How we read a framework

1. What this library is

Every major AI playbook is built to answer a particular question. This library reads each one through the same five pressures, the Five Forces of Enterprise AI Strategy, so you can see what it covers in depth, what it treats lightly, and what it leaves for you.

This is a reference library, not a ranking. We do not rank frameworks against each other. We map each one against the same five forces, then leave the choice with you.

The useful question is not “which playbook is best.” It is “what was this playbook commissioned to answer, and what does that leave open in my organisation?”

Browse the Strategy Library

What is the Five Forces of Enterprise AI?

2. The five forces we read for

The model has one internal force and four external forces. Evidence has to match the force’s scope.

Internal pressures and capabilities. What the enterprise can change inside its own walls: operating model, talent, technology, data, adoption, and the redesign of work among its own people.

Global and geopolitical dynamics. Pressures the enterprise does not control: export controls, sovereignty, cross-border data, national AI strategy, great-power competition.

Societal and regulatory pressures. Law, policy, ethics expectations, and institutional accountability. Depth here means building regulation into strategy, not explaining specific laws.

Industry and competitive pressures. Market structure, rivals, vendors, and industry adoption patterns.

Human needs and aspirations. External human pressure only: customers, the talent market, public sentiment, and the demand for work that still means something. Internal reskilling and role redesign belong under Internal pressures. They are a response to this force, not this force itself.

3. Two signals for every force

For each force we record two independent signals.

Coverage (0-3). The formal name of this scale is Coverage Depth: how far the framework develops this force.

• 0 Not addressed
• 1 Acknowledged
• 2 Developed
• 3 Deep treatment

Actionability (0-2). How far it tells you what to do.

• 0 No practical guidance
• 1 Directional
• 2 Implementable

High coverage with low actionability means the force is taken seriously as analysis, not as a recipe. That is useful information. It is not a demerit.

We rate coverage, not quality. Silence on a force is a design choice. It becomes a problem only if that playbook is the only lens you have.

Strategy frameworks also carry a total out of 25, the sum of their ten signals. Companion references carry no total. A total measures coverage through this lens, not quality.

4. Strategy frameworks and companion references

Each framework sits in one of two sections, decided by scope: a strategy framework addresses the enterprise AI strategy as a whole; a companion reference deepens one discipline within it, such as governance or risk.

The coverage map is a separate question. It shows how far a framework covers each of the Five Forces, and it says nothing about quality or completeness as a strategy. That is why a companion reference can cover more of the map than a strategy framework. Governance guides deal with regulation, stakeholders, public trust and the supply chain, which are outward-looking forces, while most strategy frameworks concentrate on the inside of the organisation. The two are designed to be used together.

5. What we read, and what we leave out

We read the framework’s own published materials.

• Core framework. Documents, tools, and papers the publisher positions as part of the named playbook. These are the only sources that can change a signal.
• Adjacent work by the same publisher. Relevant, but not part of the named playbook. We report it so you can see what else they offer. It does not change the signals.
• Third-party commentary. May corroborate a claim the publisher already made. It is never the sole basis for a signal.

Only a published framework document qualifies for the library. Service pages and marketing pages are not frameworks, and we do not assess them.

We prefer sources published or updated in the last 24 months. Older material is used only when the publisher still stands behind it. Archived, deprecated, or missing pages are not used.

Every assessment links to the original framework and lists its sources. We describe diagrams; we do not reproduce them.

6. Three questions the signals do not capture

Some choices sit at the centre of an AI strategy and are easy to leave implicit. We record them separately, as Explicit, Implicit, or Silent. They are not part of the coverage total.

• Time horizon. Does the playbook ask you to declare a planning window, or treat every recommendation as horizon-agnostic?
• Automation or augmentation. Does it treat replacing work and amplifying people as different orientations, with different risks?
• Ideas and execution. Does it notice that as execution gets cheaper, the relative value of judgement and ideas rises?

7. How a framework gets into this library

1. We check that the publisher has a published framework document. A service or marketing page is not enough.
2. We identify the publisher’s canonical page for the named framework.
3. We gather the materials that page treats as part of the framework, and we date them.
4. We check every link we cite, including all links on the framework’s main page. Pages we cannot open automatically are confirmed by hand. Dead or outdated sources are excluded.
5. We place the framework in a section by its scope: strategy framework or companion reference.
6. We rate each force for Coverage and Actionability against the scales above.
7. We write the narrative as a coverage map: where the playbook is strong, and which dimensions it leaves open.
8. We date the assessment. When the publisher updates the source, we re-read it.

Assessments are published on fiveforces.ai. They are not gated.

8. How to read a low number

A zero is not an indictment. Most playbooks were commissioned to answer “how do we execute.” They answer it well. The geopolitical, regulatory, or human-pressure question is then one you are expected to bring yourself.

If a playbook leaves a force open, look for the complement: another playbook that treats that force in depth. Complements extend. They do not replace, and they do not outrank.

9. Independence

Sapience provides advisory services that sit alongside, and sometimes in contrast to, the work assessed here. The organisations named have not reviewed or endorsed these assessments. We apply the same rubric to every entry.

This library is our opinion, based on each organisation’s publicly published framework as at the date on that assessment, linked from the assessment page. It is offered to help you see a framework’s coverage and its open dimensions, not as a comparison, ranking, or endorsement.

10. Corrections

If you believe we have a fact wrong (a date, a citation, a claim about what a document says), tell us through the contact form.

We will review promptly. Where the record is wrong, we will correct it with a visible note and a date on the assessment page.

Disagreement with a coverage judgement is welcome; it is not, by itself, a factual error. The scales and this page are the basis of those judgements.

11. Close

The library shows you what a playbook covers. The next question is whether your own strategy inherited the same shape.