Has anyone else had issues with Gramlee taking a long time to get proofreading done?
Lately, I’ve noticed some delays and even had to send reminders to get things moving. For instance, @DouglasDraper has been waiting over a week for a case study to be proofread, despite following up several times. This is pretty frustrating, especially since they advertise “Revisions in 24 hours or it’s free.”
I’ve thought about trying AI as an alternative, but I’m not sure that’s a great idea right now. This article breaks down why AI might not be the best for editing, and ChatGPT has a pretty distinct style that doesn’t always work.
@gabriel, tagging you since you might have some experience with Gramlee and could share your thoughts…
@cassie Surprising of Gramlee, which historically has been quite reliable. Maybe take them up on their guarantee, and ask for a refund on the late proofreading? That should get their attention
LLMs can help finding grammar/syntax issues too yes - they are pretty good with that. They might miss more than a human editor, so it could depend on the importance of the document to proofread.
Note that you don’t need to make the LLM write the text for you. You can ask to list the grammar & text issues for example, which you can fix then yourself the way you want. It can be quite effective, while keeping full authorship over the document.
Although it is not exactly for proofreading or fixing grammar as is, I use https://languagetool.org quite frequently. They provide more privacy than Grammarly and have fantastic suggestions for grammar and stylistic mistakes as well. Sure, the paid version is more beneficial, but the free version is good enough too!
Gramlee finally got back to us. It seem that our emails are ending up in their spam. They’ve asked us to email Rushang directly now. I’ll make an update to the handbook
@gabriel and l have made the decision to move away from Gramlee due to the lack of service and price increase. And now I need your input.
What do we need?*
Quick feedback (within 24 to 48 hours)
Easy process for sending documents
Cost-effective (@gabriel@tikr what budget range should we keep in mind?)
AI vs Human Editing
AI Editors: Fast, more affordable, good for grammar, spelling, and tone checks. But they lack human nuance, they won’t always “get” brand voice, context, or subtle meanings.
Human Editors: Great for grammar, spelling, context, and maintaining voice. More reliable for important documents, but usually slower and more expensive.
Although AI can be good as you pointed out, my eyes are burning nowadays from AI content. I would say not to go with AI content proofreaders. I know we talk about edits, not full content generation, but I can directly pinpoint those 2-3 phrases usually recommended by AI. And I’m not talking about em-dashes as I use them as well in writing, especially with Mac–right?
From the list offered, I think PaperTrue, and Scribendi could both work. I’m not sure about Proofed as we don’t really work in a setup that would require 24h turnaround.
That said, I think a strong, detailed prompt could go a long way in shaping an LLM’s output. We could specify OpenCraft’s desired tone, punctuation, style choices, and even quirks like avoiding em-dashes if we want!
Ideally, I’d prefer human editors, but that is not always realistic in terms of time and cost. So my vote is for the hybrid approach. I don’t think word count should be the deciding factor though. To me it would make more sense to base the choice on the importance of the piece, even if that is harder to define.
Thanks @Ali and @gabor for your feedback. I agree that AI has a clear way of “air-brushing” content. So I worry people won’t have the attention to detail to remove those tell-tale signs.
I also agree it should be based on the importance of the piece. I think let me try PaperTrue as they do have a free sample option, and wait on @gabriel and @tikr to give me an idea on cost.
@cassie thanks for layout all these options! In terms of cost range, all the human options that you listed are cheaper than our previous service, so there’s no blocker from a budgetary perspective.
@cassie Just to clarify, I’m not involved in tracking tool costs or coming up with budgets for third-party services, etc. If something is $-only, @gabriel is the best person to help you with any type of question you might have
I’m OK with using AI assistant tools for this task, so long as the changes are made very clear in-context. What I don’t want is an LLM to have the final say on the content, or to put an article into an LLM and just say ‘fix it’ to get the result out.
It looks like the tools you linked don’t do that, so they should be fine. Grammarly is pretty well-regarded here-- I’ve heard good things.
However, with the volume of content we’re creating, one-off submissions to proofreaders may be just as cost-effective. Especially if the user-count requires the whole org to be subscribed or automatically adds a seat any time someone does a sign-in with Google apps, for instance.
But to sum: as long as I have veto power on suggestions and the solution is reasonably good at providing them, I’m fine. I’ve voted for human editors, but my preference is slight and is banking on the fact that if we’re paying for human editors at all, it’s likely to be the most expensive option to do both, since we’ll be paying monthly subscription fees rather than usage fees for the AI tools, and usage fees for humans.
I’m waiting on my PaperTrue free sample to be returned. I already like how their web app works. The service I’ve tried is “LLM + Light Human Reviewing”. I think that might meet the sweet spot based on all the comments mentioned above – but let’s see if the copy returned is well edited.
