What 1086 AI-reviewed decks taught us
Consultants open strong. They close weak.
We ran every fully-extracted deck through an editorial AI critic. 1086 reviews · 34 firms · eight dimensions scored 0–100. One pattern rules them all.
0
AVG OPENING
→
0
AVG CLOSING
−16.0
GAP · EVERY FIRM SHOWS IT
↓ Scroll for the evidence
1. Every deck, plotted
Hover any point · click to open the deckEach dot is one deck. Its x is how well it opens, its y is how well it closes. If consulting decks were balanced, dots would hug the diagonal. They don't — they crowd the bottom-right: strong opener, weak finisher.
3. Exemplars vs cautionary tales
The highest and lowest-scored decksTop decks share three things: an answer-first opener, explicit MECE pillars, and a closing slide that makes the ask. Bottom decks are nearly always the same sin — a research report pretending to be a deck.
Top 10 · learn from these
82
American Express Investor Day 2024
A disciplined, thesis-led investor-day deck with genuine MECE pillars and metric-rich action titles -- a useful Storymakers exemplar for opening structure and pillar architecture, but its navigation bloat and missing complication act make it a partial, not whole-deck, teaching reference.
80
The future of demand
A well-structured thought-leadership deck with a clean SCQA arc and answer-first exec summary — strong Storymakers exemplar for opening/closing discipline, but the overlapping middle dividers make it a flawed template for teaching MECE pillar design.
80
Banking Consumer Study 2025
A genuinely strong Storymakers exemplar for pillar architecture and imperative recommendation titles, weakened only by a soft, metaphor-led opening and figure pages that hide their insight in the caption.
78
AUTOMOTIVE –OES
Competent Accenture research report with a legible SCQA spine and strong quantified titles, but the recommendation act is under-built relative to the diagnosis — use the opening (p.2-4) and transitions (p.19, p.22) as Storymakers teaching examples, not the resolution.
78
FULL VALUE. FULL STOP How to scale innovation and achieve full value with F
A well-structured analytical benchmark report with a clear Think→Act→Move spine and evidence-rich recommendation titles — use the numbered 'Act Like' section (p.17-22) as a Storymakers teaching example, but treat the opening and the soft p.26 close as cautionary tales of burying the stakes and under-specifying the call to action.
78
Green by Default
Well-structured Accenture thought-leadership report with clear MECE pillars and several sharp action titles — use the sectioning (p.15/20/25) and insight-title examples (p.4, p.5, p.10) as teaching exemplars, but flag the repeated generic recommendation titles and soft closing as common pitfalls to avoid.
78
Make the leap, take the lead: Tech strategies for innovation and growth
A well-architected analytical thought-leadership deck with a strong MECE pillar (Replatform/Reframe/Reach) and quantified narrative — use it as a teaching example for pillar design and action-titling, but not for opening hook or closing CTA.
78
The art of AI maturity Advancing from practice to performance
A disciplined Accenture thought-leadership deck with a genuine SCQA spine and a clean five-pillar recommend+case-study build — use the divider ladder and pillar pairing as a teaching example, but not the soft landing or the label-style analytical titles.
78
Moneyball Moment Marketing Canada
A well-structured analytical-thesis deck with thesis-bearing section dividers and number-anchored titles — use the pillar architecture and action-title craft as a Storymakers exemplar, but not the opening or the call-to-action close.
78
Beyond good intentions
A solid short-form point-of-view deck with clean SCQA bones and a strong three-pillar resolution — use p.10-13 as a teaching example of how to mirror a recommendation across slides, but nudge the opening and add dividers to make it truly exemplary.
Bottom 10 · avoid these traps
18
gol 6
This is a financial-product fact sheet with disclaimers, not a Storymakers consulting narrative — useful only as a counter-example of what happens when a document has no action titles, no arc, and no recommendation.
22
Review of efficiency of the operation of the federal courts
This is an educational primer on how the U.S. federal courts work — not a consulting argument — and serves as a counter-example for Storymakers, useful only to illustrate what happens when a deck has topic labels but no thesis, analysis, or recommendation.
25
20230316 scff portfolio details
A portfolio-disclosure reference document masquerading as a deck — useful as a counter-example of topic-label titles and missing narrative, not as a Storymakers exemplar.
