An AI post implementation review template closes the delivery loop the toolkit opened at discovery — and it exists because the review is the engagement’s most skipped instrument and its most quietly valuable one. Implementations end in one of two default modes, both wrong: the quiet fade (the install graduates, the retainer hums, nobody ever formally asks how did that go against what we said) or the victory lap (a celebration meeting where wins get toasted, misses get unmentioned, and the deck’s job is making everyone feel good about renewing) — and both modes waste the same asset: the engagement just produced a complete experiment, with a written hypothesis (the proposal’s calculator, the charter’s gates, the scope page), a full evidence record (the baseline, the delivery ledger, the monthly reports), and two parties who now know things they didn’t — and the review is the only instrument that harvests it. The template’s architecture is the toolkit’s epistemology aimed at the practice itself: the original promises laid beside the actual evidence, line by line (the calculator’s bands vs. the measured months, the gates as adjudicated, the timeline vs. the ledger — variances named in all three directions: better, worse, different); the causes binned honestly (ours, theirs, discovered — the slip protocol’s taxonomy, applied retrospectively); the client’s verdict gathered structurally, not ambiently; and every finding dispositioned — into the roadmap’s next horizon, the instruments’ changelogs, or the practice’s pattern file — because a lesson without a disposition is a feeling, and the review’s product is changed behavior on both sides.
The instrument’s market context, from the standing frame: according to McKinsey’s Superagency in the Workplace report (2025), 92% of companies plan to increase their AI investments over the next three years, yet only 1% describe their AI deployment as mature — and the maturity gap is partly a learning-loop gap: organizations (and their vendors) that ship, declare, and move on, accumulating deployments without accumulating judgment. The review is where the practice compounds differently: run at every graduation (four to six weeks after the dial turns, when steady-state evidence exists), it feeds the roadmap’s re-plan, sharpens every instrument’s next version, and — commercially — converts honest reckoning into the renewal conversation’s foundation, because per the whole library’s contrast economics, the vendor who audits their own promises is the vendor whose next promise means something. (All revenue figures in this post are illustrative business math, not guarantees; individual results vary.)
This guide is the template: the timing and the cast, the promise-versus-evidence ledger (the review’s spine), the variance taxonomy, the client-verdict structure, the disposition rule, the internal half (the practice’s own retro), and the honest realities — including the review that was a celebration wearing a clipboard.
Timing, Cast, and the Two Halves
When: steady state, not the finish line. Four to six weeks post-graduation — after the ramp the expectation scripts pre-named, when the monthly report holds real steady-state numbers — because a review run at go-live measures the launch’s adrenaline, and the instrument’s whole premise is evidence. (Engagements that ended in diagnosis rather than graduation get the review too — especially those, per the pilot playbook’s failure-protocol doctrine: the honest ending’s findings are the relationship’s remaining asset, and harvesting them well is frequently what earns the next engagement anyway.)
Who: the deciders and the doers. The client’s decision-maker and the workflow owner (the scoping session’s cast, returning), the champion (the front line’s verdict carrier), and the practice’s delivery lead — one hour, the ledger pre-circulated, per the standing no-ambush rule: nobody should meet a variance for the first time in the meeting.
The two halves. The joint review (this post’s spine — run with the client), and the internal retro (the practice’s own, run first, alone): the same ledger walked privately, the ours bin filled honestly before the client’s version could tempt softening, and the instrument-changelog candidates drafted — because the joint review’s credibility depends on the practice arriving having already told itself the truth.
The Promise-Versus-Evidence Ledger
The review’s spine: one table, three columns — what we said (sourced: the proposal’s calculator bands, the charter’s four gates, the scope page’s included list, the timeline’s committed quarter), what happened (sourced: the monthly reports, the gate adjudications, the delivery ledger, the change-order record), variance and bin — walked line by line, with the discipline that makes it an instrument rather than a mood:
The economics lines. The calculator’s low and expected bands against the measured months — with the reading rules the calculator post promised: the low case cleared or it didn’t (the renewal’s honest foundation either way), the ramp compared to the pre-named ramp (the expectation scripts, graded), and the attribution kept conservative even in victory (per the standing religion — a review that inflates its own wins is the promise spreadsheet’s sequel). Where the numbers beat the bands, the review says so plainly and asks why (the capture rate that outperformed teaches the next calculator); where they missed, the bin gets filled and the mitigation gets a date.
