To view this video, please accept marketing cookies.
Posted: Thu 19th Nov 2026
In this Lunch and Learn, Dropbox’s Rahul Kurane and Stephanie Anani from OpenAI explore how small businesses can do more with AI by connecting it to their existing business content, from drafting customer proposals to preparing for customer meetings.
Useful information is often spread across different files and folders. Bringing those materials together can help you spend less time retracing past work and more time making progress.
Rahul and Stephanie show you how to connect the Dropbox plugin, bringing your business context into the conversation with ChatGPT and using existing materials to catch up on a project, find the information you need and move from gathering context to getting work done.
They also demonstrate how Dropbox and ChatGPT can work together to turn existing business files into a first draft, check important details and save the results for your team to use.
Access a growing collection of expert-led webinars covering marketing, sales, finance, growth and more – ready whenever you are.
Transcript
Lightly edited for clarity.
Beth: Hello, everyone, and a very warm welcome to today’s Lunch and Learn. My name is Beth, community manager here at Enterprise Nation, and I’m delighted to have so many of you joining us on your lunch break today.
If this is your first time joining an Enterprise Nation session, welcome. Our entire mission is to help small businesses like yours get trusted, practical support without the jargon.
Today, we’re tackling something I know is on almost everyone’s mind: AI, of course. Over the last year or two, small business owners have constantly been told, ‘Just use AI and you’ll save multiple hours a week.’
But if you’ve actually tried using it in your day-to-day work, you’ve probably hit a very familiar brick wall. You sit down to do a task you’ve done a dozen times before.
Maybe it’s writing a proposal, handing over a project to a freelancer or prepping for a kick-off meeting with a new client. You open your AI tool of choice thinking, ‘Great, this will take two minutes,’ and then reality sets in.
Before AI can write a relevant sentence, it needs to know what you’re actually talking about. So you find yourself spending 20 minutes digging through folders, opening three different documents, searching your inbox for what the client agreed, and manually copying and pasting paragraphs back and forth into the chat window.
And by the time you finally get a draft, you’re left wondering, did AI actually save me time, or did I just spend half an hour doing manual data entry? So today is all about solving that exact disconnect.
How do you get genuinely impactful work out of AI when all of the knowledge it needs is scattered across your business? To show you how to bridge that gap, I’m thrilled to welcome Rahul Kurane from the solutions team at Dropbox and Stephanie Anani, a solutions engineer at OpenAI.
They’re going to show us live how to connect the Dropbox plugin, bringing your business context into the conversation with ChatGPT. But before I hand over, we’d love to hear from you in the chat.
Where are you joining us from today? What is your business, and what is the one administrative task that steals far too much of your time?
Drop your answers in the chat, and feel free to submit questions in the Q&A box as we go along. So with that, over to you, Rahul.
Rahul Kurane: Thank you so much, Beth, and thank you, Enterprise Nation, for having us. It’s genuinely great to be here with all of you today.
Before we dive into things, a little bit more about me and my role. As Beth mentioned, I work in the solutions team at Dropbox, and every day, I get the opportunity to work with small businesses. It’s been a real joy and a privilege.
One of the things that I’ve learned over the years is that as a small business, you don’t have the luxury of dedicated departments. You don’t have a sales department, a marketing department, a finance department or an operations team.
What you have is likely one individual who’s pitching projects, delivering the project, collecting all the materials and then really doing the operational stuff, helping to get the paperwork moving. So where you, as a small business, spend your time is absolutely crucial. Time is money for you.
What I want to start with today is a simple scenario that I think a lot of you will recognise. Imagine you’re a small design agency or a boutique consultancy, and you’ve promised a new client that you’re going to have a project kick-off call tomorrow.
So you’re probably thinking to yourself, ‘To help put together this kick-off deck, I’ve got material I need to reference that might be in a folder in Dropbox.’ You might have an onboarding checklist that’s in another folder within your Dropbox.
Maybe you have notes from previous meetings, and all of that information is dispersed across your Dropbox account. So before you even begin drafting this kick-off deck, you’re probably thinking, ‘I’ve got to search for that content first. I’ve got to find it.’
You’ve got to check whether the document you’re using as input is the latest version. And then you’ve got to remember everything that you agreed upon in prior calls. It’s a lot of information that you need to pull together.
