Summary:
Claude Fable 5.1 builds well-structured slide decks from a prompt, a PDF report or an existing PowerPoint file, but none of the three decks we generated were client-ready by McKinsey or BCG standards.
The world of slide creation is changing fast with the introduction of AI slide generation tools. On September 1, 2026, Anthropic released Claude Fable 5.1, which it describes as its most advanced model for coding and knowledge work. AI platform for finance, Hebbia, said Fable 5.1 "produced the best decks of any model we've tested."
With all this high praise, we were of course curious to see if it makes better consulting slides than its predecessors, Fable 5 and Opus 5.
In this article, we test Fable 5.1 on three different tasks: creating a deck from a prompt, turning a text report into a deck, and improving an existing deck. We score every output against nine criteria for consulting-grade decks and explain where and why the outputs fall short. Finally, we share tips for getting better decks out of Fable 5.1.
If you'd like, you can see our head-to-head test of Claude Fable 5.1 vs. GPT-6 Astra for making consulting slides in our related article here. We also compare specialized consulting tools and general tools in our latest article here.
Before we start testing, we need to define what a "good" deck is. Our area of expertise is creating consulting-style decks like the ones top-tier firms such as McKinsey, BCG and Bain produce.
A consulting-grade deck states one clear message per slide, backs every message with evidence, and follows a storyline a client can read without a presenter. These decks are typically information-dense, heavy on charts and self-contained enough to be left behind as reports for clients.
To measure the output of Fable 5.1, we rated each deck against these nine questions:
We rated each criterion Yes, Partially or No.
We ran three tests in the Claude desktop app in September 2026, using Fable 5.1 on the Medium effort setting.
| Test | Input | Prompt | Output |
|---|---|---|---|
| A: Simple prompt | None | "Create a project kickoff deck for our new strategy project" | 14-slide deck |
| B: Report to deck | 37-page PDF report by Monitor Deloitte (2023) | "Turn this document into a deck" | 10-slide deck |
| C: Improve a deck | PowerPoint deck with a fictional business case | "Make this deck better" | 17-slide .pptx deck |
For the first test, we wanted to see what Fable 5.1 produces from a completely raw starting point. We used the following prompt:
Create a project kickoff deck for our new strategy project
The prompt we used to test how well Fable 5.1 creates slides from nothing but a simple prompt
Fable 5.1 started by asking clarifying questions: who the client is, who the audience for the deck is, how long it should be, and whether there are details on timeline, team or format. We asked it to keep the deck general with placeholders and to use the recommended length.
Fable 5.1 returned a 14-slide kickoff deck with a logical storyline and clean, balanced layouts. It had no action titles and far too little detail to pass as a consulting deck.
Score: 5 Yes, 0 Partially, 4 No.
The output from our first test follows a familiar structure but lacks detail and depth.
| Criterion | Rating | What we saw |
|---|---|---|
| Layout | Yes | The layouts are simple but useful for communicating the message and purpose of each slide. |
| Focus | Yes | Each slide only carries information relevant to its message. |
| Action titles | No | There are no action titles. |
| Tone | No | The slides are too light, with not enough detail to be consulting slides. |
| Visual balance | Yes | The deck is fairly balanced. |
| Storyline | Yes | The deck follows a logical storyline: purpose and aim of the project, then approach, timeline and deliverables, ending with an ask to the audience. |
| Coherence | Yes | Messages and numbers are consistent. |
| Client-readiness | No | The deck is not client-ready. |
| MBB look and feel | No | The slides look AI-generated and lack the detail, complexity, thoroughness and purpose of consulting slides. |
Second, we gave Fable 5.1 a 37-page PDF report, Electric Two-Wheeler Supply Chain Landscape in India, written by Monitor Deloitte for the Society of Manufacturers of Electric Vehicles (SMEV) in 2023. We asked it to simply turn the report into a deck.
Before building the deck, Fable 5.1 asked how comprehensive the deck should be and who it was intended for. We chose the recommended 10 slide option and an internal audience.
Like before, Fable 5.1 asks a few questions to better understand who the deck was intended for.
Fable 5.1 picked sensible layouts for list-type content, used charts for numerical content and kept the numbers consistent with the report. It followed the report page by page instead of building a storyline, added information that did not support each slide's message, and wrote descriptive rather than insight-driven text.
Score: 1 Yes, 3 Partially, 5 No.
