For consultantsTemplates

Compare options and get a recommendation, for consultants

Read each option, pick the factors that matter for your decision, compare them in one table, drop the factors where the options tie, and get a recommendation that also argues the runner-up's case.

This is the version for consultants. The steps and checks are the same as in the general template; each prompt carries one added line about your setting, and the examples are from your kind of work.

Sample results: runs and scores on this page are illustrations while testing is set up. How we test

  • Who it is for: Independent consultants and small-firm partners who weigh vendors, tools or work plans for a client and give one recommendation in a memo.
  • Time: 15 minutes

Download it, in the format you use

  • Prompt chain

    Copy each step into ChatGPT or Claude, one after another.

  • Checklist

    Follow the steps and check off the checks by hand.

  • Claude skill

    A folder Claude loads, so it follows the steps by itself.

  • n8n file

    Import into n8n and connect your own AI account.

Using a download:How to use this in n8nHow to install a Claude skill

How it runs

Select a part to see its prompt, input, output and check.

  2. Runs on each piece

The same job, with and without the template

Sample

One plain prompt

Which of these three options should we recommend to the client? Cost over three years and logging from a client site matter most. (The same options and must-haves are pasted below.)

I would recommend the desktop suite. It is by far the cheapest over three years, at about 6,000 dollars in total (3,900 dollars once plus 700 dollars a year for updates), compared with about 20,000 dollars for the cloud tracker and more than that for the spreadsheet option once the accountant's time is counted. Its one drawback is that staff can only log time in the office, but they can enter their hours when they get back from a site, so it still works for a team that is partly on the road. Set-up takes about 8 weeks, which should still fit before February 1. Overall it gives the best value.

Problem:It picks the desktop suite for its low price and treats logging time only in the office as good enough, though staff on client sites must be able to log time, and it assumes 8 weeks of set-up fits before February 1.

With this template

The same input, run through the steps.

Swipe sideways to see all columns

FactorCloud trackerDesktop suiteSpreadsheet plus accountantDifference
Meets the must-havesYes: staff log time from anywhereNo: staff can log time only in the officeYes: staff can open it on any devicediffers
Cost14 dollars per user a month3,900 dollars once, plus 700 dollars a yearNo license cost, plus the accountant at 60 dollars an hour, about 10 hours a monthdiffers
InvoicesMade automaticallyMade automaticallyBuilt by handdiffers
Set-upAbout 3 weeksAbout 8 weeksAbout 1 weekdiffers
What reviewers sayReports easy to read, phone app slowReliable but datedNobody has reviewed itdiffers

Dropped because the options tie: Handles all 40 staff (all three do)

Recommendation: Cloud tracker Why: It meets the site must-have: staff log time from anywhere, including a client site, and its cost fits the budget (Meets the must-haves). Whether about 3 weeks of set-up fits before February 1 depends on a start date the options do not give. Of the two options that meet that must-have, it costs less over three years, about twenty thousand dollars for 40 staff against about twenty-two thousand for the spreadsheet (Cost). It makes invoices automatically, so no accountant hours go on building them by hand (Invoices). What you give up: The phone app is slow, and set-up takes about 3 weeks against about 1 week for the spreadsheet. Runner-up: Spreadsheet plus accountant

Case for the runner-up: The spreadsheet has no license cost and the shortest set-up, about 1 week, so the firm could start logging time almost at once. Staff can open it on any device, so it also meets the site must-have. The accountant's time is the main cost, and it is a known amount at 60 dollars an hour, about 10 hours a month. What would change my mind: How long building invoices by hand really takes, because if it is clearly fewer than 10 hours a month the spreadsheet could become the cheaper option over three years, and if it is more the cloud tracker's lead grows. Also whether the slow phone app stops site staff from logging time, which would count against the cloud tracker.

  • Check:Every number comes from the options: All the prices, user limits and set-up times are copied as written.
  • Check:Each factor is marked tie or differs: Six factors compared: five marked differs and one marked tie.
  • Check:Only differing factors are left: Handles all 40 staff was removed as a tie, and the five rows left are all marked differs.
  • Check:A recommendation is named: Cloud tracker.
  • Check:The facts that could change the answer are named: Two facts to find out: how long invoices take to build by hand, and whether the slow phone app stops site staff from logging time.
  • Check:Every number in the recommendation is in the options: No total was added; the three-year costs are written in words.
  • Check:Every reason names a factor from the table: The must-haves, the cost and the invoices are all rows of the table.
  • For your read:You make the final call: Does anything you know about these tools or the client's staff change the answer?

Runs and scores

Sample

Claude, Sonnet model

October 4, 2026, typical sample

  • Check:Every number comes from the options

    All the prices, user limits and set-up times are copied as written.

  • Check:Each factor is marked tie or differs

    Six factors compared: five marked differs and one marked tie.

  • Check:Only differing factors are left

    Handles all 40 staff was removed as a tie, and the five rows left are all marked differs.

  • Check:A recommendation is named

    Cloud tracker.