And then you bring AI into that process. Now you find yourself spending 15 minutes, 20 minutes, even an hour gathering all these different files, copying text from one file, pasting it into the prompt window in ChatGPT and then building the deck that you need. That can be a pretty cumbersome task.
What Stephanie and I are going to be talking about today is how we can help you simplify that process. In fact, at Dropbox, we recently did a study where we surveyed 500 professionals and asked them various questions about how they were using AI today.
I’ve got two really interesting nuggets from this survey. 84% of the professionals said that they agree AI helps them move faster, but it’s helping them move faster than the underlying processes can keep up with. Keep that in mind – that’s something I know a lot of us struggle with.
But this other number, I think, is even more crucial. 74% of that population told us they are manually moving data and text, copying and pasting content from different apps into their AI tool, into ChatGPT, at least three times during a single working day.
So while ChatGPT can easily draft something within seconds, the amount of time you’re spending just gathering all that information introduces a ton of manual friction. It becomes really hard to build the information set you need to actually deliver on that client deck.
So I have a question for all of you. When you think about AI and your use of it, what is it that you’re struggling with the most? Is it finding files, explaining the project background to the AI tool, or verifying the results?
I’d love to hear from you, so I’ll take a pause here to see if anyone chimes in. When I talk to clients, I ask them this very question, and what they tell me is that it’s finding the files and explaining the background to the project that takes the most time.
I see this across the board. It doesn’t matter if you’re a design agency, a consulting firm or a small mom-and-pop shop.
So when it comes back to that project kick-off, your goal is to start with an approved proposal, pull together all the notes you’ve collected over the many meetings you might have had and then bring them back to your AI tool.
From a more practical standpoint, instead of copying text between all these different windows, you simply want to say, ‘Hey, ChatGPT, use the approved proposal, use these discovery notes, use this information, and generate this kick-off deck for me.’ So with that, I’d love to loop in Stephanie here.
Stephanie, when you think about all the work you’ve done with clients over the years, why does an AI tool struggle when we only give it a generic request? And why is business context, as I just described, so fundamentally important to a small business?
Stephanie Anani: Thanks, Rahul, and hello, everyone. It’s lovely to be here with you today.
To answer that question, we need to take a little bit of a step back and think about how AI tools like ChatGPT are trained. They’re trained on a vast amount of general knowledge, but without specific context, it knows nothing about your business, your pricing or your client agreements unless you supply that foundation.
So if you go into ChatGPT and just say, ‘Write a client kick-off brief,’ ChatGPT has very little information to go off. It’s all about making sure that it has access to the right information.
I’ll use an example from my daily life. I work with some of the largest companies in Europe to help them make AI valuable, and oftentimes, the context with these customers changes quite frequently.
We have multiple meetings a week, and things change within their business. They might have earnings reports and all these different things that really change the trajectory of the work that we do.
So what I do is make sure that all that information automatically goes into the appropriate folder and the appropriate space. For this example, it could be a folder in Dropbox.
When I go to ChatGPT, open my project for the specific client and ask, ‘What did we discuss last time? What did we agree on? Has anything changed in their business?’ All of the right context is sitting there, and I actually get a very specific answer.
For you to create a useful brief with AI, there are three things that are, in my opinion, crucial. The first is what you’re actually trying to achieve.
I know it’s sometimes tempting to go to AI without a clear idea of what you want, but starting with a clear vision of what you’re trying to achieve helps guide whatever AI tool you’re using – hopefully, ChatGPT – to the right answer.
The next is the actual source material. The idea is to get ChatGPT connected to the actual source. This reduces what we were talking about earlier – having to move materials from one source to another, and having to update context when things have changed.
And lastly, there’s an agreed scope. ChatGPT needs to know what matters and what good looks like.
Another thing that’s really important is to use ChatGPT to flag gaps. It’s all well and good to have all the context available to you, but the truth is that it might not be all the context ChatGPT needs to achieve the task to the standard you’d like.
So one thing I always ask ChatGPT to do is flag gaps, and I have a specific workflow where I ask ChatGPT to grill me on certain things. That really helps open up my understanding of what isn’t there as well.
One example of something you could say is: ‘Use the approved proposal, customer notes and checklist in Dropbox, and create a short kick-off brief covering what’s been agreed, who’s doing what and what we still need to confirm.’
Another tip I like for document editing: sometimes you might give ChatGPT a draft and ask it to refine it or add specific information. It gives you something back, and you know it’s different, but you’re not quite sure exactly what changed.