25
mi guide to the markets uk
A best-in-class market reference atlas with consistent grammar and rich callouts, but as a Storymakers exemplar it is the opposite of one — use it to teach exhibit hygiene and footnote discipline, never to teach narrative, action titles, or how to land a recommendation.
25
guide to the markets asia
A best-in-class market reference book judged against its own genre, but a near-zero Storymakers exemplar — use it to teach how reference decks differ from narrative decks, never as a model for action titles, SCQA, or pillar structure.
25
Second Quarter 2023 Results
This is an earnings-disclosure deck, not a consulting argument — topic-label titles, no SCQA arc, and a closing half built entirely of reconciliation tables; useful as a counter-example of what Storymakers principles are designed to replace, not as an exemplar.
25
Third Quarter 2024 Results
A standard Gartner earnings/IR deck — competent as a reference document but a near-anti-pattern for Storymakers, useful only as a 'before' example to demonstrate why topic titles and appendix-heavy structures fail to tell a story.
25
Second Quarter 2024 Results
A standard investor-relations earnings deck — competent as an IR document but a weak Storymakers exemplar: use it only as a negative example of topic-label titling and appendix-as-ending, not as a structural model.
25
Scott Anthony Press Kit
A functional press kit, not a Storymakers exemplar — useful only as a negative example of topic-label titles and a missing CTA close.
28
guide to the markets au
An exemplary reference data-book and a poor Storymakers exemplar — use it to teach taxonomic MECE structure and chart cadence, but use it as a counter-example for action titles, opening thesis, section dividers, and closing recommendation.
4. What the AI editor kept asking to fix
Click a theme to see real suggestionsWe asked the AI critic to write 3 concrete improvements per deck. Grouped by theme, the same five asks dominate. Click any theme to see the verbatim editorial notes.
"Rewrite topic-label titles (p.10, p.23, p.24, p.32-33) into declarative insights, and add a one-line thesis recap on p.29 above the CTA"
"Add a prioritization/roadmap slide before p.45 that ranks the 10 trends by impact and urgency, and a concrete next-steps page replacing the current generic 'Continue the conversation' CTA"
"Signpost p.13-18 as four explicit pillars (one divider + four titled pillar slides) so the recommendation structure is visible in the deck's skeleton, not just inferred"
"Replace the p.21 closing with a concrete 90-day or first-move recommendation tied to the roadmap on p.20 — a Storymakers close names the next action, not the risk"
"Move p.31 'CREATING AN AGENDA FOR CHANGE' before the Appendices divider so the close ends on a forward-looking agenda, not mid-appendix"
"Insert a single CTA slide between p.26 and p.27 with 3 prioritized actions for the reader's first 90 days, then let p.27 serve as the emotional close"
"Collapse the three framework restatements (pp.6, 13, 14) into one canonical DNA slide and reclaim two pages for a recommendation-and-roadmap block"
"Add an explicit closing CTA slide after p.36 that names 2-3 decisions the reader must make this quarter and reconnects to the $371B opportunity from p.4"
"Promote p.45's 'watershed year' thesis to p.2 and restructure the 10 trends as 3-4 MECE pillars (e.g., AI-led productivity / Risk & regulation / Tech & operating model) with a synthesis map up front"
"Signpost p.13-18 as four explicit pillars (one divider + four titled pillar slides) so the recommendation structure is visible in the deck's skeleton, not just inferred"
"Add explicit 'STEP TWO: SHIFT FOCUS FROM INSTITUTIONS TO INDIVIDUALS' and 'STEP THREE: EMPOWER VULNERABLE LEARNERS' divider slides before p.22 and p.25 to deliver the three-step MECE promise made on p.16"
"Make the three-pillar model (Boundaryless / Adaptable / Radically Human) explicit on one framework slide before p.13-15 so readers can hold the MECE structure while reading each pillar"
"Rewrite the three 'Practical considerations...' titles as distinct one-line insights unique to each pillar (e.g., 'Start with lifecycle redesign, not packaging tweaks')"
"Align section divider wording with the Design/Drive/Deliver framework so each pillar uses the same verb family"
"Insert 2-3 named pillar dividers (e.g. 'E-commerce acceleration', 'Refurbished shift', 'Omnichannel stickiness') so the analytical middle reads as MECE rather than a stream"
