Buying AI

An owner at a desk in a small office on a weekday morning with two laptops open side by side showing two different software dashboards, a printed twelve-week project plan with handwritten dates pinned to a noticeboard behind her

Switching AI vendors without burning everything down

Switching AI vendors is harder than switching traditional SaaS because the data, integrations, prompts and staff habits are all tool-specific. Eight to twelve weeks run as a project, not a one-week migration.

An owner at her office desk in the early afternoon, a printed contract open in front of her with passages highlighted, a notepad to her right showing a handwritten dated timeline, leaning forward and reading carefully, a closed

When an AI engagement goes wrong, the escalation playbook

AI engagements go wrong predictably. The recovery paths for an owner-operator are narrower than enterprise, but they exist and they follow a sequence.

An owner sitting at a kitchen table on a Sunday afternoon, a printed vendor contract open in front of her with a yellow highlighter on the page and a one-page handwritten exit-readiness note on a separate sheet to her right, a

Exit clauses and switching costs, planning departure on day one

The most expensive part of an AI vendor relationship that goes wrong is usually the exit, not the engagement. Three contract-stage questions prevent most of it.

An owner sitting at a kitchen table in early evening, a printed vendor contract open in front of her with a highlighter in her hand and sticky tabs on four pages, a notepad with handwritten questions to her left, leaning over the

Reviewing an AI vendor contract, the things owner-operators get wrong

Owners commonly review an AI vendor contract for the wrong things. The legal language is fairly standard. The new questions sit in operational and commercial provisions a non-lawyer can read.

An owner at her kitchen table in early evening, working through a handwritten one-page cost estimate with five cost lines and two columns of figures, a vendor proposal set to one side and a calculator at the edge of the page, pen

Estimating the total cost of AI ownership before you sign

Headline AI vendor pricing is typically 40 to 60 per cent of the actual first-year cost. A one-page TCO across five layers, done before signing, makes the decision cleaner.

An owner sitting at her office desk in late morning, three printed proposal documents fanned in front of her and a notepad with handwritten questions to her left, pen in hand and leaning forward to compare them, a coffee mug and

Buying AI for owner-operated businesses

Most AI buying advice is written for enterprise procurement teams. This is the proportionate version for a 5 to 50 person owner-operated business making the same decisions with none of the same resources.

Owner-operator at a desk reviewing a printed page of vendor notes next to an open laptop, daylight through a side window.

Managing an AI vendor as an owner-operator

A single page of standing notes per vendor, plus a thirty-minute quarterly review, covers what many owner-operators silently lose to poor vendor discipline.

A business owner at a desk reading a printed contract with a pen in hand, laptop to one side.

Data and IP clauses in AI contracts, in plain English

Two clause categories AI contracts treat differently from traditional SaaS, what they actually say, and the six questions an owner should be able to answer from the text.

An owner at her office desk in late afternoon with a phone to her ear, taking handwritten notes on a printed list of reference questions, three vendor contact names listed in a notebook beside her, a closed laptop and a mug at the edge of the desk

Reference checks for AI vendors, what to ask and how to verify

Default reference checks fail because the vendor picks the references and the references say nice things. Five peer-to-peer questions and four verification routes produce real information.

A business owner and a vendor representative sitting across a table with a laptop open between them, the owner pointing at the screen during a software demo

Six questions to ask in an AI demo before you sign anything

AI demos are inherently staged. Six vendor-neutral questions narrow the gap between what you see in the room and what you get in production.

A woman taking notes in a small notebook across a meeting-room table from a man with an open laptop, mid-conversation, daylight from a window.

Reading an AI sales pitch and separating evidence from marketing

AI vendor pitches mix three layers: marketing language, verifiable claims, and unverifiable ones. Four questions and a clean closing ask separate them.

A small firm owner at a kitchen table with a printed list of AI tools and an open notebook split into free and paid columns, a mug of tea beside her

When the free AI tier is enough, and when you actually need to pay

Free AI tiers have got genuinely good. The decision to pay still has four real triggers, and many owner-led firms cross one within three months without noticing.

A small firm owner at a home desk comparing three printed invoices with a notebook and calculator in front of her, an evening lamp lit and a mug of tea beside her

Why your AI bill changes month to month, and how to budget for it

AI tool bills move in ways traditional software bills do not. The volatility is structural, and a quarterly contingency budget handles it better than an annual forecast.

A woman in her forties sitting at a kitchen table comparing two printed software quotes, with a notebook and a mug of coffee beside her

Choosing between per-seat and usage-based AI pricing

The two dominant AI pricing models in 2026 each look cheap until they don't. Here's how to match the model to your team and avoid doubling your bill.

A small business owner at an office desk comparing two printed proposals, a laptop with a calculator open beside them, holding a pen

When to hire a freelancer rather than an agency for AI work

An agency proposal at fourteen thousand pounds, a freelancer quote at four thousand for nominally the same scope. The default is to suspect the freelancer of being too cheap. Often the freelancer is the right answer.

A woman at a kitchen table looking at a laptop screen with a notebook open beside her, late afternoon light through a window.

Reading 'AI-powered' claims on vendor websites

By 2026, 'AI-powered' on a vendor site tells you almost nothing. Four questions surface what the product actually does.

A woman in her late forties sitting across a meeting-room table from a vendor, her open notebook in front of her, listening as he speaks

Spotting an AI vendor who doesn't know what they're doing

Real AI vendor expertise versus visible inexperience is observable in the first thirty minutes of a sales conversation. Five tells, applied consistently, filter the worst of the field before anyone signs anything.

A small business owner at a kitchen table with three printed quotes spread in front of her, a notebook open beside them, holding a pen

Choosing between an AI consultant, an agency and a SaaS tool

Three quotes on the desk, one from a solo consultant, one from an agency, one from a SaaS tool. The choice is not which is best, it is which fits the job. Here is how to tell.

A small business owner at a kitchen table reading a printed comparison sheet with a notebook of handwritten notes and a closed laptop beside her

Build, buy or hire, matching the path to your situation

Three paths exist for getting AI into your business. Technical owners over-build, non-technical owners over-buy, time-poor owners over-hire. Four variables decide which path actually fits.

An owner at her kitchen table writing on a printed page of four questions, her closed laptop pushed aside and a mug of tea next to her, late afternoon light through the kitchen window

Four questions to answer before you buy any AI tool or service

The four questions an owner-operator should answer in writing before any AI vendor demo, so the demo serves the question rather than reframing it.

html

If any of this sounds familiar, let's talk.

The next step is a conversation. No pitch, no pressure. Just an honest discussion about where you are and whether I can help.

Book a conversation