DFIRST Reviews

The tool itself is generally as good as any other AI generation workflow tool.

It has its moments and hiccups from time to time, and for the past few weeks all the icons of each generation box has been missing for me so I have to blindly click and hope I hit the right button to download my generations.

The reason for the 1-star review is that DFIRST have violated their own terms of this LTD.

Sumolings are supposed to get access to all future features of their plans, yet they are running a summer promotion for unlimited generations that is inaccessible to Sumolings.

If you have a Tier 1 LTD, this matches their Starter account, yet you do not get access to their summer promotion for Starter accounts. You appear to only be eligible for this if you have an ongoing subscription.

If this is not available to Sumolings, they should create a new account tier and make the promotion available to that, rather than effectively saying "If you are on Tier 1 LTD you have a Starter account, but not a 'real' starter account unless you pay the ongoing subscription!".

This instantly loses all credibility for me.

TLDR: A solid 3.5 from me, will be higher once the points below are addressed.

Full disclosure - my AI prompting and leveraging skills are early stage, therefore I can only relate my my experience of Dfirst as a AI wrapper and what value it adds to my usage of AI within the remit of Dfirst.

The frustration of using individual models and cobbling together the data into actions is the overlong LLM repetative steps, long outputs and very scattered data and research across queries, sessions and AI models.
Dfirst is a wrapper which aims to add value to other ai tools. It's an attractive prospect to have all the best models available, workflows, persistent assets and a whiteboard to work on that appears to address the issues, however the devil is in the detail, only through real world use can I see if it works better.
And thats the key for me at the stage is the interface delivering sufficient efficiency and quality to be worthwhile? I am not wholly convinced yet, the whiteboard for my project has become rather difficult to work with, its very large - so large I have to continually zoom in and out to move around it or know where I am in it. Its a bit of a mess but with plenty of quality within it but its become a rather disorganised mess. On the one hand If it hadn't been so much work to get to this stage I would start again but on the other hand I'm really no clearer how to use it properly so there is not much point to me doing that.
My attempt to produce a video failed, it seemed to want me to make a separate prompt and video for every scene then stitch it all together after (too much work and not confident it would be usable)
Another irritation is UI, large whiteboards are difficult to work with, (a very large dark warehouse with a tiny torch) that requires continual zooming in/out, - exporting the good bits of content out of Dfirst didnt work well, with multiple copy/pastes for each paragraph, I gave up on it because that wasnt working.
Other issues I had were errors that I could not correct etc just a bit frustrating and with the amount of confusion I did start to wonder if it was worthwhile using Dfirst at all.

However, the objective of Dfirst is extremely valuable, it is needed.
I am hopeful that we (Dfirst and myself!) will get there;

Heres what I think
- whiteboard interface : the UI is not yet working well for me, it soon became confusing and awkward to use, possibly I went too wild with agentic ai on a startup project, im not sure.

Training materials:
- its not intuitive enough: I think this should be addressed with a multiprong user education strategy: within the app by embedding tool tips etc and user guides, help articles etc.

- also some video workshops we can refer too, the rather dry "user guide" type of information is always a priority to give a clear "bottom line" on how things work. what it can/cannot do, but i feel with a tool like this and "over the shoulder" workshop of an expert using the tool and talking through what he is thinking/deciding/doing would be excellent - Appsumo (not Jay) already did something like this which was great, I really rate that guys videos BUT he was a novice with the software exploring it and not an expert demonstrating it.

- I 'd like mutiple over the shoulder demos by an expert with Dfirst/prompting/AI/marketing demonstrating all Dfirst's capabilities with real world examples multiples of them actually, they can just be raw not overedited videos for speed - so users can get to extract value from the app as quickly as possible and also become proficient users of it so that they can provide valuable feedback that will actually help the DFirst team rather than burden them with support tickets, or provoke negative reviews.

