Home/Blog
From the bench

How this work is actually done.

Our consultants use frameworks that most firms keep behind a subscription. These are those frameworks, written out in full: vendor grids, market sizing, evaluation scorecards, cost models and go-to-market strategy. No gates, no lead-capture forms.

Buyer's guide

The best forward deployed engineer services in 2026

Every major AI vendor now sells engineers who move in with the customer. Who each option suits, what they cost, and the one piece of work to finish before any of them start billing.

12 min read
Buyer's guide

The best AI vendor evaluation services in 2026

Selection, risk and compliance are three different jobs sold under one label, at prices two orders of magnitude apart. Which firm does which, and how to buy the one you actually need.

11 min read
Buyer's guide

The best Gartner alternatives in 2026

Nobody actually needs "a Gartner". They need the grids, the inquiry calls or the data, and each of the three has substitutes priced from free to a fraction of a seat.

11 min read
Buyer's guide

The best technology scoping and discovery services in 2026

Most scoping is sold by firms bidding to win the build it scopes. The market in 2026, what each tier costs, and how to keep the estimate honest.

10 min read
Buyer's guide

The best AI agent development companies in 2026

The most crowded label in software services, sorted: who is actually worth evaluating, who each firm suits, and the four questions that separate agent builders from rebrands.

11 min read
Analyst firms

What a Gartner subscription actually costs

Gartner publishes no rate card, so the question gets answered badly. Here is what is publicly reported, what the money buys, and when it is worth paying.

9 min read
Costing

How much market research actually costs

Research is priced by method, not by output, which is why two quotes for the same brief can differ by an order of magnitude.

9 min read
Costing

What does it cost to build an AI agent?

Most quotes price the build and ignore the running, which is why agent budgets are wrong in a consistent direction.

9 min read
Procurement

An AI vendor evaluation checklist

AI systems are probabilistic, they change when the vendor updates a model, and your data may travel further than you agreed. A general checklist scores none of that.

10 min read
AI delivery

Why AI pilots fail to reach production

The striking thing about failed AI pilots is not that they fail. It is that they usually work, and stop anyway, somewhere between the demo and Tuesday.

10 min read
AI integration

The forward-deployed engineer, and why AI integration needs one

The hard part of an AI project is never the model. It is everything around it, and that can only be learned from inside the customer's workflow.

12 min read
Go-to-market

A go-to-market framework for products that already exist

Most GTM advice is written for launches. This is for the more common situation: a real product, real users, and growth that has stopped explaining itself.

11 min read
Context engineering

context.md and llms.txt: teaching a coding agent your stack

Coding agents fail on unfamiliar frameworks for a boring reason: they are guessing from training data that predates your version. A context file fixes it.

9 min read
Scoping

What belongs in a technology scoping report

The document that turns "we should probably do something about this" into work a team can start on Monday, and a number a sponsor can approve.

10 min read
Costing

Costing an AI feature before you build it

Token pricing is the visible cost and rarely the largest one. A defensible estimate models the unit economics, the retry tail, and the humans who stay in the loop.

11 min read
Decision frameworks

Build vs. buy: a decision framework that accounts for the second year

Most build-versus-buy analyses compare a build estimate against a first-year licence fee. That comparison is wrong in both directions, and predictably so.

10 min read
Procurement

Building a vendor evaluation scorecard that survives scrutiny

Weighted criteria, anchored scales, and the four failure modes that quietly turn a scorecard into a rubber stamp for a decision already made.

11 min read
Market sizing

TAM, SAM and SOM without the hand-waving

Top-down and bottom-up market sizing, why serious analysts always run both, and how to present a number that a CFO cannot dismantle in thirty seconds.

10 min read
Research

How to write an analyst-grade market research report

The nine-part structure that credible market reports share, and the sourcing discipline that makes them survive a hostile read.

11 min read
Vendor evaluation

Magic Quadrant vs. Forrester Wave vs. IDC MarketScape: what each grid actually measures

Three analyst firms, three two-dimensional grids, three different questions. Knowing which is which decides whether your citation strengthens an argument or quietly undermines it.

12 min read
Rather have it done?

Brief a consultant instead of reading nine articles.

Every framework here is one the bench applies for real. Describe what you need and a specialist returns the finished artifact: report, plan, scorecard or deck.

Start a conversation →