AIat workStart an enquiryEnquire

Home / Tools

Which tool for which jobA side-by-side of the jobs Claude tends to suit and the jobs ChatGPT tends to suit, with a third band for work that belongs in neither.Two tools, taught side by sideWhich one you reach forWe sell neither. The useful skill is knowing which to open before you start typing.ClaudeReach for it when the input is long or the outputhas to sound like you.A 60-page contract you need the gist ofSustained drafting in your own voiceA document bundle, read togetherThinking something knotty through in onethreadChatGPTReach for it when the job is short, repetitive, orstarts as a voice note.A voice note on the drive home into afinished draftTwenty listing descriptions before lunchQuick reformatting and tidy-upsJobs you do the same way every dayAnd when to use neitherIdentified client records, clinical judgement, financial advice, tax positions and safety sign-off.Knowing where that line sits is part of the training, not a footnote to it.
We hold no reseller agreement, partner status or referral fee with either vendor.

Which tool for which job

Task by task, with the reason. Disagree with the reason and you'll pick correctly anyway, which is the point.

By

Reviewed September 2026. This page names fast-moving vendor detail, so it carries a review date.

The general mapping

JobToolBecause
Summarise a contract or bundleClaudeLong input, narrow question, verifiable locations.
Draft forty follow-ups before lunchChatGPTVolume work; per-item bar is clarity, not craft.
Write a report from voice notes in the carChatGPTSpeech to structured text is its strongest everyday habit.
Keep fifteen vendor reports sounding like one agencyClaudeSteadier voice across sustained drafting.
Restructure a messy client spreadsheetClaudeHolds structure across long, inconsistent input.
Reply to reviews and write website copyChatGPTShort, high volume, low risk.
Draft a policy that must satisfy three obligationsClaudeMulti-constraint reasoning you can inspect.
Turn a photo of a form into textChatGPTSmoother photo handling in practice.
Decide a clinical, legal, tax or safety questionNeitherRegulated harm. Stays with the qualified person.
Work on an identifiable client recordNeitherNot until tool, tier and written rule are settled.

When the answer is neither

Two tests. First: if this output were wrong and nobody caught it, is the consequence a bad afternoon or a regulatory problem? Second: does doing this job require information that's on your never list? Either one means the tool isn't the answer, no matter how well it would perform.

Cost, honestly

Most small businesses end up paying for one tool properly rather than two, and which one depends on the shape of the work above. Where it genuinely warrants both, that's usually two or three people, not everyone.

The tier decision is the important one and it isn’t about price — tiers and what happens to your data.

The four questions that settle it

  1. How long is the input — a paragraph, or a bundle?
  2. How many times will I do this today — once, or thirty?
  3. Does it have to sound like us, sustained, all the way through?
  4. If this output were wrong and nobody caught it, what's the worst outcome?

The fourth is the one people skip and the one that matters. If the honest answer involves a regulator, a patient, a court or a roof, the job is in the third category and the tool choice is irrelevant.

Two mistakes to avoid

Treating long as important. A two-line email to your largest client is high-stakes and short. Stakes decide how carefully you check; length decides which tool you open. Two different axes, and conflating them is why some teams over-engineer short work and under-check long work.

Using the third category as an excuse. "Our work is regulated so none of this applies" is an expensive sentence. The regulated judgement stays with the qualified person. The tender, the induction pack, the standard explanatory letter and the monthly report do not — and that's where the hours are.

Twenty minutes on a call usually sorts your jobs by shape and ends the internal argument.

Start an enquiry

The questions we get asked about this

Does this change with every model release?

The shapes don't. Specific strengths shift, which is why the rule is about input length, volume and consequence rather than benchmarks.

What if we only want to use one tool?

Then learn which jobs sit on its weak side, and check those outputs harder. That's a reasonable position, as long as it's a decision rather than an accident.

Get a training plan built for your business. Nine questions about how your business actually runs. You get a plan ordered for your work, a reading on where you're exposed, and one clear place to start. No account needed. Build your plan.

Start an enquiry

Tell us what's eating the week

One form, one reply from a person. We'll tell you which engagement fits, roughly what it costs, and if we're the wrong fit we'll say so.

Start an enquiry
Start an enquiry