To be clear these demos would ideally be a mix of :
1) Cradle to grave Demos : fully featured new startup projects, EVERYTHING that Dfirst can do for you, how to do in Dfirst (how to organise it into 1 whiteboard or multiple?)
2) Discrete Task Demos: eg. "How to create a logo for your new start up" , "how to create a promo video" "how to create a high converting sales page from scratch"

So overall I think the potential of Dfirst is massive, the UI and UX is not delivering yet but I am hopeful it can do - but I think it needs some work. As a user I need more support materials to be able to use this effectively at all and also to make it worthwhile from an efficiency perspective.
3) I would suggest recording user webinars too and adding them to your members materials or youtube chanel. Powerful software like this lends itself to webinars especially in the format of : Part 1 demo/tutorial of specific tasks/features
Part 2 user questions answered by the expert with an online demostration.
I think this would also be a perfect way for the team to fully test the UX and UI and refine it too because sometimes it feels like either I'm doing something wrong or the team cannot have used this tool for what I want.

DFIRST is a powerhouse tool that adds a ton of potential for how freelancers and agencies handle client creatives. I've spent plenty of time trying to organize files, remember what creatives worked for inspiration, and tried to keep opening all my related files to make sure I keep my brand voice, look, and strategy all in line.

DFIRST is a great one-stop shop for simplifying the workflow and bringing it all together while researching and building creatives.

Need to research a social profile or website to see what the brand is about, what strategy it is using, who the audience is, and break down the things that makes that audience tick? DFIRST gives you the building blocks and tool to research with ease.

The default research template workflows do a great job at getting you the basics if you need to do some quick research, but the real gem here is the ability to go through with a fine-tooth comb to edit the research output (if needed), build your own workflow, or add your own context as you like.

While building strategy and creatives, you can use previous items, such as research, brand guides, and more as reference material for what you want to generate. No AI is perfect, so you must always do your due diligence to make sure your output is useable. But the process of going from research to implementing a plan has gone from days, weeks, and months (perhaps even years) to - in many cases - just minutes.

The credit limits certainly feel low when getting into video or heavy amounts of image generation, but it also gives me more hope that this tool will be sustainable going forward compared to others I have tried in the past.

I appreciate all the updates that have been coming out recently, such as their new Brand DNA feature. The areas I'd love to see some improvement are related to UX with the whiteboards and the image editor and perhaps some improved walkthroughs when you first get into the tool, as it can feel a little overwhelming at first and can feel like looking for a needle in a haystack when wanting do do something specific - but I also appreciate all that is currently in place.

The workspace, whiteboard, and social media scraping limits alone are generous in this deal - making it a no-brainer in my book. As long as improvements continue to roll out, DFIRST has a bright future.

5 tacos out of 5

DFirst is definitely the best purchase I’ve made on AppSumo, or at least one of the best. I liked that the agent is specialized on marketing projects. You’d probably fall even more in love with it if you tried it yourself. I bought Tier 5 and I’m going to upgrade to tier 6 very soon.

But I want to give feedback for improvements:

The things that need improvement are the workspace usability. It seems like it gets stuck when you want to move it around. The assistant sort of lays out the whole plan at once, and then you give feedback bit by bit. It listens to that feedback initially, but if you have a whole list of feedback that you want to go through step by step, it doesn’t really let you. It just takes the first piece of feedback, changes that, and then you basically have no choice but to approve it because it won’t let you keep writing more feedback. You have to approve it or approve it, basically, and that’s something I didn’t like.

And for longer tasks, like if I want to define a brand launch campaign, for example, there are many little steps before getting to the final part, like making assets. There are all these micro-tasks: planning, defining the persona, branding, and so on. If I disagree with the result at any step, it’s a hassle because changing one specific result affects many things that come after it, and every time you make a change, it consumes a lot of credits in the long run. I think it should let you go step-by-step, show you things gradually, and let you approve or tweak them bit by bit instead of forcing you to approve everything at once.

I need this tool for defining launch campaigns, creating the foundations and identity of a brand when it’s being established. I want it to set up marketing plans on at least an annual level, and ideally with a three-year perspective as well. It’s important that it helps define the core elements of the marketing strategy, like which channels to use, what messages to convey, and how to differentiate in the market.

