Guide · Published November 14, 2026 · Reviewed November 14, 2026 · Ingenious AI team · Not legal advice
ADMT Notice Kit vs Spreadsheet Tracking: Where Spreadsheets Break and What Replaces Them
The short answer: A spreadsheet can start an ADMT programme; it cannot finish one. It genuinely handles a register (the tool inventory) and a simple log at low volume. It breaks at four points the duties make inevitable: per-person document generation (every notice and explanation is a document, not a row), per-decision clocks (Colorado's 30 days runs per case, silently, including weekends and absences), tamper-evidence (a spreadsheet can be edited without a trace, which is the opposite of what the record is for), and linked follow-ups (corrections, re-runs, and reviews must attach to the decision they concern). The ADMT Notice Kit exists to cover exactly those four: generated documents, a deadline tracker, a SHA-256 hash-chained record, and a review queue on the same chain.
Check your duties and dates — free Build your first notice — free
What is this comparison for? It is an honest tooling assessment for a business deciding how to run the duties in this series. It is written by the kit's own team and says so — 'Ingenious AI team' is on the byline — so read its claims the way you would any vendor's: the checkable ones are the product facts (generation, clocks, chain, queue, plans), and the free generator and the Free plan exist so you can verify them before paying anything.
What a spreadsheet genuinely handles well
Credit where due: the inventory register is a spreadsheet-shaped problem — rows per tool, columns for decisions, states, weights, review dates — and a spreadsheet serves it well at any scale this series contemplates. A low-volume operator (a handful of decisions a month, one state) can also run a simple sent-log on a sheet honestly, provided the documents themselves are kept with it and the dates are entered on the day.
The failure mode is not the tool but the growth: volume rises, a second state arrives with a different clock, the log-keeper goes on leave, a correction needs the original decision's reasons. The question is never 'spreadsheet or system?' in the abstract — it is 'at what point do the four breaks below cost more than the system?' For most operators in these guides' examples, that point is the second state or the fiftieth monthly decision, whichever comes first.
Break one and two: documents and clocks
Notices and explanations are documents addressed to named people, generated from a profile and per-case facts. In a spreadsheet world they are mail-merge projects with version chaos: which template version went to March's applicants, which contact was current, which reasons were entered for this file? The kit generates each document from the business profile and the case facts, and files the generated document in the chain — the version question answers itself.
Clocks break second. Colorado's 30-day count is per decision and unforgiving of batch habits; a spreadsheet can compute it, but the sheet does not watch itself — somebody must open it, sort it, and act, every working day, including the days nobody does. The kit's tracker exists so the approaching deadline is a property of the system, visible on the dashboard, not a property of someone's diligence. Connecticut's 45-day answers have the same shape at lower urgency.
Worked example: Willow Creek Home Care
Example — fictional business, real duty structure. Willow Creek, the fictional agency from the notice-writing guide, migrated its programme off a spreadsheet after its first quarter. The breaks it hit, in order:
- Documents: by month two, the mail-merge had produced 3 template versions in simultaneous circulation, one with a departed employee's email as the review contact — found by an applicant, not by the firm.
- Clocks: a two-week absence by the log-keeper produced 2 explanations sent after day 30 (days 34 and 38) — the sheet had computed the dates correctly; nobody was watching it.
- Linkage: a correction referencing a decision from the prior quarter took 5 hours of cross-tab archaeology to assemble — the case existed, in pieces, in four places.
- Migration: the register imported as-is in an afternoon; 19 open cases backfilled with documents and dates; closed cases archived as a referenced log with its limits stated.
- After: the next quarter runs 96 decisions fully in the kit — generated notices, tracker-watched clocks, corrections opening against cases — and the one-hour test passes in 11 minutes.
Break three and four: tamper-evidence and linkage
A compliance record's value is its credibility to a later reader. A spreadsheet's edit history — where it exists at all — is an account-level feature, not a property of the file that gets exported. The kit's record is hash-chained: each entry's SHA-256 fingerprint incorporates the previous, so alteration is detectable in the export itself. That difference is the difference between 'trust our file' and 'check our file', and it is the reason the records guide treats the chain as a first-class feature rather than an implementation detail.
