Guides
New ideas to help you build the future.
Plain-language field guides from Magma on the ideas worth understanding before they become obvious, from AI and strategy to the craft of shipping. Written to be useful first, and to show the thinking behind the work.
Sample Reports: One Method, Three Readers
Three free sample reports, for a pricing lead, a strategy lead and a bank's AI governance lead. Public sources only, every source graded, and every call signed by a named person.
Understanding the Forward-Deployed Engineering Model
A buyer's field guide to the delivery model behind the 2026 AI wave: a senior engineer embedded with the business side, turning intent that is not yet a specification into running systems. Where it came from (Palantir's Deltas), how engagements are really structured, the objections taken seriously, and eight questions to vet any vendor.
OSFI E-23 Readiness for AI Models
E-23 takes effect on 1 May 2027 and covers AI and third-party models. The Review gives each use case one record, inside the model-risk review you already run, with a written triage of the questions that decide its route. See a sample result first.
Sample: the governance record of a live AI assistant
The intake behind the E-23 Review, run on a live public AI assistant by the person who built it, where no regulator required it: a proposed conditional yes, three open conditions with owners, two dated rounds, and the 8-page reviewer pack.
AI Governance: The Three Questions Your Board Will Ask
Could AI leak our data? Can we prove to a regulator that we are in control? Do we even know what is running? All three start with the same record. A field guide for regulated organizations, with Canada's May 2027 deadline and how the US, UK, EU and Singapore ask the same things.
How Do You Know Your AI Answers Are Good?
A practical guide on how to test AI search for your website, or an AI chat, and make its answers better in a way that lasts: put real answers side by side, rank them, and write down why.
What Is Strategic Foresight?
A field guide to seeing structural change early enough to do something about it: weak signals, why your dashboards find out too late, and where AI fits (and where it does not). Grounded in Ansoff and the joint OECD and World Economic Forum research.
Sample: AI Use Case Triage highlight report
What you keep after a 90-minute AI Use Case Triage, shown on a fictional company: eight AI ideas graded PROCEED, CAVEAT or REFUSE, the one to build first, the two to refuse and why, and a map of where AI would touch sensitive data.