The output from our second test of Fable 5.1
| Criterion | Rating | What we saw |
|---|---|---|
| Layout | Partially | For slides on types of challenges and policy actions, the model chose suitable pillars and number balls, and it used charts for slides with numbers. On many slides, however, it highlighted parts of the slide that the message did not refer to. |
| Focus | No | Slides often contain confusing additional information that doesn't support the main message. The reader has to study the slide in detail to understand how the information is relevant. |
| Action titles | Partially | Some slides have action titles; others don't. |
| Tone | No | The text is descriptive rather than taking a point of view, as consulting slides do. Some slides use labels like "So what", which never appear in real McKinsey or BCG deliverables. |
| Visual balance | Partially | Somewhat balanced, but heavy on boxes and number balls. |
| Storyline | No | The deck is a direct read of the report and lacks a coherent storyline. |
| Coherence | Yes | Messages and numbers are consistent. |
| Client-readiness | No | Not really client-ready. |
| MBB look and feel | No | The slides are too simplistic and bluntly structured. They lack small details like call-outs and chart annotations, and bigger things like a compelling storyline and clean message slides. |
Below are some specific examples showing the patterns:
A closer look at two of the slides from Fable 5.1 that show how the information on the slide is not immediately related to the main message.
For our last test, we gave Fable 5.1 a PowerPoint deck with a made-up business case and asked it to make the deck better.
As in the other tests, the model asked clarifying questions: which content the improved deck should carry, and how far the redesign should go. It recommended keeping the example business case and doing a full redesign as a new .pptx file.
Fable 5.1 kept a rough consulting structure, consistent numbers and only relevant content on each slide. The slides stayed sparse, the action titles improved only partly, and the text did not read like a consulting deck.
Score: 3 Yes, 3 Partially, 3 No.
| Criterion | Rating | What we saw |
|---|---|---|
| Layout | Partially | The financial slides use fitting layouts, such as KPI tiles, cash flow charts and a sensitivity chart, but many other slides fall back on equal-width boxes of bullets that list content rather than show the message. |
| Focus | Yes | Slides only contain relevant information, but the information is sparse. |
| Action titles | Partially | It has partial action titles, but they could be much better. |
| Tone | No | The text is not consulting-like at all. |
| Visual balance | Partially | The deck is clean and consistent, but text-light box slides leave large empty areas, while the cost model and savings slides pack small text into dense tables. |
| Storyline | Yes | The deck follows a consulting structure in rough terms. |
| Coherence | Yes | Messages and numbers are consistent (the numbers are dummy numbers). |
| Client-readiness | No | The deck is not client-ready. |
| MBB look and feel | No | The slides do not look like McKinsey or BCG slides. |
Fable 5.1 is a capable slide tool that creates excellent first drafts of decks, but it is still far from actual consulting slides.
Here, we sum up the results:
| Criterion | A: Prompt | B: Report | C: Improve deck |
|---|---|---|---|
| Layout | Yes | Partially | Partially |
| Focus | Yes | No | Yes |
| Action titles | No | Partially | Partially |
| Tone | No | No | No |
| Visual balance | Yes | Partially | Partially |
| Storyline | Yes | No | Yes |
| Coherence | Yes | Yes | Yes |
| Client-readiness | No | No | No |
| MBB look and feel | No | No | No |
| Total (Yes / Partially / No) | 5 / 0 / 4 | 1 / 3 / 5 | 3 / 3 / 3 |
Three patterns stand out from our Fable 5.1 test:
Fable 5.1 is far better than previous models and with a few tweaks the output starts to come closer to real consulting decks. To get the results from okay to good, here are our best tips.
Fable 5.1 defaults to light, generic slides. If you explictly ask it to create decks that have the level of detail and complexity you'd expect from a McKinsey or BCG deck, it immediately produces better results.
The output is much improved just by asking the model to make the slides the same level of detail and complexity as consulting slides.
Fable 5.1 imitates the density and structure of what you show it. Give it some good references that have visual diversity and match the level of detail that you'd like to get a better outcome.
Upload your company's PowerPoint template and ask Fable 5.1 to build on that slide master. The output then uses your fonts, colors and layouts, and you spend much less time reformatting.
Claude Fable 5.1 is a step change from previous models. It builds decks with a logical structure, consistent numbers and sensible and visual diverse layouts, and it asks the right questions before it starts.
That being said, it still needs massaging and prodding before the output is client-ready. Across our three tests, it produced no client-ready deck, no consistent action titles and no consulting-grade text. Treat Fable 5.1 as a fast first draft: plan to rewrite the action titles, tighten the storyline and add the call-outs and annotations that make a slide consulting-grade.
Yes. In our tests, Claude Fable 5.1 built complete decks from a one-line prompt, a 37-page PDF report and an existing PowerPoint file, and it delivered the improved deck as a new .pptx file.
No. None of the three decks we generated were client-ready. Each needed stronger action titles, more insight-driven text and more detail before it could go to a client.
In our testing, Fable 5.1 produces better decks than earlier Claude models, but it still falls short of consulting standards on action titles, text tone and storyline.
Ask explicitly for consulting-level detail, upload good slides as references, and have it build on your own slide master.
Real consulting slides from houses like McKinsey or Bain are denser, more visually diverse, more clearly tied to key messages, more structured, and with more details like chart annotations than the slides produced by Fable 5.1.
The best AI tools out there right now to produce consulting-level slides or professionals slides in general are Claude Fable 5.1 and Opus 5.5. These two models can get very far with good prompts, references, and clean up afterwards.
