Seven workflows, from document to checked output
Each Analect workflow answers one question and leaves a record somebody else can check. They sit beside your AI, not in the path of it: nothing here handles a live request or changes a model.
They fall under three questions: what your AI is fed, what it costs, and whether what it says is supported by its source. An account is given the workflows its work needs and sees only those.
What your AI is fed
What is in the documents your AI reads, who may see them, and does sending a model less still get the same answers?
What is actually inside this document, and how do its parts connect?
Reads a document once and shows what it asserts: the things it names, how they relate, and the sentence each statement came from.
On the platform, with access set up by ETTAbout Analect Explorer Analect GovernanceWho may open each document, how sensitive is it, and who has looked?
Gives every stored document a sensitivity label worked out from what it contains, and uses it to decide who may open the document, what is withheld from whom, and what is logged. It governs who may read stored material. It does not yet stand in front of a model call.
On the platform, with access set up by ETTAbout Analect Governance Analect StoreWhat does each form of this material cost to keep?
Weighs your material in actual bytes in each form it could be kept in, the original file, its text and the structured forms built from it, and prices each across the kinds of storage you might use. The original is kept.
On the platform, with access set up by ETTAbout Analect Store Analect ProofDoes sending the model less still get the same answers?
Tests, on your own documents, whether a model answering from retrieved passages does as well as the same model reading the whole document, and counts the input each route used. Its verdict is one of three: holds, does not hold, or not shown.
On the platform, with access set up by ETTAbout Analect ProofWhat it costs
Which model fits each job, what will running it cost, and what will the platform under it cost?
Which models fit this work, and what will running it cost?
Prices every AI workload across the model catalogue, from your bills or a plain description, with the reasoning behind each pick and the catalogue's date on every figure.
Part of the AI Spend ForecastAbout Analect Forecast Analect CompareWhat will a Databricks lakehouse cost over five years, with Snowflake kept or pulled out?
Models what a Databricks lakehouse costs over sixty months, with Snowflake kept beside it or consolidated into it. Every price comes off a dated rate card with its source, every input is a range, and it makes no recommendation.
On the platform, with access set up by ETTAbout Analect CompareWhether what it says is supported
Does the source support what the output says, what did the output leave out, and can anyone check that we said so?
What goes in, and what comes out
| Workflow | The question it answers | You hand it | You get back |
|---|---|---|---|
| Analect Explorer | What is actually inside this document, and how do its parts connect? | Documents as PDF with a text layer, Word, plain text or Markdown. | What each document asserts, as small structured statements. |
| Analect Governance | Who may open each document, how sensitive is it, and who has looked? | The documents your account has stored through the other workflows. | A sensitivity label for every document: public, internal, confidential or restricted, with the counts it rests on shown. |
| Analect Store | What does each form of this material cost to keep? | A document you have stored on the platform. | The size in bytes of each form: the original file, its text, and the structured forms built from it. |
| Analect Proof | Does sending the model less still get the same answers? | A document, uploaded or from your library. | A verdict for each model: holds, does not hold, or not shown, with the interval it rests on. |
| Analect Forecast | Which models fit this work, and what will running it cost? | Usage or billing exports from your AI providers, read in your browser and never uploaded. | A model for each workload, with the reason it won and the models it beat. |
| Analect Compare | What will a Databricks lakehouse cost over five years, with Snowflake kept or pulled out? | Your workload assumptions, each as a low, a base and a high figure. | Two scenarios side by side, month by month, over sixty months. |
| Analect Evidence File | Which claims in this AI output does the source support, and which does it contradict? | An AI output and the source it was written from, as pasted text, PDF with a text layer or Word. | A Certificate of Accuracy for each output: every claim, its verdict and the passage the verdict rests on. |
See it on worked material, then on yours
Thirty minutes with the team that runs Analect: the AI Spend Forecast and the Accuracy Evidence File on worked examples, and what either would look like on your bills or your outputs.