"Insert sub-pillar dividers inside act 3 — e.g. 'Tech', 'Talent', 'Agility', 'Culture' — so the 15-slide block reads as four MECE moves rather than a topic dump"
"Replace p.1 slogan with a single-sentence thesis cover ('Only 2 in 10 Auto-OES companies scale digital PoCs successfully — here is what the Champions do differently')"
"Rewrite topic-label titles (p.10, p.23, p.24, p.32-33) into declarative insights, and add a one-line thesis recap on p.29 above the CTA"
"Promote p.29 and p.30 out of the appendix into the opening context block (around p.4-6) - they establish why the $1,920 matters at a population level"
"Promote p.45's 'watershed year' thesis to p.2 and restructure the 10 trends as 3-4 MECE pillars (e.g., AI-led productivity / Risk & regulation / Tech & operating model) with a synthesis map up front"
"Compress p.2-5 into a single executive-summary slide that states the thesis ('Champions convert ~X% more revenue to recurring services by doing four things') so the reader has the answer before the evidence"
"Cut two of the three opening quote slides (p.3–5) and move the $11.5tn thesis from p.7 forward to p.3 so the answer leads"
"Consolidate slides 4-8 into a two-slide Situation+Complication spread so the opening moves faster to the 'two paths' framework on p.10"
"Move the '46% revenue-at-risk by 2023' hook from p.25 to within the first three slides so the stakes are established before the methodology"
"Promote p.29 and p.30 out of the appendix into the opening context block (around p.4-6) - they establish why the $1,920 matters at a population level"
"Cut two of the three opening quote slides (p.3–5) and move the $11.5tn thesis from p.7 forward to p.3 so the answer leads"
"Move p.31 'CREATING AN AGENDA FOR CHANGE' before the Appendices divider so the close ends on a forward-looking agenda, not mid-appendix"
"Collapse the three framework restatements (pp.6, 13, 14) into one canonical DNA slide and reclaim two pages for a recommendation-and-roadmap block"
"Consolidate slides 4-8 into a two-slide Situation+Complication spread so the opening moves faster to the 'two paths' framework on p.10"
"Collapse Chapter 2: either lead with the four plays (the 'what') and demote the three pivots to a how-to appendix, or vice versa — do not stack two frameworks plus a multiplier model across 15 slides"
"Add an explicit 'Where to start' or diagnostic checklist slide between p.28 and the appendix so the deck closes on an imperative, not a restatement"
"Replace p.14 with a genuine 'so-what' slide: 3 prioritised actions for retailers/OEMs derived from the p.12 diagnosis, with owners and timing — move Contacts/About/Methodology behind a real appendix divider"
"Split p.20-21 into four parallel practice slides, each with its own action-title insight — the resolution deserves equal real estate to the diagnosis"
"Rewrite every 'Trend N | Topic' label as a declarative action title that states the so-what (e.g., replace p.23 'Trend 4 | A whole new way of working' with 'Only 26% of bank CEOs have a future-ready strategy — talent reinvention is the 2024 bottleneck')"
"Rewrite topic-label titles into action titles: 'Barriers to holistic transformation' → 'Change fatigue, budget, and talent gaps are the top three blockers'; 'Conclusion' → state the core recommendation in one line."
"Split pp.17-19 into one-lever-per-slide so each recommendation has a standalone action title and owner, rather than compound titles like p.19's "Grow revenue through AI; manage multiplatform inventory...""
"Rewrite every Figure title as an action title: 'Figure 3' becomes 'Four plays cover every revenue pool banks can still claim'; 'Figure 9' becomes 'Main banks already own 82% share — the upside is share-of-wallet, not share-of-customer'"
"Rewrite slide 3 to lead with the answer: make '152 platform adopters earned 2.1 pp higher margins' the action title, and move the definition to a sidebar"
"Convert each section title into an action title — replace '1. Growth levers' with the actual finding ('New revenue sources, not cost cuts, are the consensus path to growth — 73% agree')"
"Rewrite p.11 and p.12 action titles to state the finding ('Average maturity is just 56/100 — a third of the way to resilient') rather than reference figure numbers"
"Compress p.2-5 into a single executive-summary slide that states the thesis ('Champions convert ~X% more revenue to recurring services by doing four things') so the reader has the answer before the evidence"
"Move p8 ('Modern IT as competitive advantage') to directly before the recommendation block (around p18-19) so the narrative reads problem → evidence → answer, not answer → problem."