The gates and the adoption lines. The four gates as adjudicated, plus the adoption metrics’ steady-state read (usage sustained? workarounds extinct or reborn? the escalation path trusted?) — the change template’s watched-like-uptime doctrine, formally revisited, because adoption decay between graduation and review is the earliest churn signal the practice will ever get.
The delivery lines. Timeline against the ledger (the slips with their bins — the discovered items especially, because discovered scope is the instrument-sharpening goldmine: what did the map miss, what did the substrate checks not catch, which milestone definitions bent), and the dependency ledger’s client-owed record read factually (the symmetric tracking cashing its check: the access that aged, the approver that bottlenecked — named kindly, because half the client’s own learning lives here).
The variance taxonomy’s third direction: different. Not just better/worse — different: the value that showed up where nobody predicted it (the front desk’s reclaimed calm that the calculator never priced, the recall engine that outearned the intake line), captured deliberately, because unexpected value is the expansion map’s raw material and the testimonial’s honest substance (gathered under the standing consent-and-composite rules where it travels into content).
The Client’s Verdict and the Disposition Rule
The verdict, structured. Three questions, asked directly and recorded: what should we keep doing (the practices to protect), what should we change (the friction the practice couldn’t see from inside), and what almost made you doubt us (the question nobody asks and the answer worth the whole meeting — the moment in week five, the invoice that surprised, the silence that worried — surfaced now, while it’s a finding, instead of at renewal, when it’s a decision). The champion gets the same three questions separately, for the front line’s version — per the standing cross-level discipline, because the decider’s verdict and the doer’s verdict diverge exactly where the next engagement needs to know.
The disposition rule: every finding lands somewhere, in writing. Into the roadmap (the different-value findings and the shelved near-misses re-ranked — the review feeding the quarterly re-plan directly, per the toolkit’s mesh); into the instruments (the map template’s new question, the calculator’s recalibrated capture rate, the onboarding script that would have pre-framed the week-five moment — each a changelog entry with the review as its citation); into the SOW anatomy (the clause the friction taught — the quarterly counsel review’s inbox); or into the pattern file (the vertical’s accumulating playbook, anonymized per the standing rules). The dispositions get owners and dates — the risk register’s format, aimed at the practice’s own improvement — and the next review checks them, which is what makes the learning loop a loop. The review’s deliverable: the ledger, the verdict, and the dispositions on two pages, shipped to the client within 48 hours (the workshop’s memo discipline) — the document that opens the renewal conversation having already done its hardest work. We do not build the AI. We implement it — and the review is where the implementing practice implements on itself.
Why the Honest Mirror Wins
The structural recommendation: run the review at every ending — promises beside evidence, variances binned in three directions, the verdict asked structurally, every finding dispositioned with an owner — because the engagement’s last instrument decides whether the practice compounds judgment or just accumulates deployments.
The reasoning is structural:
- The review is the measurement religion’s integrity test: a practice that baselines clients, gates pilots, and evidences milestones — then exempts its own promises from the same audit — has built the phantom list’s blind spot at the center of its brand; the review is where the epistemology proves it was never just for show.
- The honest variance record is the renewal’s strongest possible foundation: the client who watched the practice name its own misses, bin them as ours, and date the fixes has evidence about the one thing references can’t prove — how this vendor behaves about being wrong — which, per the standing trust economics, is what the second engagement is actually purchased on.
- The disposition rule is the compounding mechanism the toolkit was missing: seventeen instruments that never change are a museum; the review is the changelog’s feeder, converting every engagement’s friction into the next engagement’s sharpened edge — the practice’s judgment as a versioned asset, growing on the same cadence as everything else in this library.
- And the almost-doubted question is the churn radar nothing else provides: clients rarely announce eroding confidence — they renew smaller, then don’t; the structured verdict surfaces the erosion while it’s still a finding, which makes the review, quietly, the retainer book’s best insurance.