At first glance, DFIRST seems amazing, so I was excited to test out the tool. Before diving in, I watched all their YouTube videos (which isn't much tbh), and then tried using the tool on some ongoing projects we have.

Initially, the tools seems fantastic: it does an amazing job at understanding what I want, and lays out a clear path (that I can also edit!) to achieve the goals I'm looking for. The fact that it also uses multiple AI models, and always recommends the best model for the job is also great.

So, that's the good part...

Now the bad:

1) The training is VERY surface level and incomplete. This means that I a) either need to be a pro at AI agents & workflows already to make things work, or b) waste TONS of credits to figure things out on my own. I'd appreciate a comprehensive onboarding & training experience to get up to speed with DFIRST without having to waste multiple hours of my time, and hundreds (or even thousands) of credits on trying to figure things out on my own with substandard results.

2) The AI workflows/agents aren't as smart as they seem... And by this, I mean many things – just a few examples:
• I asked them to scrape a website and Instagram to understand the brand's colors, language, fonts, and content pillars. They scraped both (apparently successfully), but still used DFIRST's colors, still used English as the language (despite the particular website being in a different language), still just guessed the fonts (incorrectly), and completely missed the content pillars, formats, and visual style. And all this happened multiple times...
• The tool fails to establish a clear, logical workflow... By this, I mean that oftentimes it doesn't reference crucial information or inputs on its own, and wastes A LOT of credits by doing so. Yes, I could technically go in manually and tweak things afterwards, but a) that takes time, and b) that means I've just wasted ~100-200 credits on useless garbage.
• Its foreign language skills are not excellent. Not bad either per se, but it does make errors in languages other than English (so far, I've tried 2 other European languages). Mostly weird capitalization errors, so not that big of a deal, but still...

3) The output is just plain bad. For customer research, I've achieved better results using a single prompt in Gemini Pro alone. The social media posts created in DFIRST are downright terrible. In fact, even within a single workflow, the visual style, the tone of voice, and even the language (!!!) of some of the posts generated are completely different. Let that sink in: the same brand, the same prompt(s), the same workflow. And the 3 posts generated look like posts for 3 different brands, operating in 3 different geographical regions (and one of those completely missed the brand's product, displaying it as a supplement, when in fact, it was a diagnostics product). Not to mention that the posts I've managed to create are terrible as well: lazy, boring, and definitely no potential for engagement, virality, etc.

4) DFIRST is very credit-hungry. I bought the top tier, and already – just by playing around with a few basic workflows (no videos, and barely any images), I've spent about 20% of my monthly credits. Yes, I know there's a relatively inexpensive way to top up my credits, and I also know the tool needs to be sustainable in the long-run... But for the ~$800 I've spent on this tool, I'd be able to buy A LOT of credits, or even multiple years of unlimited subscriptions to mature, established AI tools. At this credit consumption rate, DFIRST feels almost like a ripoff.

5) And finally: while I wasn't affected personally, I see the founders launched a similar tool on AppSumo a few years ago that they've since abandoned. I don't know what happened to their old customers, but after a quick Google search (Digitalfirst Appsumo), I've found the following:
• Customers suddenly losing access to their dashboard (apparently without a warning) – Does this mean they've also lost access to all their generated assets, prompts, databases, and references?
• Customers complaining that there's no clear way to transfer their account to DFIRST, despite it being more-or-less a mature & rebranded version of the same product. – Does this mean that those customers were abandoned?
• Customers complaining about terrible customer support. – Does this mean we can expect a similar experience with DFIRST, especially once the AppSumo campaign (and refund period) are over?
Look, I'm not saying those customers are necessarily right (I have no idea), and I'm not saying us DFIRST customers are in for a bad experience (I certainly hope not) – but this is definitely a red flag that I think all Sumolings should be aware of.

All-in-all, I'll see how DFIRST evolves over the upcoming weeks. What I'm really looking for is a complete onboarding/training experience to ensure I can use the tool as intended. And I also hope they can improve the quality of the AI outputs significantly.