"Add an answer-first executive summary slide at p.2 or p.3 stating Accenture's 2-3 point of view on how COVID-19 permanently reshapes the consumer, before the supporting data"
"Replace every 'A quick take on X' title with a declarative claim (e.g. p.24 → 'Banking's data-lake playbook shows how to break siloes') and consolidate redundant quick-takes to tighten act 3"
"Tag each case study (pp.12-17) with the specific practice it illustrates (e.g. 'Travelers — Practice 1: Evangelize') so the case chain reads as evidence for the framework rather than a tour of logos"
"Replace the p.4 Drucker quote with a bridge slide that previews the flywheel components, or cut it and go straight from p.2 stakes to p.5 data"
"Thin the overcrowded slides (p.5 especially) by splitting multi-threaded messages into one-idea-per-page, or demote supporting data to an appendix so the title's claim stays the dominant voice"
"Rewrite the repeated 'What's happening / What agility looks like / Where it's working' titles into pillar-specific insights (e.g. p.11 → '57% of finance leaders cannot access the data real-time close requires')"
"Split p.20-21 into four parallel practice slides, each with its own action-title insight — the resolution deserves equal real estate to the diagnosis"
"Add a prioritization/roadmap slide before p.45 that ranks the 10 trends by impact and urgency, and a concrete next-steps page replacing the current generic 'Continue the conversation' CTA"
"Rebalance the arc: either compress the diagnosis (sections 2-3 are effectively the same argument told twice — incumbent side then client side) or expand recommendations to three full slides matching the three pillars, giving ecosystem partnerships its own page"
"Move the '46% revenue-at-risk by 2023' hook from p.25 to within the first three slides so the stakes are established before the methodology"
"Move p8 ('Modern IT as competitive advantage') to directly before the recommendation block (around p18-19) so the narrative reads problem → evidence → answer, not answer → problem."
"Replace p.14 with a genuine 'so-what' slide: 3 prioritised actions for retailers/OEMs derived from the p.12 diagnosis, with owners and timing — move Contacts/About/Methodology behind a real appendix divider"
"Collapse p.6-9 into one or two slides: the reader doesn't need four establish-context pages after the $3.2T prize on p.5 — move faster into the diagnosis on p.11"
"Replace the p.4 Drucker quote with a bridge slide that previews the flywheel components, or cut it and go straight from p.2 stakes to p.5 data"
"Rewrite the seven repeated 'Why alignment matters' titles with the specific insight each chart proves (e.g., p.15 → 'Aligning cyber with the business would save Business Blockers $294K per attack')"
"Collapse the geo spotlight (NA/Europe/Growth Markets) and the chart pack into a single MECE structure — e.g. Banking stress / Demand / Cost side / Financial conditions — so every divider earns its place relative to the 'financial fissures' thesis"
"Collapse the seven 'pulse check comparison' tables (p19-25) into a single trend-over-time visual plus 1-2 supporting tables; move the rest behind an Appendix divider."
"Reorder or re-title p.6-p.8 as an explicit 'value at stake' pillar and p.9-p.10 as a 'digital as enabler' pillar, with visual dividers, so the MECE structure is legible"
"Collapse the mid-deck frameworks (p.11, p.12, p.14, p.17, p.19) into one master 'levers' diagram and have the subsequent slides zoom into each lever, to reduce framework fatigue."
"Rewrite every 'FIGURE X' title (pp.9, 17, 18) as a declarative finding ('Expanding the customer base is CMOs' #2 priority for 2023' rather than the chart label)."
"Convert the question-title chart slides (p.22, p.24, p.25) into answer titles — e.g. p.22 'CEO-CFO alignment is the top lever CFOs use to overcome business challenges'"
"Rewrite ~40 topic-labeled titles as insight-bearing action titles — use p.43 as the template (state the so-what, not the subject of the chart)"
Most-used imperative verbs in suggestions
Add765
Rewrite729
Replace647
Insert374
Move146
Collapse137
Promote130
Convert89
Expand44
Tighten35
Rename35
Compress30
5. The vocabulary of praise and blame
What AI reviewers reach for↑ Praised
action titles644
quantified416
mece306
pillar262
headline195
metric155
action title152
opening107
framework106
hook64
benchmark50
scqa40
↓ Blamed
closing517
appendix346
no recommendation250
topic label223
topic labels217
buried108
missing81
weak63
filler43
weakening36
soft33
vague25