Can AI build consultant-grade decks?
I ran 13 model configurations across 4 consulting test cases and 53 published runs. Within each test case, every tool received the same prompt, source files, and constraints; each valid deck was scored on six dimensions: visual design, factual accuracy, narrative quality, clarity, editability, and data visualization.
Primary benchmarkCase 4 · AI and workforce transitionMaster-following test
- 13
- model configurations
- 4
- test cases
- 53
- published runs

Claude Code / Sonnet 5Case 4 · AI and workforce transition81/ 100
01 — LEADERBOARD
Quality first. Then time and cost.
Case 4 is the primary test: can an AI agent build an editable deck while following a supplied PowerPoint master and slide library? Its score combines task quality, factuality, narrative, clarity, visual design, and editability.
Best quality
Case 4 overall quality · Higher is better
9 configurations · Case 4 only
Fastest
Average generation time in minutes · Lower is better
8 configurations · Case 4 only
Cheapest
Reported cost per complete deck · Lower is better
4 configurations · Case 4 only
Reported cost captured for 4 of 9 case 4 configurations. Missing values are not ranked.
Cost not captured: Opus 5 · Sonnet 5 · Sonnet 5 · GPT-5.6 Sol · DeepSeek V4 Flash Latest
Supporting evidence
General presentation capability
Cases 1–3 · no PowerPoint master supplied
Analysis, storytelling, charting, and source handling across the unconstrained test cases. This score is intentionally separate from Case 4 master adherence.
Limited evidence: Fable 5 (1 of 3 cases) · GPT-5.6 Sol (3 of 3 cases) · GPT-5.6 Terra (2 of 3 cases) · DeepSeek V4 Pro (1 of 3 cases) · Gemini 3.7 Flash (2 of 3 cases) · Grok 4.6 (1 of 3 cases) · Kimi K3 (1 of 3 cases) · MiniMax M3 (2 of 3 cases) · Qwen 3.8 Max (1 of 3 cases). These configurations are shown for coverage, not ranked.
Permanent evidence
Every Test Case 4 result stays visible.
Legacy publication data · legacy release view
The prompt is unchanged across releases. Each result keeps its generation version, review status, compliance findings, reported cost, API-equivalent estimate, and elapsed time, even when it is not part of the current ranking.
| Result | Configuration | Version | Quality | Compliance | Reported cost | Estimate | Time | Evidence |
|---|---|---|---|---|---|---|---|---|
| 1Recorded | Sonnet 5Claude Code / Sonnet 5Historical rubric | local-v16local-v7 | 81 / 100Overall score | UnverifiedChecks not published | Not recordedProvider reported | Not recordedAPI-equivalent | 9m 43sGeneration | View deck preview for Sonnet 5 View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 2Recorded | Opus 5Claude Code / Opus 5Historical rubric | local-v16local-v7 | 69 / 100Overall score | UnverifiedChecks not published | Not recordedProvider reported | Not recordedAPI-equivalent | 11m 41sGeneration | View deck preview for Opus 5 View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 3Recorded | GLM 5.3 FlashOpenCode + OpenRouter / GLM 5.3 FlashHistorical rubric | local-v16local-v9 | 68 / 100Overall score | UnverifiedChecks not published | $0.05Provider reported | Not recordedAPI-equivalent | 42m 00sGeneration | View deck preview for GLM 5.3 Flash View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 4Recorded | GPT-5.6 SolCodex / GPT-5.6 Sol / MaxHistorical rubric | local-v16local-v7 | 67 / 100Overall score | UnverifiedChecks not published | Not recordedProvider reported | Not recordedAPI-equivalent | 10m 01sGeneration | View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 5Recorded | Grok 4.6OpenCode + OpenRouter / Grok 4.6Historical rubric | local-v16local-v16 | 65 / 100Overall score | UnverifiedChecks not published | $0.77Provider reported | Not recordedAPI-equivalent | 8m 45sGeneration | View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 6Recorded | Sonnet 5Claude Code / Sonnet 5Historical rubric | local-v16local-v16 | 60 / 100Overall score | UnverifiedChecks not published | Not recordedProvider reported | Not recordedAPI-equivalent | 17m 51sGeneration | View deck preview for Sonnet 5 View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 7Recorded | DeepSeek V4 ProOpenCode + OpenRouter / DeepSeek V4 ProHistorical rubric | local-v16local-v16 | 60 / 100Overall score | UnverifiedChecks not published | $0.24Provider reported | Not recordedAPI-equivalent | 13m 56sGeneration | View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 8Recorded | Qwen 3.8 MaxOpenCode + OpenRouter / Qwen 3.8 MaxHistorical rubric | local-v16local-v7 | 60 / 100Overall score | UnverifiedChecks not published | $0.81Provider reported | Not recordedAPI-equivalent | 16m 35sGeneration | View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
| 9Recorded | DeepSeek V4 Flash LatestOpenCode + OpenRouter / DeepSeek V4 Flash LatestHistorical rubric | local-v16local-v7 | 51 / 100Overall score | UnverifiedChecks not published | Not recordedProvider reported | Not recordedAPI-equivalent | Not recordedGeneration | View ledgerScore ledger
PowerPoint complianceCompliance checks are not published for this historical result. |
02 — COMPARE
What do the decks look like?