I graduated from Vanderbilt. Almost went straight into investment banking. I spent years at Vanderbilt University reading the same labor reports and McKinsey decks that documented the trends now defining 2026 — and I came away with one inescapable conclusion: a salary has a ceiling. Inflation doesn’t.
I decided not to try and outrun inflation with a salary. I replaced my corporate salary by implementing pre-built AI tools we leverage — Intercom AI, Helios AI, and n8n at the core, plus the broader implementation stack — for service businesses with operational gaps they can’t fix on their own.
What Most Articles Won’t Tell You About Reviews
A few honest realities:
The failure mode with your name on it is the Victory Lap. It’s the review run as celebration — the deck of wins, the numbers curated toward their best angles, the misses reframed as learnings-without-owners, the meeting’s real agenda being the renewal it precedes — and it fails precisely because it usually works in the room: everyone leaves warm, the renewal often signs, and the two costs land later, where they’re expensive. First, nothing was learned: the calculator’s optimistic capture rate goes uncorrected into the next proposal, the map’s blind spot ships into the next field week, the week-five doubt goes unspoken until it’s a non-renewal — the practice accumulating engagements without accumulating judgment, which is the 92/1 gap’s vendor-side mechanism. Second — subtler — the client half-knows: operators can tell a curated deck from an audit, and the victory lap teaches them that this vendor’s self-reports require discounting, which quietly re-prices every future claim the practice makes, including the true ones. The tell is a review with no ours bin entries and no almost-doubted answer; the cure is the template run whole — the internal retro first, the promises sourced, the variances binned in the open, the dispositions dated — plus the sentence installed where the renewal’s warmth tempts: the celebration is allowed — after the audit; the mirror comes before the toast, or the toast is the only thing anyone will remember.
The diagnosis-ending review is the practice’s character on display. The pilot that didn’t clear its gates gets the fullest, kindest, most rigorous review of all — findings harvested, the substrate’s lessons scoped, the relationship’s next honest move named — because per the standing declination economics, how endings are handled is the vertical room’s favorite story.
Reviews feed content under the standing rules only. The different-value findings and the cleared low cases become case material through the composite-and-consent discipline — labeled, permissioned, illustrative — never lifted raw from the ledger; the review is an instrument, not a testimonial farm.
Small engagements get the small version — never none. The governance-page engagement’s review is fifteen minutes and three questions; the ritual scales down, the exemption doesn’t — because the smallest engagements are where tomorrow’s retainer relationships are auditioning both sides. The standing arithmetic (3-5 clients = full-time corporate-equivalent income working a few hours a week once implementations stabilize) holds with the review as each relationship’s learning heartbeat. You learn a skill instead of buying into a business model — and in reviews, the skill’s signature is the client who watched you grade your own promises and signed the next one because of it. (Illustrative math throughout; results vary.)
According to McKinsey’s Superagency in the Workplace report (2025), 92% of companies plan to increase their AI investments over the next three years, yet only 1% describe their AI deployment as mature. The consultants who own the ending in 2026 are not the ones with the best celebration decks. They’re the ones who laid the promises beside the evidence, binned their own misses first, and dispositioned every finding — and whose practices got measurably sharper with every engagement, which is what compounding actually looks like.
Run the First Mirror This Month
The action sequence for ai post implementation review template:
This week: The template assembled — the three-column ledger format, the variance bins, the three verdict questions, the disposition format with owners and dates.
This month: The most recent graduation reviewed — internal retro first, ledger pre-circulated, the hour held, the two-pager shipped in 48.
Per engagement: The review calendared at graduation (four to six weeks out); the champion’s separate verdict gathered; dispositions into the roadmap, the changelogs, the pattern file.
Ongoing: The next review checking the last one’s dispositions; the instruments sharpening on citation; the lap declined every time the renewal’s warmth offers to replace the audit. (Illustrative trajectories; results vary.)
The engagement was an experiment with a written hypothesis — so harvest it. Promises beside evidence. Variances in three directions. The verdict asked out loud. Every finding dispositioned, owned, dated.
The mirror before the toast — that’s the whole template, and it’s how a practice compounds judgment instead of just deployments.
Pick the industry. Take the first step. If you want to see the playbook fully in action – tap here to start.
