Questions we answerAI and automation
Should I let AI write proposals and client-facing work?
Split the document. Structure and boilerplate, usually yes. Pricing, scope and anything you would be held to, no — and expect drafts to converge with your competitors'.
Free · a few minutes · starts from “I think AI could help”
The short answer
Split the document before deciding. Structure, boilerplate and first drafts are usually fine. Pricing, scope and anything that creates an obligation are not, and the reason is not writing quality — it is that the judgement is not yours if you did not make it. Expect a second effect: if you and a competitor start from similar briefs and similar models, the proposals converge. Keep the parts that make you distinctive written by a person.
The useful version of this question is not whether AI writes well enough. It usually does. It is which parts of a client-facing document carry commitment, and whether anybody is reading them before they go out.
There is a second, quieter risk that vendors have no reason to mention. A drafting tool that everybody uses produces work that reads like everybody's. In a competitive proposal, sounding like the other submissions is a commercial problem rather than a quality one, and it is invisible from inside your own document.
Is it really this?
Worth checking before you spend anything. These are the signals that tend to tell the two apart.
Signs it is
- A large share of each document is genuinely the same each time.
- Somebody is named as the person who signs it off, and does.
- The time cost is real and is displacing work you would rather be doing.
Signs it's something else
- Every proposal is genuinely bespoke, in which case the drafting is the thinking.
- The slowness is approvals and chasing, not writing.
- You have no templates, which is usually the cheaper fix and gets tried second.
What's usually behind it
- Proposals taking days that the business would rather spend on delivery.
- A backlog of similar documents where most of the content genuinely repeats.
- Somebody having already started doing this quietly, and a wish to set a rule before it goes wrong.
What people usually get wrong
- Letting a draft carry a price or a deadline that nobody consciously agreed to.
- Judging output on how well it reads rather than on whether it commits you to anything.
- Pasting client-confidential material into a tool without checking what the account terms say about it.
The options, and when each one fits
In no particular order. Which of these fits depends on your situation, not on the question you typed.
Draft with AI, review before sending
Structure and repeated content, with a named person reading it before it leaves.
Keep the commitment sections human
Always, for price, scope, timescale and anything a client could hold you to.
Improve the templates first
Most of the repetition would be solved by a good template, which is cheaper and more predictable.
Keep it entirely human
The document is the differentiator, and sounding like everybody else costs more than the time saved.
When doing nothing is the right call
Where proposals are few and each one genuinely wins or loses on its thinking, leaving this alone is defensible. The time saved is small and the thing at risk is the part that wins the work.
What the diagnostic asks
- Where you think AI could help, and how repeatable that work is.
- How much judgement it needs and how often it comes round.
- What happens if it goes wrong and how sensitive the information is.
- Whether software or templates already exist for it.
“I don't know” is always an answer. Nothing is assumed on your behalf.
The routes it weighs
Every diagnostic weighs all ten routes. These are the ones that tend to matter for this question — a description of the rules, not a promise about your answer.
AI Assist
Keep a person in charge and give them AI for drafting, summarising, checking or analysis.
Process improvement
Change how the work flows — remove steps, name an owner, fix the hand-offs.
Do nothing yet
Record the issue, set a revisit date, and decide what would change your mind.
Recruit
Recruit a new person for a durable, sizeable need that is central to the team.
Outsource
Hand a defined, ongoing scope of work to an external provider.
Specialist support
Bring in a freelancer, consultant or specialist for a defined piece of work.
What changes the answer
- Sections that create an obligation — price, scope, timescale, warranty — should stay with a person whatever the volume.
- A reviewer who genuinely reads before sending changes the risk profile of everything else.
- Client-confidential material changes what may be pasted anywhere, and that question comes before the drafting one.
Get your own read
A handful of short questions, then the strongest route for your situation, what's also credible, what we still don't know and what would change our view. No account, no setup.
Related questions
- Can AI help my team?Often, but the useful question is where. AI that helps a person draft, summarise or analyse is a different decision from AI that runs a process.
- Should I automate this work?If it is repeatable, frequent, low-judgement and written down, quite possibly. If any of those is missing, fix that first.
- Can AI handle customer enquiries?Part of them, well, with a person holding the line. The split that matters is between enquiries with known answers and those needing judgement, empathy or authority.
- Quality is slipping in my team. Is it people, process or workload?More often process than people. Slipping quality is usually a sign that the work has outgrown the way it is done, not that anyone has stopped trying.
- Should I hire or automate this role?It depends on what the work actually is. Repeatable, frequent, low-judgement work can often be automated; durable, growing, judgement-heavy work usually needs a person. Most real roles are a mix, and the honest answer is often to automate part of it first.
- Who maintains our automations when they break?Whoever built it — and if they have left or are too busy, you do not have an automation, you have a liability nobody is watching.