Skip to content
CRAFTDEXA

Theme guide: STRATEGY

Technology decisions that compound.

Strategy is deciding what not to build, what to buy and what to keep flexible. Good decisions compound, because each one makes the next cheaper. Bad ones are paid for again in every project that follows.

What is moving now

Jun 2025

Gartner expects many agentic AI projects to be cancelled

Gartner predicted that more than 40% of agentic AI projects will be cancelled by the end of 2027, citing rising costs, unclear business value and inadequate risk controls, and warned of vendors relabelling older products as agentic. Its advice is to pursue agentic AI where it delivers clear value.

SourceGartner press release (opens in a new tab)

Apr 2026

Technology Radar: a return to engineering fundamentals

Volume 34 of the Thoughtworks Technology Radar (15 April 2026) argued that as AI speeds up the writing of code, the hard part moves to technique: managing cognitive debt, evaluating tools in an agentic world, and keeping to fundamentals such as sandboxing and defence in depth.

SourceThoughtworks Radar vol. 34 release (opens in a new tab)

Dec 2025

Open standards lower the cost of changing your mind

With the Model Context Protocol now under the Linux Foundation's Agentic AI Foundation, one way of connecting agents to tools is governed in the open. For a technology decision, that is a reason to build on the standard rather than around one vendor's interface.

SourceAnthropic announcement (opens in a new tab)

Summarised from public sources as of November 2026. Dates and details change, so follow the link to the source before relying on one.

A practical checklist

  1. 01

    Write the decision and the alternatives

    One page: the options, the choice, why, and what would make you revisit it.

  2. 02

    Separate reversible from irreversible

    Move quickly on the first kind. Slow down on data models, vendors and public APIs.

  3. 03

    Price the whole life, not the build

    Running, securing, upgrading and eventually replacing a system is most of its cost.

  4. 04

    Buy the commodity, build what differentiates

    Be honest about which is which, and look again once a year.

  5. 05

    Name an owner and a review date

    Every platform choice gets a person responsible and a date to ask whether it still earns its place.

  6. 06

    Prove it small first

    A thin slice in production teaches more than a long plan.

Where it shows up in our work

Related capabilitiesTechnical Strategy ·System Architecture ·Cloud Infrastructure ·Code Reviews ·Engineering Workflows

Working on something in strategy? Tell us what needs to exist.