← All posts

How I built the workflow template packs

By Jameson Daines · 2026-04-13 · 10 min read

Advisor Prep Hero ships with profession-specific workflow template packs. Each template runs you through a structured set of questions about a matter, a client, or an engagement, and produces a set of real Markdown files in your workspace. The documents you'd have written yourself if you'd had two more hours in the day.

Picking which templates made the cut was harder than building any of them. The practitioners I talked to during research had strong opinions, and most of those opinions were different. This post is the criteria I used, the templates that made it, the runners-up, and the ones that explicitly didn't make it.

The question I started with

The question wasn't "what templates should an AI tool ship." It was more specific than that: what documents do professionals in these fields actually need to produce, where AI can meaningfully help, and where existing tools fail them?

That's a different question. The first version of it produces a generic list that any productivity SaaS would ship. The second version produces a much shorter list of things that are actually hard, where the gap between what AI can do and what existing tools do is real.

The criteria

I had four hard rules.

Rule 1: It has to be a document the practitioner actually writes, not one they should write. A lot of "professional templates" include things a consultant invented to charge you for a workshop. Real attorneys and CPAs don't produce these. Cut.

Rule 2: It has to be something the AI can meaningfully improve over a blank page. If the document is mostly proprietary knowledge specific to this one client and this one situation, the AI's contribution is small. The templates that benefit most from AI help are the ones where the practitioner has the raw facts but the structure is missing.

Rule 3: It has to have a clear "done" state. Templates like "ongoing client relationship notes" don't have an end. Templates like "new matter intake memo" do. The done-state matters because it lets the workflow have a satisfying conclusion: you start the workflow, you answer the questions, you get the document, you're done.

Rule 4: It has to be something the practitioner does primarily alone. Anything that requires real-time coordination with a client, opposing counsel, or a partner is a bad fit for an AI workflow because the AI can't model the other party's responses.

The law practice templates

Listed roughly in the order they'd come up in a matter lifecycle.

1. New Matter Intake

What it produces: A matter intake memo with the client's situation, the legal issue, relevant statutes or case types, conflict check notes, and a list of open questions. Why it's the flagship: Every matter starts here. The template turns a 30-minute intake call into a structured document you can actually use as a foundation for the work that follows.

2. Client Communication Draft

What it produces: A draft letter or email to a client explaining a situation, a decision, or a next step in plain language. Why it matters: The hardest part of client communication is usually the translation, taking something legally precise and making it legible to someone who didn't go to law school. That translation is exactly what AI is good at when given clear input.

3. Research Summary

What it produces: A structured memo summarizing research on a legal question: the relevant rule, the key cases or statutes, how courts in the relevant jurisdiction have applied it, and what's unsettled. Why it matters: Research summaries get written once and then consulted repeatedly. Having them in a real file with version history is meaningfully better than a chat log.

4. Matter Status Review

What it produces: A weekly or monthly snapshot of where a matter stands: what happened, what's pending, what the client needs to know, and what the next action is. Why it matters: The status review is the single highest-payoff document for staying on top of a busy docket. The template makes it as cheap as possible to produce consistently.

5. Engagement Letter Outline

What it produces: A structured outline of the scope, fee arrangement, communication expectations, and mutual obligations for a new engagement. This is an outline for review, not a finalized legal document. Why it matters: Most engagement letter failures are scope failures. Having a structured thinking exercise before you write the letter catches the gaps.

The tax and accounting templates

6. Client Onboarding Checklist

What it produces: A customized document request list and onboarding memo for a new tax client, based on their situation (individual, business entity type, specific issues). Why it matters: The onboarding checklist is the most frequently repeated document in a tax practice. Having a workflow that generates a tailored version in five minutes instead of thirty is genuinely time-saving at scale.

7. Tax Issue Analysis

What it produces: A structured memo working through a specific tax question: the issue, the relevant code sections or regs, how the IRS has approached it, the risk level, and the recommendation. Why it matters: Tax issue memos get written, filed, and consulted again later. Having them in a real searchable workspace file is better than a chat window.

