A ClearPoint Strategy alternative you can evaluate before the sales call

ClearPoint does reporting better than we do. Automatic report generation and distribution, data pulled in from multiple systems, scorecards and dashboards at volume — that is a mature discipline in their product and a thin one in ours.

And we will not claim to be the only tool here that mentions MCP: ClearPoint's own pricing page lists "API and MCP access" as something you can include, and describes connecting to existing on-premise and cloud tools through API, MCP and data-loader services. Anyone telling you the incumbents ignore agents entirely is not reading their pages.

The distinction is what a buyer can do with that today. Ours is public: an npm package, a GitHub repository under MIT, an entry in the official Model Context Protocol registry, a documented tool contract, and a read-only demo token published in our llms.txt so your assistant can connect to a real matrix in minutes. Theirs, as documented publicly in July 2026, is a scoped option discussed on a pricing call.

ClearPoint's pricing page is built around choices: user bands from 10-25 up to 500+, support from email-only to a dedicated manager, integrations from none to custom API access, SSO from basic to SAML 2.0, scorecards from a handful to 500+, training from self-guided to full-service setup. That is a defensible enterprise approach — you buy the shape that matches your organization.

It also means you cannot know what you will pay, or precisely what you will get, without a conversation. For a 40-person company evaluating on a Thursday afternoon, that is the friction that ends the evaluation.

We publish a free tier and a public demo. You will run out of our product's depth long before you run out of things you can check on your own.

This is the philosophical split, and it decides which tool you want. ClearPoint's centre of gravity is the report: assemble the data, format it, distribute it, on schedule.

Ours is the judgement. The X-Matrix is scored out of 100 on five metrics — coverage, correlation density, orphan elements, complete vision-to-KPI chains, and the balance of leading versus lagging indicators — and the score names what is wrong. This initiative is connected to nothing. This vision never reaches a measurement. This objective is tracked only by lagging indicators, so you will find out too late.

A beautiful report of an incoherent strategy is still an incoherent strategy, printed. If your problem is producing the monthly pack, buy the reporting tool. If your problem is that nobody notices the strategy has drifted until the quarter is gone, buy the thing that tells you, weekly, unprompted, through the assistant you already talk to.

• The scoring specification is public and the engine is MIT-licensed — you can audit how your strategy is judged.
• Exposed as an MCP tool, so the verdict lands in your assistant's answer rather than in a PDF.
• Following the score's own suggestions measurably raises it — we regression-test that, because a score that cannot be acted on is decoration.

A meaningful share of ClearPoint's user base sits in government and public administration, where balanced-scorecard reporting against published objectives is often a statutory obligation rather than a management choice. If that is your context, ClearPoint's report-generation depth and its long institutional track record are the right criteria, and we are not a serious contender.

We fit the other shape: a leadership team of 3 to 15 people who steer with a small number of objectives, want the review preparation to happen by itself, and are comfortable being driven through an AI assistant.

You cannot configure your way to "the data never leaves the building". hoshin.app has a sovereign local mode: the same 14 tools and the same score, running against plain JSON files on your own machine. No account, no network, no telemetry, no vendor to trust — the code is MIT and auditable before you run it.

That mode exists because a real prospect — a governance and compliance partner at a consulting firm — refused hosted-by-default on architectural principle, treating client knowledge and method as trade secrets. For that buyer, no amount of SOC 2, encryption or regional hosting is responsive to the objection. Files on their own disk are.