional system). Either side can pause the build to renegotiate; neither side absorbs hidden overages without explicit agreement. The default is to ship the original scope and address scope expansion in a separate engagement.

The system continues to run inside the operator's cloud tenant. Models, prompts, and integration code are versioned and the operator has the source. When the underlying foundation model improves (a new release from the model vendor, a new open-weight option), the operator can swap the component without renegotiating the engagement. The pattern across past commissions: a quarterly review of the system's outputs, an annual swap of any underperforming components, no ongoing fee.

The right call is sometimes a product (when the workflow matches a product's calibration target), sometimes an internal hire (when the operator has a five-year horizon and a $5M AI runway), sometimes a Big Four engagement (when the operator is large enough that the strategy-then-build separation makes sense), sometimes no AI right now (when the operator's leading constraint is not actually addressable with AI). We tell prospects when their constraint falls into one of those buckets and route them to whichever path fits. The never-overbook rule is real; the firms that get one of those four slots are the firms where the commission is the right buying motion.

Operators who want to test the fit before ordering the $499 AI-Ready Audit can run a five-minute self-check on six questions. First, is the business established enough that a $10,000-plus system pays for itself inside a year. Second, is there a named workflow where time or money is leaking measurably. Third, has the operator tried an off-the-shelf product and either rejected it or hit a misfit ceiling. Fourth, is the operator comfortable running the system inside their own cloud tenant under NDA. Fifth, can the senior operator commit to the 20-minute call that ends the $499 AI-Ready Audit. Sixth, is the budget for a custom build from $10,000 real this quarter.

Six yes answers means the $499 AI-Ready Audit is worth ordering. Three or fewer yes answers means the right next step is probably one of the alternatives. Four or five yes answers means the call surfaces whether the missing one is addressable.

Two artifacts make the call substantially more productive. First, a one-page description of the leading constraint, written in the operator's words, naming the workflow and the rough dollar or hour leakage. Second, a list of the systems the operator uses for the workflow (the system of record, the related tools, the integration boundaries). Neither artifact has to be polished. The point is to surface the constraint quickly so the audit call's twenty minutes are spent on the findings, not exposition.

Operators who arrive at the audit call having run the sequence usually commission the build that same week. The sequence asks four questions in a specific order. First, is the leading constraint actually addressable with AI, or is it a process problem, a staffing problem, or a stack problem that AI would not solve. Second, if AI is the right intervention, is the right buying motion a custom commission, an off-the-shelf product, or an internal hire. Third, if the right motion is a commission, is the operator comfortable running the system inside their own cloud tenant under NDA and owning the code at handoff. Fourth, is the budget for a custom build from $10,000 real this quarter.

Operators who answer yes to all four book the call. Operators who answer no to any one of them either change the question (the leading constraint is different, the budget moves, the cloud posture changes) or take a different path. We do not push operators who land at a "no" on any of the four into a commission they will not be served by.

Twelve months post-handoff, three signals tell the operator whether the commission performed against the target written down after the audit. First, the dollar or hour delta on the workflow the commission addressed, measured against the pre-engagement baseline. Second, the percentage of the workflow the AI layer now handles autonomously versus the percentage that still routes to a human reviewer. Third, the number of times the operator's team has modified the build's prompts, models, or integration code on their own without ColabContent involvement. All three should be improving over time. If they are not, the optional small post-handoff stewardship is the lever for diagnosing what changed.

A commission is not the right answer for every operator. The mid-market operator with a workflow that matches a horizontal SaaS product's calibration target is better served by the product. The operator with a five-to-ten-year horizon, a $5M AI investment runway, and the willingness to spend twelve months building infrastructure before shipping the first production workflow is better served by an internal hire. The operator at $500M-plus revenue with stakeholder counts that justify a Big Four engagement is better served by that motion. We will tell the operator which of those alternatives fits if a commission does not.

The honest case for a commission is narrow on purpose. Established operators with a named workflow constraint, with stack systems that the product market does not represent well, with the budget runway for the fixed fee, with the cloud posture to run the system inside their own tenant. Operators in that narrow band are where the math works.

Most consulting houses do not publish ranked comparisons against their competitors, do not publish the boundary of what they will not build, and do not publish fixed-fee pricing bands. We publish all three because the operators we want to commission for are the operators who reward that transparency with a faster booking. The never-overbook rule means we are not optimizing for top-of-funnel volume. We are optimizing for the right four operators each quarter. Publishing the comparisons, the rankings, and the boundaries selects for those operators.

Intuit Assist is the generative layer inside Intuit Tax Advisor, and it is the closest thing to Lacerte AI that Intuit itself ships. Intuit describes it as using generative AI in Intuit Tax Advisor to suggest client-specific tax strategies from ProConnect Tax or Lacerte returns. In practice Tax Advisor pulls client data out of Lacerte without manual re-entry, shows which tax-saving strategies apply to that client, and applies them in one click with the projected saving attached. Two things to be clear about before you compare it with anything else. It is a planning and advisory layer, not workflow automation: it helps a preparer advise, it does not chase organizers, move documents or clear diagnostics, which is where most of the hours actually go. And Intuit ties AI availability to specific products and tiers, so confirm what your own Lacerte licence includes rather than assuming the marketing applies to it. Read on accountants.intuit.com, 16 September 2026.