8. Client Update Letter

What it produces: A plain-language letter explaining a tax situation, a filing status, or a planning recommendation to a client. Why it matters: Same translation problem as legal client communications. The technical content is the practitioner's job. The plain-language framing is where AI actually helps.

9. Engagement Scope Summary

What it produces: A short document confirming what services are in scope for an engagement, what's out of scope, and what the client is responsible for providing. Why it matters: Scope disputes are the most common source of client friction. Writing this down clearly before the work starts prevents most of them.

The consulting and advisory templates

10. Project Kickoff Document

What it produces: A kickoff memo with the client's situation, the stated goal, the success criteria, what's in scope, what's out of scope, and the timeline. Why it matters: The kickoff document is the thing consultants wish they'd written better after every engagement that went sideways. The template forces the clarity conversation before the work starts.

11. Client Situation Analysis

What it produces: A structured writeup of a client's current state: where they are, what they're trying to change, what the obstacles are, and what you know versus what's still unclear. Why it matters: This is the document consultants are always doing in their heads and rarely writing down. Getting it on paper in a real file, at the start of an engagement, is worth every minute.

12. Recommendation Memo

What it produces: A memo presenting a recommendation to a client: the situation as you see it, the options you considered, the recommended path, the rationale, and the risks. Why it matters: Recommendation memos are high-stakes documents. Having an AI that can help structure the argument when you've already done the thinking is different from asking an AI to think for you.

13. Weekly Engagement Review

What it produces: A short end-of-week document: what got done, what didn't, what the client needs to know, and what the next focus is. Why it matters: The weekly review is the single highest-payoff document for staying on top of a consulting engagement. Same argument as the law practice version.

14. Deliverable Outline

What it produces: A structured outline for a client deliverable: the goal, the sections, the key points in each section, and the supporting evidence or analysis that needs to go in each. Why it matters: Most consultants spend too much time writing and not enough time outlining first. The template inverts that, and the quality difference shows.

15. Engagement Close-Out Summary

What it produces: A close-out memo documenting what was accomplished, what was not in scope, what the client owns going forward, and what follow-up work might be relevant. Why it matters: Close-out memos protect both parties. They're also genuinely useful for the practitioner's own future reference, when the same client returns six months later and you need to remember where things stood.

The runners-up

These templates almost made the cut. Most will end up in future pack updates if they keep coming up in user requests.

The templates that explicitly aren't here

These are the templates I deliberately did not build, because they're a bad fit for the AI workflow model.

How the workflows actually work in the app

Each template is a TypeScript file in src/modules/workflow/templates/ that defines:

  1. The interview questions. 5-15 questions the workflow asks before generating anything.
  2. The system prompts. The AI prompts that take the interview answers and produce the documents.
  3. The output schema. A list of files the workflow produces, with the path and the content type for each.
  4. The follow-up actions. What you can do after the workflow finishes, links to other templates, options to refine specific sections, or re-run with different inputs.

The runtime takes the answers, calls your AI provider directly via BYOK (Claude, GPT, or Gemini), and writes the resulting files to your workspace folder. Every action is logged in the audit log so you can trace exactly which answer produced which paragraph in the output.

The templates are intentionally simple. Most are 100-200 lines of TypeScript. Adding a new one is a few hours of work. If you have a workflow you wish was in a pack, email me and tell me what it should look like. Real user requests are how the template list grows.

What I learned about templates, briefly

Three things, in case you're building anything similar:

  1. The interview questions matter more than the prompts. A great prompt with bad input produces generic output. A clear question set with thoughtful answers produces something useful even with a mediocre prompt.
  2. Templates compete with blank pages, not other templates. The benchmark isn't "is this better than the equivalent template somewhere else." The benchmark is "is this better than the practitioner starting from scratch." That's a much more useful framing.
  3. Don't build a template until you've been asked for it multiple times. Every template I built because I thought it would be cool got cut. Every template I built because three different practitioners asked for it survived.

That's how the packs came together. If you've read this far, you probably want to see them in action.

See the workflow template packs in Advisor Prep Hero

If you're an attorney, CPA, or consultant and there's a workflow you wish was in Advisor Prep Hero, email me at [email protected]. The template list grows based on what practitioners actually ask for, not what I think they should want.