# Inside Playbooks: A Tutorial on the 217 Pre-Engineered Research Workflows Built for Institutional Investors

> A tutorial on Playbooks: 217 pre-engineered research workflows across 12 categories, from morning briefs to sector-specific KPI math. The prompt does the thinking before you arrive.


```tldr
- **217 pre-engineered prompts** with structured inputs and visible logic
- Organized into **12 phases of research**, from morning briefs to sector-specific KPI math
- The prompt is already smart by the time you arrive; you're not prompting, you're configuring
- The point is not that the AI is smart; the point is that the workflow is
- A guided tour through how the system is built and where the leverage actually lives
```

## Sometimes the blank prompt is a problem

Most analysts know that AI is transforming industries and changing the way people work and live. And yet, many who have tried an AI research assistant have had mixed results. At times, they stare at a blank box not really knowing what to ask and how to ask it. When they prompt something like "find me five interesting names in semis" what comes back is almost never useful and rather disappointing. Nothing you can't find with a simple Bloomberg screen. They might be tempted to think that the technology is not yet there for real-life investment research.

In reality, the tech is fine. The prompt is the problem.

An investment analyst does not have the time and patience to think and interact with AI like a seasoned prompt engineer. They vaguely know what they want but are not ready to expend the energy to provide all the context the model needs to really nail that reply.

Fund managers and analysts who are using AI successfully usually lean on prompts that have already been engineered for the job. This skips a big step, takes the onus off the analyst, and yet still gives pristine results.

Analysts don't have time to prompt engineer. That's why we built Playbooks.

That is what Playbooks are. Pre-engineered prompts, with structured inputs and customizable context, organized by the phase of research you are in. We have shipped 217 of them across 12 categories, and they are shortening the time to value with AI. This post is a guided tour through how the system is built and how to use it for real work.

## The shape of the system

Open the Playbooks tab in the matterfact terminal and you will see the entire workflow of an institutional analyst broken into twelve numbered phases:

```playbook-grid
```

The first time you open Playbooks should feel familiar, as these workflows mirror an analyst's day. It is the same sequence you would run through if you had infinite time in the day. All these "jobs" are organized in buckets and when you click one, it acts like a filter, focusing on related workflows that would normally take hours of research and data gathering.

Playbooks collapse the time and effort into a few clicks.

```youtube
url: https://youtu.be/1cqTDTodd8A
title: A quick tour of matterfact Playbooks
caption: A quick walkthrough of the Playbooks library inside the matterfact terminal.
```

## The anatomy of a Playbook

Click any card and you get a modal with three things: a short description, a set of structured input fields, and a Prompt Preview that shows you exactly what gets sent to the model. The Prompt Preview teaches you how to correctly ask AI for what you need in real time.

```playbook-example
```

Take Morning Brief, the first card under Daily Briefs. The inputs are simple: your watchlist, an optional focus area, and a delivery schedule. The Prompt Preview shows you what the system will actually do with those inputs and the context it needs to do a great job:

> Generate my morning brief for today. Watchlist: AAPL, NVDA, MSFT, GOOGL, META, TSM, ASML, AMD. Focus: AI capex datapoints and Fed signaling. Cover, in order, with bullet density: (1) overnight equity / rate / FX / commodity moves and what drove them, (2) major news and filings on the watchlist names since yesterday's close, (3) today's economic calendar (data releases, Fed speakers) and earnings calendar with consensus expectations, (4) podcast and street chatter from the last 24 hours relevant to my watchlist, (5) ranked top 3 things to watch in today's session and the implied trade. Schedule sending me a daily report on this to my email at: daily at 6:30 AM ET.

The five sections are ordered by what you most likely need before market open. It also gives you a flexible delivery schedule that turns market intelligence into a repeatable process.

You can override or tweak any of it, or you can just hit Run. The point is that someone who knows what a good morning brief looks like has already done the structural work. You are not prompting anymore, just configuring.

## Deep dives by coverage: Sector Frameworks

The Sector Frameworks category has 53 Playbooks, which is a quarter of the entire library.

Sector-specific math is where most generic AI tools fail and where analyst time gets spent. A general-purpose model asked to analyze a hotel REIT will give you a polite essay about strong brand and demand recovery. A Playbook called Hotels RevPAR will give you ADR × occupancy by tier, group / transient / business mix, and fee economics, because the prompt was written by someone who has actually modeled a hotel.

## Operationalizing podcast intelligence

We have [written before](https://www.matterfact.com/blog/why-podcasts-why-now) about why podcasts are now a primary research input. The Playbooks under Podcast Insights are how that insight becomes a workflow.

Topic Expert Consensus is the one I would point a new analyst to first. You give it a topic, say, "AI infrastructure capex cycle and hyperscaler ROI" and an optional time window. The Prompt Preview shows what a great prompt looks like:

> Sweep podcast transcripts for expert commentary on: AI infrastructure capex cycle and hyperscaler ROI. Time window: last 90 days. Generate 4-6 diverse search angles to maximize recall (topic synonyms, adjacent terms, specific company / person names tied to the topic). For each distinct expert who has spoken substantively, capture their view, the underlying argument, and the timestamp / episode. Then synthesize: (1) the consensus view, (2) dissenting / contrarian takes, (3) variant perception (where the buy-side debate hasn't caught up to the experts), and (4) the strongest single argument on each side. Cite episodes by name and date. Schedule sending me a weekly expert-consensus digest on this topic to my email at: weekly Sunday at 8:00 AM ET.

First, the recall step of generating diverse search angles before retrieving addresses the "feature" of the model to give you what matches your established assumptions. Second, the synthesis step explicitly separates consensus from variant perception, which is where the alpha actually lives. Third, citations to specific episodes mean you can verify or escalate to a longer listen on anything that matters.

Run it on Sunday morning, get a digest in your inbox by 8:00 AM ET, walk into Monday already several hundred episodes ahead.

There are several other categories modeled after real investment and finance workflows and engineered to both save time and improve the results you get.

The point of Playbooks is that the mental load to structure the perfect prompt is done before you arrive. What is left is the thing only you can do in seconds: read, decide, and run it.

```artifact
url: https://app.matterfact.com/playbooks
eyebrow: Inside matterfact
heading: Open the Playbooks library.
description: Browse all 217 pre-engineered research workflows across 12 categories, from morning briefs to sector-specific KPI math.
buttonText: Open Playbooks
```

Analysts love Playbooks, and if you haven't tried them yet, we are sure you will love them too. Find the perfect Playbook for you.

```request-access
heading: See Playbooks on your own coverage.
description: matterfact is deployed with select institutional partners. Request access to run the full Playbook library against the names and sectors you cover.
buttonText: Request access
```