Guides by Job

Iternal Guides, Indexed by the
Job You Are Doing

Ten jobs buyers arrive with, and the guides that settle each decision inside them. Pick the job, not the product.

Built from real buyer questions in our sales meetings

These guides are organized by the job a buyer is doing, not by product and not by industry. Ten jobs run the list below, each one a decision somebody is already in the middle of when they start searching: keeping regulated material inside a boundary, getting messy data into a shape that answers correctly, funding the work, standing it up on hardware you already own, getting people to use it. Every guide underneath a job settles one question inside it and hands the rest to the guide that owns them, so the fastest route in is the job that matches the sentence you would use to describe your own situation.

The Ten Jobs, and What Sits Under Each One

AI for Confidential Data That Cannot Go to the Cloud

Whether the material may leave the boundary that already governs it is settled by the classification, the contract and the regulator that apply to it, not by a product choice. Four questions follow, each answered on its own page.
Getting Messy Enterprise Data Ready for AI: The Sequence

You cannot fix messy data at retrieval time, so the answer is a sequence rather than a purchase: inventory what you hold, de-duplicate it, restructure it, govern it, and only then retrieve against it.
Where to Start With AI When Nobody in the Building Knows

Start from a business process that is still done by hand and has a measurable cost, not from a technology or a model, and prove it on your own data before you buy anything at scale.
How to Build an AI ROI Business Case Finance Believes

Baseline the current cost of the process in hours and error rate, state the post-AI cost the same way, and show the delta net of license, hardware and the human review that remains.
Run AI on the Hardware You Already Own: Three Routes

Almost every deployment is judged on one question before any other: can this run on the machines we already have, and what do we do if it cannot.
Supported LLM Models and Bring Your Own: How to Choose

Choose the smallest model that clears your accuracy bar on your own documents rather than the largest one your machine will technically load, and keep the option to bring your own.
AI Literacy Across the Workforce: Which Barrier Is Yours?

Four different problems arrive described as “our people are not using AI”. They have four different owners, and the first job is to say which one you actually have.
Manual Business Process Automation: What Automates First

The processes that automate cleanly share one shape: a document goes in, a person applies a known rule or template, and a document comes out.
How Private AI Differs From ChatGPT and Cloud Assistants

Three differences decide it: where the inference happens, what the answer is grounded in, and how the cost behaves over the life of the deployment.
Build an AI Practice: Where a Partner Starts Investing

Start from a captive, tangible outcome you can deliver repeatedly rather than an open-ended proposition — a concrete deliverable is far easier for sellers to take to market than a general capability.

Elsewhere in these guides

Best AI Tools by Job — shortlists that compare named options for one specific job, with the limit on each one stated.
Sector Guides by Industry — the same jobs seen from inside one sector, in the words that sector used.
Book a demo — for the questions a guide cannot settle, run it against your own documents.

Iternal Guides, Indexed by the Job You Are Doing

The Ten Jobs, and What Sits Under Each One

Iternal Guides, Indexed by the
Job You Are Doing