  • Check:The facts that could change the answer are named

    Two facts to find out: how long invoices take to build by hand, and whether the slow phone app stops site staff from logging time.

  • Check:Every number in the recommendation is in the options

    No total was added; the three-year costs are written in words.

  • Check:Every reason names a factor from the table

    The must-haves, the cost and the invoices are all rows of the table.

  • For your read:You make the final call

    Does anything you know about these tools or the client's staff change the answer?

Fixed: Nothing needed fixing: every check passed on the first pass.

Met 8 of 8

1 of these is for your read: the tool hands it to you.

ChatGPT, GPT model

October 5, 2026, tricky sample

  • Check:Every number comes from the options

    All the prices, customer numbers, languages and report times are copied as written, and Agency B's start date is listed as not stated.

  • First pass failed:Each factor is marked tie or differs

    Two rows had maybe in the Difference column.

  • Check:Only differing factors are left

    No factor tied, so all six rows are kept and each is marked differs; the dropped line reads None.

  • Check:A recommendation is named

    Agency B, named as the closest because no option meets every must-have.

  • Check:The facts that could change the answer are named

    Two facts to find out: Agency B's start date, and whether an online-only survey can reach the customers the client needs.

  • Failed:Every number in the recommendation is in the options

    The recommendation gives a gap of 4,700 dollars between Agency A and Agency B; 4,700 is worked out from the options and not written in them.

  • Check:Every reason names a factor from the table

    The reasons are the must-haves, the price and the languages, all rows of the table.

  • For your read:You make the final call

    Can you confirm Agency B's start date, since the date for the results depends on it?

Fixed: The table step ran again once with the rule to write only tie or differs, and every row then carried one of the two words, which cleared the first failure. The recommendation step ran again once and still gave a gap in dollars worked out from the options, so the numbers check failed again and the run stopped there and flagged it for you.

Met 7 of 8

  • Not met:Every number in the recommendation is in the options: The recommendation still gives a gap of 4,700 dollars, a figure worked out from the options and not written in them.

1 of these is for your read: the tool hands it to you.

n8n, import format 2.20

October 6, 2026, easy sample

  • Check:Every number comes from the options

    All the costs, call lengths and start times are copied as written.

  • Check:Each factor is marked tie or differs

    Four factors compared: three marked differs and one marked tie.

  • Check:Only differing factors are left

    Recording was removed as a tie, and the three rows left are all marked differs.

  • Check:A recommendation is named

    Video calls.

  • Check:The facts that could change the answer are named

    Two facts to find out: whether in-person visits would give richer notes, and whether the sites could be booked sooner than 3 weeks.

  • Check:Every number in the recommendation is in the options

    No total was added; every number in the answer is in the options.

  • Check:Every reason names a factor from the table

    The must-haves, the cost and the interview length are all rows of the table.

  • For your read:You make the final call

    Does anything you know about these stakeholders, such as who needs to be seen face to face, change the answer?

Fixed: Nothing needed fixing: every check passed on the first pass.

Met 8 of 8

1 of these is for your read: the tool hands it to you.

The steps, and the check on each

  1. Step 1: Read each option

    List the facts stated about each option, and what is not stated.

    Hands back: For each option, a list of stated facts and a list of facts that are missing.

    Check:Every number comes from the options

    A plain rule. If it fails, run the step again (the n8n file does this for you).

  2. Step 2: Build the comparison table

    Choose the factors that matter for this decision and compare every option on each.

    Hands back: A comparison table with a column for each option and a Difference column marked tie or differs.

    Check:Each factor is marked tie or differs

    A plain rule. If it fails, run the step again (the n8n file does this for you).

  3. Step 3: Drop the factors where options tie

    Remove the factors where the options are the same, so only the real differences are left.

    Hands back: A table of only the factors where the options differ, and a list of the ones dropped.

    Check:Only differing factors are left

    A plain rule. If it fails, run the step again (the n8n file does this for you).

  4. Step 4: Recommend one option

    Recommend one option with reasons tied to the factors that differ, and say what you give up.

    Hands back: A recommendation with reasons, what is given up, and the runner-up.

    Check:A recommendation is named

    A plain rule. If it fails, run the step again (the n8n file does this for you).

  5. Step 5: Argue the runner-up's case

    Make the best honest case for the runner-up, and say what would change the recommendation.

    Hands back: The runner-up's best case and the facts that could change the recommendation.

    Check:The facts that could change the answer are named

    A plain rule. If it fails, run the step again (the n8n file does this for you).

After the last step

  • Check:Every number in the recommendation is in the options

    A plain rule. If it fails, run the step again (the n8n file does this for you).

  • Check:Every reason names a factor from the table

    A second look by the AI. If it fails, run the step again (the n8n file does this for you).

  • Check:You make the final call

    You check this one. If it fails, it is flagged for you.

What you need first

  • The options(needed)

    Paste what you know about each option. Give each one a name and put a blank line between options.

  • What you are deciding and for whom(needed)

    Say what the choice is, who it is for, and what matters most to them.

