The buyer's guide, chapter 3

The buyer's guide, chapter 3: Decide what kind of help you need

1:46

The buyer's guide, chapter 3: Decide what kind of help you need (1:46)

Key points

  • An early decision is which kind of AI help you need. The phrase covers very different work, the right kind depends on your aim, and no kind is best for everyone.
  • Advice helps you choose where AI might help and what to try first. Discovery is a brief, paid look at one task, ending in a recommendation you can use, whether or not you keep the same supplier.
  • Setting up ready-made tools you have or can buy often gives a first project the most for its money. Building something new costs more, with more upkeep and a bigger handover.
  • Training often decides whether staff use a tool or ignore it, and support maintains live systems. A good first project often pairs a brief discovery with a modest set-up or build, plus training.
  • A freelancer is one person: direct and usually cheaper, but with limited time, and you rely on them alone. A consultancy is a team with broader skills and better continuity, usually at greater cost.
  • Agencies tend to suit fixed jobs more than open questions. A software vendor's advice usually leans towards what it sells; that is fine if you know. These are loose patterns, so judge suppliers by the work they offer, not their label.
  • With under ten staff, a modest first step at a clear price usually works best, such as a few days setting up a tool you already pay for, plus training. The owner often holds the budget, uses the tools and checks the work, so keep the project small enough to fit beside running the firm.
  • Firms of ten to forty-nine can usually run one real trial, with one staff member in charge. Large organisations, with procurement rules and security reviews, face slower buying; the same principles hold, but more people must agree.

Transcript

An early decision is which kind of AI help you need. The phrase covers very different work, the right kind depends on your aim, and no kind is best for everyone. Advice helps you choose where AI might help and what to try first. Discovery is a brief, paid look at one task, ending in a recommendation you can use, whether or not you keep the same supplier.

Setting up ready-made tools you have or can buy often gives a first project the most for its money. Building something new costs more, with more upkeep and a bigger handover. Training often decides whether staff use a tool or ignore it, and support maintains live systems.

A good first project often pairs a brief discovery with a modest set-up or build, plus training. A freelancer is one person: direct and usually cheaper, but with limited time, and you rely on them alone. A consultancy is a team with broader skills and better continuity, usually at greater cost.

Agencies tend to suit fixed jobs more than open questions. A software vendor's advice usually leans towards what it sells; that is fine if you know. These are loose patterns, so judge suppliers by the work they offer, not their label. With under ten staff, a modest first step at a clear price usually works best, such as a few days setting up a tool you already pay for, plus training.

The owner often holds the budget, uses the tools and checks the work, so keep the project small enough to fit beside running the firm. Firms of ten to forty-nine can usually run one real trial, with one staff member in charge. Large organisations, with procurement rules and security reviews, face slower buying; the same principles hold, but more people must agree.