Expanded published deck comparison
03 — HOW I TESTED
How I tested each deck.
What each AI tool was asked to do
Structured
Test case 1 · Freight modal split · Storyline suppliedInterpretive
Test case 2 · Road freight activity · Find the storyMulti-source
Test case 3 · Shipment operations · Reconcile evidenceStrategic
Test case 4 · AI and workforce transition · Build the argument
Case 1 · Freight modal split — full benchmark prompt
Create one editable eight-slide PowerPoint deck for the management board of a German industrial manufacturer that ships most of its finished goods by road.
Follow the storyline in source/briefing-notes.md exactly, in the order given. Each numbered section of the briefing notes becomes one slide, in sequence:
1. why the board is being asked to look at transport mode now;
2. how freight splits across modes in Germany today;
3. how that split has moved since 2008;
4. how Germany compares with the EU27 average;
5. what the rail share means for the company's decarbonisation commitments;
6. the operational barriers to shifting volume to rail;
7. a staged shift-to-rail approach;
8. the decision the board is asked to make.
Use only the supplied case files. Build the mode-share trend and the Germany-versus-EU27 comparison as native, editable PowerPoint charts driven by the values in source/freight-modal-split.csv. Every number that appears on a slide must be present in the supplied files; do not invent tonnages, costs, emissions figures, customer names, or measured outcomes. Keep facts, estimates, and recommendations visibly distinct. Put source provenance in the speaker notes of every slide that makes a data-backed claim, wrapped in a [Sources] ... [/Sources] block.
Design contract, to be followed exactly: white background; dark grey body text; one deep-blue accent colour used for emphasis and chart series; a single thin rule under each slide title; left-aligned text; body text no smaller than 14pt. Slide titles must be full sentences that state the takeaway, not topic labels. No gradients, no stock or generated imagery, no decorative icons, no pill-shaped badges, no repeated dashboard cards.
Save the required artifact exactly at output/deck.pptx, relative to your current working directory. Do not ask questions. Do not send a follow-up message. Do not use skills, plugins, custom benchmark instructions, or general web access.
Case 2 · Road freight activity — full benchmark prompt
You are preparing a point-of-view deck for the operations leadership of a logistics consultancy.
Analyse the monthly German truck toll mileage index in source/lkw-maut-monthly.csv and answer one question: what has happened to German road freight activity since 2008, and what should our industrial clients take from it?
The storyline is not given to you. Work out from the data itself which movements matter — the turning points, the sustained trends, the level the index returns to — and build the deck around what you find. Decide the structure yourself.
Read source/index-methodology.md before interpreting the series. The index measures activity, not tonnage or revenue, and the adjusted and unadjusted columns answer different questions; choose deliberately which one each chart uses and say which you used.
Constraints on claims. Every figure on a slide must be computed from the supplied CSV or stated in the methodology note. You may not attribute a movement in the index to any external event, policy, or economic condition as established fact, because no such information is supplied — where you offer an explanation, mark it clearly as interpretation and keep it visibly separate from what the data shows. Do not invent tonnages, freight rates, company names, client outcomes, or forecasts.
Deliverable: one editable PowerPoint deck of six to ten slides, saved exactly at output/deck.pptx relative to your current working directory. Use native, editable PowerPoint charts built from the supplied values — at least one chart of the full series. Slide titles should state the takeaway as a full sentence. Put source provenance in the speaker notes of every slide that makes a data-backed claim, wrapped in a [Sources] ... [/Sources] block.
Choose a professional design yourself: it should look like something a technology and logistics consultancy would put in front of a client. Do not ask questions. Do not send a follow-up message. Do not use skills, plugins, custom benchmark instructions, or general web access.
