Workflow card WC-14.1: Score your current stack and make a keep, cut, or try list
From AI for Real Estate Agents (2026 Edition). Facts in, draft out, licensee signs.

Time: estimate: about 30 minutes

Trigger. You pay for more than one AI-related tool and are not sure which earn their keep, or a new pitch has landed and you want to decide with evidence.

Inputs.

- Every AI-related subscription from the last twelve months of card statements: name, category (Figure 14.1), monthly cost, billing cycle, renewal date
- Your current time log (T-01.1 baseline and the latest week), with minutes per task
- Your filled scorecards (T-14.1), one per tool, scored by you, not by the assistant
- Each tool's data policy and export terms, one line each
- Any new tool you are considering, with its pricing page copied as text

Brief. (paste everything between the lines)

------------------------------------------------------------
Role: You are helping a licensed residential real estate agent review the software tools the agent pays for. You only use facts I give you. If a fact is missing or unclear, ask instead of guessing. Do not add prices, features, or claims about any product from general knowledge; use only what I paste.
Facts: [paste the subscription list with costs and renewal dates] [paste the time log, minutes per task, before and current] [paste the ten scores for each tool from T-14.1] [paste the one-line data policy and export notes per tool] [paste the pricing page text for any tool under consideration]
Audience: Me. This is a working document for my own decision, not for clients.
Voice: Plain. Short. No sales language. Tables where they help.
Rules: Use only facts I give you; ask instead of guessing. Sort every tool into Keep, Cut, or Try. A tool with a score of 1 on question 4 (data policy), 6 (off switch), or 7 (compliance fit) goes to Cut regardless of its total. A tool goes to Keep only if the time log shows a task it touches got faster. Put at most one tool in Try, and say what the thirty-day test is. Do not recommend any product I did not list. No language about who should live anywhere or who the neighbors are. Where any tool produces first-contact material, note that the output must carry "[license disclosure line]".
Format: (1) A table: tool, category, monthly cost, annual cost, total score, lowest-scoring question, verdict. (2) For each Cut, one sentence saying why and the renewal date to beat. (3) For each Keep, the task in the time log that justifies it and the minutes saved per week. (4) For the one Try, the thirty-day test: the task, the minutes it costs today, and the number it must beat. (5) The annual total before and after cuts. After the output, list every fact you used and which input it came from.
------------------------------------------------------------

Check.

- Read the fact list at the end. Any price, feature, or claim not in your inputs gets cut.
- Confirm every Cut for a 1 on question 4, 6, or 7 matches your own sheet.
- Open the time log yourself for every Keep. If the minutes saved are not there, move it to Cut.
- Check each renewal date against your card statement; the assistant cannot see the statement.
- Make sure the Try column has one entry or none.

Ship. Cancel the cuts this week. Put the Try tool's thirty-day scorecard date on your calendar. Copy the annual totals to T-14.2. Save the brief with the tool list removed to your brief library as "Tool stack review, v1," and run it again at the quarterly review (Chapter 15).

Compliance check. Ch. 4.5 (privacy and client data), 4.6 (calls, texts, and bots), 4.2 and 4.3 (altered images). Paste no client names, transaction details, or never-paste data into this brief; the tool list and your scores are all it needs.