Linkage breaks last and worst. A correction arrives referencing a decision from two months ago; the sheet has the decision on one tab, the correction on another, the re-run nowhere. Reconstructing the case means human archaeology across tabs and inboxes. In the kit, the correction opens against the case, the re-run appends to the chain, and the export produces the whole story in order — the property every other guide in this series quietly depends on.
The honest middle: what the kit does not do
The kit does not make decisions, judge fairness, or give legal advice — its documents say so on their face. It does not replace the inventory thinking, the vendor negotiation, the reviewer staffing, or the threshold analysis; those remain the business's work, guided by the other guides in this series. It does not cover regimes outside its data — its checker states the GDPR is a different regime and outside it.
Pricing is stated plainly so the comparison can be completed: Free (Colorado, one decision type, core record log), Pro $49/month (all three states, explanation builder, review queue, exports), Team $99/month (multi-seat review queue, all states). Against the spreadsheet's real cost — staff time per document, clock-watching, reconstruction hours, and the risk position of an editable record — the comparison is arithmetic each business should run on its own volumes. The full plan detail is on the pricing page.
Migration: moving from a sheet without breaking the chain
Businesses that started on a spreadsheet should migrate deliberately: import the register as-is (it is already the right shape), backfill open cases (decisions whose clocks or follow-ups are live) with their documents and dates, and let closed historical cases remain as an archived sheet referenced from the new record — with the archive's limits stated honestly in the file (it is a log, not a chain).
From the migration date, everything new is generated in the kit: notices, explanations, the tracker, the queue. The one-hour test from the records guide is the acceptance check — pick a live case from last month and produce its full file from the new system. If the business is still deciding, generate a notice in the free generator and compare its structure with your mail-merge: the documents are the product, and they can be inspected in two minutes.
- Keep the spreadsheet for what it is good at — the inventory register — and be honest about the four breaks for everything else
- Count your real volume and states: the break point is usually the second state or the fiftieth monthly decision
- Price the spreadsheet honestly: staff time per document, daily clock-watching, reconstruction hours, an editable record's risk position
- Verify product claims before paying: generate documents in the free generator, and start on the Free plan if your shape fits it
- Migrate deliberately — register as-is, open cases backfilled, closed history archived with its limits stated — and accept on the one-hour test
Sources, as carried in the kit's own data: the duty summaries in app/rules.py and the checker data in app/checker_data.py, verified against primary or authoritative sources on October 28, 2026. Run the free duties checker for your jurisdiction, decision type, and the tool's role, and compare the kit's plans for the record, builder, and review queue. More from Ingenious AI is on the hub.
Related guides
- Colorado AI Act Automated Decision Notices: What to Send
- Adverse-Outcome Explanations in Plain Language
- ADMT Deadline Tracker: Which State Duties Land When
- California CCPA ADMT Requirements: What Your Business Must Do by January 1, 2027
See all ADMT guides for the full set.
Frequently asked questions
Can we run ADMT compliance on a spreadsheet?
Partly, at low volume: the inventory register and a simple sent-log work on a sheet. It breaks on per-person document generation, per-decision clocks that need daily watching, tamper-evidence, and linking follow-ups to their decisions — the four things the kit is built around.
What does the kit do that a spreadsheet cannot?
Generates each notice and explanation from the business profile and case facts, tracks Colorado's 30-day clock per decision, keeps a SHA-256 hash-chained (tamper-evident) record, and runs the human-review queue on the same chain, with exports.
What does the kit cost?
Free covers Colorado, one decision type, and the core record log. Pro is $49 per month (all three states, explanation builder, review queue, exports). Team is $99 per month (multi-seat review queue, all states).
What does the kit not do?
It does not make decisions, judge whether a decision was fair, or give legal advice — its documents state that. Inventory analysis, vendor negotiation, reviewer staffing, and threshold questions remain the business's work.
How do we migrate without losing history?
Import the register, backfill open cases with their documents and dates, and archive closed cases as a referenced log with its limits stated. Run the one-hour test on a live case as the acceptance check.