Resources · Comparison
Lunar8Ops vs ChatGPT for tenders and payment claims
ChatGPT is better than Lunar8Ops at open-ended writing and reasoning from a blank page. It cannot hold your rates permanently, guarantee it will not contradict a price you have certified, act on work arriving in your inbox, or route a draft to the person who must sign it. Those four are the product.
Last reviewed
This comparison exists because the objection is fair and we get it constantly. A subcontractor who has pasted a tender question into a chat window and got back a decent paragraph has correctly worked out that most of the writing problem is solvable with a good model and a clear prompt.
They are right about that part. The argument is not that the model in this product is better — the underlying models are broadly the same ones, and on any given paragraph a careful person with ChatGPT will often do as well. The argument is about what sits around the prompt.
So the comparison below is not chat-versus-us on writing quality. It is about the four things a chat window structurally cannot do, and the two things it does better than we do.
Where ChatGPT and other chat assistants wins
What we are not better at.
It is better at anything genuinely open-ended
Thinking through a commercial position, arguing both sides of a contract clause, drafting something you have never written before with no precedent to work from. A chat window with a strong model and a person who prompts well beats a system optimised for producing your standard documents consistently.
It costs a fraction of this and starts working immediately
Twenty dollars a month and no setup, against a subscription plus, for the subcontractor tier, a setup engagement. If your document volume is low and your documents are not similar to each other, the cheaper tool is the right tool and we would rather say so than sell you the wrong one.
It is not one more system to administer
No account for the office to manage, no library to keep current, no approver group to configure. For a one-person business that writes a handful of documents a year, that is a real advantage and not a small one.
Side by side
Where the difference actually is.
| ChatGPT and other chat assistants | Lunar8Ops | |
|---|---|---|
| Your rates and prices | Pasted in each time, or remembered inconsistently. Nothing stops it quoting a figure that contradicts your price list. | Held permanently as certified tables. A figure that contradicts a certified price is removed before the model ever sees it — a comparison, not the model being careful, so it cannot come out wrong. |
| Claims it cannot evidence | Will write a plausible sentence about your ISO certification whether or not you hold one, because a plausible sentence is what it was asked for. | Flags anything it cannot evidence from your own material rather than writing around it. The flag is the feature — it is what stops an unevidenced claim reaching a prequalification response. |
| Work that arrives while you are on site | Nothing happens. It is a thing you have to remember to open, and forgetting at 4:47 on a Thursday is the entire problem. | Point it at an inbox or folder with a rule in plain words. When something matching arrives it reads it, does the work, and leaves the draft waiting. |
| Who signs it | Whoever had the browser tab open. | A draft routes to a named approver, and every document carries a sign-off trail: who approved it, when, and against what. |
| Your writing voice | Described in a prompt, or inferred from whatever you pasted this session. | Learned from documents you have already sent, and it stays learned. |
| What happens to your data | Depends on the plan and the settings. Consumer tiers have historically defaulted to training on conversations. | Uploads deleted within 24 hours. Nothing you write trains anyone's model, on any tier. |
| Output format | Text in a browser window. Formatting is yours to do. | PDF or Word on your letterhead, in the format your past submissions used. |
Choosing
Stay with ChatGPT and other chat assistants when
- —You write a handful of documents a year and none of them look like each other.
- —The work is thinking, not producing — a strategy, a negotiating position, a one-off letter.
- —You have nobody to route approvals to and no rates worth maintaining.
Lunar8Ops earns its place when
- —The same five or six document types come around every month and the prices in them matter.
- —An unevidenced claim in a prequalification response would cost you a place on a panel.
- —The work arrives when the person who could answer it is on the tools.
- —Someone other than the drafter has to approve what goes out.
Fair questions
On this comparison.
Can I just give ChatGPT my price list every time?
You can, and for a small price list it works. What it does not do is guarantee consistency: nothing compares the drafted figure against the certified one, so a quote that contradicts your list reads exactly like one that does not. That difference matters at the point where a wrong figure becomes a contract.
Aren't you just a wrapper around the same model?
The drafting step uses a general model, yes — and you can bring your own provider account on the paid plans, which makes that explicit rather than hidden. The product is the part around it: the certified rate tables, the refusal to state what it cannot evidence, the rules that act on arriving work, the approvals routing, and the measurement of whether any of it paid off. None of that is prompting.
What if I already have a prompt that works well?
Bring it. A format you describe in a sentence, or two examples of something you already write, becomes a reusable document type with its own fields. The prompt engineering you have already done is not wasted.
Judge it on one of your own tenders.
Tell us your company name and we will send back one page on a package that is open right now: the questions where your published evidence does not cover the answer, with the closing date at the top. You send us nothing, and it is useful whether or not you ever reply.