So I have a workflow where I ask ChatGPT, ‘If you’ve cancelled anything I’ve written, strike it through. If you’ve added anything of your own, either highlight it or put it in italics.’ This really gives me a lot of clarity about what ChatGPT has changed.
Another thing you can do when you’re reviewing these documents – say, for example, a key decision is buried at the bottom of a brief or a date looks questionable – is work with ChatGPT to figure out where it got that information from.
Rahul: Absolutely love those tips, Stephanie. You hit the nail on the head when you said that to build context, you’ve got to go straight to the source.
And to your point, you’ve got to ask not only for what you need but for what you don’t need. How do you plug the gaps? How do you put guardrails in place to make sure you’re pulling the right information for that kick-off deck or customer brief?
So what we’re going to do now, team, is show you how connecting Dropbox and ChatGPT makes all the difference. As Stephanie just described, Dropbox is where your business files are stored.
All the files you’re using on a day-to-day basis – whether that’s your discovery notes, your onboarding checklist or your proposal docs – are going to be in Dropbox. Then, using the Dropbox plugin for ChatGPT, you can search for that information directly within ChatGPT.
Not only can you search for that information, but you can retrieve the information you need, create shared links and create file requests. So there’s a lot you can do to simplify the different workflows you need to pull together kick-off decks and anything else you might need.
What we’re going to do now is show you a prerecorded demonstration I put together this past week that walks you through how to do this, step by step, from start to finish. So with that, let me go ahead and share my screen, and I’ll be providing a voice-over as we go along.
The first thing we’re going to do is set up the Dropbox plugin for ChatGPT. One thing I want to call out is that this plugin is available across all ChatGPT plans. It works for both Dropbox business plans and individual plans, so you can get started today.
Once you’ve installed the plugin, the first task you want to perform is finding your project files. So you can add Dropbox directly within ChatGPT and search for your client’s files.
In this case, I’m a design agency, and my client is Stackcraft Solutions. The Dropbox plugin has retrieved those files, and now, within ChatGPT, I can see I’ve got the approved proposal, the discovery notes and the onboarding checklist.
What I want to do now is look at that proposal doc and review it. I want to see whether it has the information I need to prepare the kick-off deck – the version, whether it’s been approved or not and when it was approved.
As Stephanie mentioned earlier, it’s really important to review the material to make sure you have exactly what you need. One thing being highlighted here that’s really cool is that you can preview this proposal doc directly within the ChatGPT interface.
You don’t have to jump into Dropbox and back, so a lot of the context switching you’re probably experiencing today goes away.
The next step is to turn those files into a deck. I want to use that approved proposal, those discovery notes and that checklist to prepare a kick-off deck for my meeting tomorrow.
As you can see here, I’m typing in a very simple prompt. To Stephanie’s point earlier, I’m also providing specific instructions: cover the client’s goals, the agreed scope, responsibilities, timelines and decisions; keep it concise; and flag missing and conflicting information.
That is key. As we all know, when we get into that kick-off meeting the next day, there are going to be certain things we need to confirm or verify.
As you can see here, the deck has been generated. I can now flip through the slides within ChatGPT and make sure all the content I need is there. You can see there’s a decision slide, an agreed scope slide and so on.
I brought up the example of a design agency, or even a consulting firm, for that matter. You probably have your brand, your templates and a certain way of presenting information, and you can apply that to this deck as well by providing it as part of your instructions.
What we’re going to show you as we go through this is that there are a couple of decision points that have to be confirmed. One of those is the agreed start date, and another is who the approver is.
So we’re going to verify those details by going back to that proposal doc one more time to find the information we need. In this case, it’s telling us the agreed start date is proposed but not confirmed, so that’s something we want to confirm during the kick-off.
So now we’re going to give ChatGPT a prompt to say, ‘Add some additional information to the decision slide: one, the final client approver, and two, the agreed start date.’
What’s really nice about this is that, yes, you’ve generated the PowerPoint deck, but now you can also update that deck, and that becomes really powerful. Think about the friction you’re experiencing on a day-to-day basis.
A lot of it has to do with copying information from one place and putting it in another place. That goes away.
Now you can see that the slide’s been updated. Those two pieces of information I want to confirm are now in the deck.