Case 3 · Shipment operations — full benchmark prompt
Our client sent over the attached files ahead of Monday's steering meeting. They want to know what is going on in their shipment operations and what to do about it.
Put together a deck of six to ten slides, saved exactly at output/deck.pptx relative to your current working directory.
Work only from the supplied files. Where they disagree with each other, say so on the slide rather than picking a number silently. Do not invent figures, customer names, service levels, or outcomes that are not in the files.
Design is up to you. It should look like something a technology and logistics consultancy would put in front of a client. Do not ask questions. Do not send a follow-up message. Do not use skills, plugins, custom benchmark instructions, or general web access.
Case 4 · AI and workforce transition — full benchmark prompt
Create one editable eight-slide PowerPoint deck for the CEO, CHRO, COO, and technology leadership of Meridian Signal. The theme is how AI is changing the modern workforce: not simply by removing jobs, but by redistributing tasks, decision rights, skills, and accountability.
Use source/workforce-transition-brief.md, source/org-structures.csv, and source/work-activity-shifts.csv as the complete evidence base. Treat the proposed future organization and activity mix as a planning scenario, not as a forecast or a measured outcome. Do not introduce external statistics, named companies, workforce reductions, productivity claims, or implementation results.
Build this eight-slide narrative:
1. the executive thesis on how AI changes work and organization design;
2. which activities are automated, augmented, or remain human-led, using a native editable chart built from source/work-activity-shifts.csv;
3. "Before AI" — a native editable organization chart of the current structure in source/org-structures.csv;
4. "After AI" — a separate native editable organization chart of the proposed AI-enabled structure in source/org-structures.csv;
5. how roles, skills, and management responsibilities change;
6. the governance model that keeps humans accountable for decisions and exceptions;
7. the 90-day transition plan in the source brief;
8. the executive decisions required now.
The two organization charts must be on separate slides, use native PowerPoint shapes and connectors, show the reporting hierarchy clearly, and remain fully editable. Label the slides with the exact phrases "Before AI" and "After AI". Make clear that the after-state is proposed and that the scenario keeps total workforce at 420 while reallocating 16 roles across functions; do not imply layoffs.
The supplied Meridian Signal design system is mandatory, not optional. Start from reference/Meridian-Signal-Slide-Design-Templates.pptx or reference/Meridian-Signal-Master.potx and preserve the embedded Meridian Signal master and layouts. Use reference/Meridian-Signal-Master.pptx, reference/Meridian-Signal-Design-System.md, and the PNGs under reference/master-preview/ to understand the intended layouts. Select and adapt appropriate slide-library patterns rather than drawing an unrelated theme. Remove all bracketed instructional placeholders and sample data from copied templates.
Follow the Meridian visual language throughout: warm cream and beige surfaces, selective charcoal section fields, graphite structure, Golden Gate orange emphasis, Arial typography, disciplined left alignment and grid spacing, hairline rules, two-part content-slide titles, and restrained native diagrams and charts. Avoid gradients, purple or blue AI styling, robot or sparkle imagery, repeated dashboard cards, pills, decorative icons, heavy shadows, and text below the 11pt floor.
Put source provenance in the speaker notes of every slide, wrapped in a [Sources] ... [/Sources] block. Cite the exact local files used on each slide; cite the Meridian reference file or design contract when a visual pattern is adapted.
Save the required artifact exactly at output/deck.pptx, relative to your current working directory. Do not ask questions. Do not send a follow-up message. Do not use skills, plugins, custom benchmark instructions, or general web access.
Every submission follows the same six-stage pipeline.
- Lock the task
Freeze the exact prompt, redistributable sources, requirements, and reference boundary.
- Generate one deck
Start a fresh local agent session and submit the task exactly once.
- Check validity and editability
Inspect OOXML integrity, required content, source notes, and editable structure.
- Render in PowerPoint
Export locally through Microsoft PowerPoint and preserve raw visual evidence.
- Blind-judge quality
Send a blinded, sanitized bundle to the release-pinned judge models.
- Rank eligible results
Rank every configuration with at least one eligible deck and incorporate all eligible repetitions.
Only Test Case 4 used Meridian Signal’s corporate template.
Test Case 4 was the only test case supplied with Meridian Signal’s corporate PowerPoint system. The reference library is split into the master layouts that define structure and the slide templates that show how to use them. Cases 1–3 had no corporate template, so their visual-design scores are not directly comparable with Case 4.
Structure
18 master layouts
Reusable PowerPoint layout definitions.
Application
49 slide templates
Example slides from the companion design library.