  • Must-haves and limits(optional)

    List anything an option must meet to be allowed, such as a budget or a deadline. Leave empty if there are none.

Ways to adapt it

  • To weigh one factor more than the rest, say so in "What you are deciding and for whom", for example "Cost matters twice as much as anything else".
  • To compare more than five options, first run the template with all of them, then run it again with only the top two.
  • If you do not know the facts yet, paste what you have anyway. The first step lists what is missing, so you know what to ask.
  • To get the answer in a format for your team, replace the shape at the end of the recommendation step.

Change any step or check in the Editor, then download it again.

Words that fit consultants

  • Option, vendor, bid and alternative mean the same thing here. The names in the options text become the column heads of the table.
  • A must-have is a condition the client sets, such as a budget or a start date. An option that fails one is shown as failing and is not averaged away.
  • A tie is a factor where the client would not notice a difference. Those rows are dropped and listed.
  • Not stated means the options text is silent. Ask the vendor before you tell the client.
Try it with a sampleSample

Three options for a client whose staff work partly on client sites; one is the cheapest but logs time only in the office, so it fails a must-have, and the other two are close on cost.

The options
Cloud tracker: 14 dollars per user a month, with no user limit. Staff log time on a phone or a laptop, from anywhere. Invoices are made automatically from the logged time. Set-up takes about 3 weeks. Reviews say the reports are easy to read but the phone app is slow. Monthly contract, cancel any time. Desktop suite: 3,900 dollars once for up to 40 users, plus 700 dollars a year for updates. It runs on the office server, so staff can log time only in the office. Invoices are made automatically from the logged time. Set-up takes about 8 weeks. Reviews say it is reliable but dated. Spreadsheet plus accountant: no license cost. Staff fill in a shared online spreadsheet each week, which they can open on any device. The firm's accountant builds each invoice by hand, at 60 dollars an hour, about 10 hours a month. Set-up takes about 1 week. Nobody has reviewed this setup.
What you are deciding and for whom
Time tracking and invoicing for a 40-person engineering firm, half of whom work on client sites. Cost over three years and logging from a site matter most
Must-haves and limits
Staff must be able to log time from a client site. Budget is 30,000 dollars over three years at most. Must be running by February 1.

The scoring guide

Sample

One criterion for each check, marked met or not met. (The guides in Quality rate a whole piece of work on three levels.) Under each one is how the three sample runs scored.

Download the scoring guide (Markdown)

  • Every number comes from the options

    A plain rule

    Meets:Every number in the factors comes from your options text.
    Does not meet:A number is not in the options, or was worked out or rounded.
    Claude:Met
    ChatGPT:Met
    n8n:Met

  • Each factor is marked tie or differs

    A plain rule

    Meets:Every factor is marked tie or differs.
    Does not meet:A factor has no mark, or another word.
    Claude:Met
    ChatGPT:Met
    n8n:Met

  • Only differing factors are left

    A plain rule

    Meets:Only factors marked differs are left in the table.
    Does not meet:A factor marked tie is still in the table.
    Claude:Met
    ChatGPT:Met
    n8n:Met

  • A recommendation is named

    A plain rule

    Meets:A line starting Recommendation: names one option.
    Does not meet:There is no recommendation line, or it names more than one option.
    Claude:Met
    ChatGPT:Met
    n8n:Met

  • The facts that could change the answer are named

    A plain rule

    Meets:A line starting What would change my mind: names the facts that could change the answer.
    Does not meet:The line is missing, or it only repeats the recommendation.
    Claude:Met
    ChatGPT:Met
    n8n:Met

  • Every number in the recommendation is in the options

    A plain rule

    Meets:Every number in the recommendation is in your options text.
    Does not meet:The recommendation has a total, average or other number that the options never gave.
    Claude:Met
    ChatGPT:Not met
    n8n:Met

  • Every reason names a factor from the table

    A second look by the AI

    Meets:Every reason points to a factor that is in the table.
    Does not meet:A reason rests on something the options never said.
    Claude:Met
    ChatGPT:Met
    n8n:Met

  • You make the final call

    You check this one

    Meets:You read the recommendation and the runner-up's case, and no fact you know changes the answer.
    Does not meet:You know something the options text does not say that changes the answer.
    Claude:For your read
    ChatGPT:For your read
    n8n:For your read

Claude: met 8 of 8 (including 1 for your read)

ChatGPT: met 7 of 8 (including 1 for your read)

n8n: met 8 of 8 (including 1 for your read)

Learn how it works:Several anglesAdapting a template

Practice:Scoring guidesCheck it yourself

Which one fits?

Related templates

All for Consultants

Write

Write

Write

Last run
October 6, 2026: n8n, import format 2.20 (sample run)
Tools and versions
Claude, Sonnet model; ChatGPT, GPT model; n8n, import format 2.20
What was checked
The files, against each tool’s import format, and 3 sample runs, each check’s finding recorded.
Version
1.0
What changed
First release.

Report a problem with this templateHow we test