So at this point, the deck’s been created. What’s our final step? Our final step is to save this finished deck back to Dropbox, so we can close the loop.
This is where the Dropbox plugin for ChatGPT is so powerful. I created this deck, and all I need to do is say, ‘Save this presentation deck back to the Stackcraft Solutions folder in Dropbox.’
As I’m doing here, I’m giving it a name. I get confirmation that the deck’s been created, and I can go over to Dropbox to verify that, yes, the PowerPoint has indeed been created.
So for all of you watching this right now, think about your next project. Think about the files you need to generate that kick-off deck, that proposal brief or that update for your executive team.
I’m going to stop sharing now and bring this back to our conversation, but I hope that was helpful. Stephanie, any thoughts from you as you saw that demonstration?
Stephanie: I really like the way you showcased that. I really appreciate that we were super clear about the decision points and the data, because that’s how you get the most from AI.
My advice for anyone here who wants to take a first step would be to start with a specific task. Pick something you know and understand really well so that you can judge the output – a routine task you do every week, for example.
Because you know what good looks like, you’ll immediately be able to spot what’s accurate and what’s missing. You might even find there are sources of information you have access to that you still need to connect to really supercharge your AI processes. And then, again, compare the results with your usual process.
One thing I really like to do as well is give AI samples of really good work I’ve done. For example, I have a very specific way I like to write my emails.
So instead of letting ChatGPT write my emails out of the box, I asked it to go through all my emails, gather what my style looks like and then, for example, write emails. It really improved the quality of what I was able to do.
So just to summarise, pick something you know really well, evaluate the outcome and ask yourself: did that actually save me time? In most cases, it will save you time, but there are always ways to optimise that.
Did I get value from this? Was it genuinely useful? That’s what will tell you if the approach is worth replicating.
Rahul: 100%. I love that iterative approach you’re recommending. It’s so important to start somewhere.
As Stephanie mentioned, start with something you know well, then build from there and iterate. You’ll be surprised – within a week, you’ll have something you’re really proud of.
I want to mention a couple of additional use cases. Today, I talked about client onboarding and kick-off, but there are so many different use cases for ChatGPT and the Dropbox plugin for ChatGPT.
I’ve got a few examples I’ve pulled from past projects I’ve worked on. Number one is preparing customer proposals.
This isn’t so much about preparing a proposal using information you already have – maybe you want to use a proposal from a few months ago. You can use the Dropbox plugin to pull up that old proposal and use it as a foundation for the next proposal you’re building, because maybe you have important references or service examples you want to highlight in this new proposal. That’s a great use case.
Another one I love and have seen work really well is project handovers. As a small business, the next time someone on the team has to go on vacation, before that person goes, they can use the Dropbox plugin to put together a handover doc.
They can pull together information from all the projects they’ve been working on – the different change logs, all the information they know – that they want to provide to someone else. This is another great example of where you can create a project handover doc.
My last one, and probably my favourite, is this idea of a future-self handover. You can use the Dropbox plugin for ChatGPT to reference past work you’ve done – say, information from a client project from six months ago.
You can use this set-up to instantly refresh your memory: ‘What did I work on? What were the key tasks? What were the milestones? What were some of the pitfalls or gaps?’ Hopefully, that gives you some great ideas.
I want to talk about an important topic, Stephanie, that a lot of clients ask us about, which is security and especially permissions. A question I often get is, ‘Rahul, when I connect my Dropbox to ChatGPT, does ChatGPT now see everything within my Dropbox? Do I now instantly have access to all the information in my Dropbox team?’
The short answer is no. When you connect your Dropbox to ChatGPT, we make sure the Dropbox permissions and ACLs (access control lists) you set at source are honoured.
What that means is that when I search for a document or list folders or files, as you just saw in the demonstration, I’m only going to be able to see what I have access to. I’m not going to have access to that confidential finance folder that hasn’t been shared with me.
That’s really important to remember, and I know it’s an important question that comes up – it’s a concern all of us have, so hopefully that helps. Stephanie, from the OpenAI side, is there anything you want to share on that front?
Stephanie: One thing I’d also like everyone to know is that if you have a business or enterprise OpenAI account, we do not train on your data, and that’s stipulated. So you can rest assured that your data is secure and that we’re not training on it.
Rahul: Fantastic. Thank you, Stephanie. So, everybody, we’ve come to the end of our presentation.
Before we jump into Q&A, I do want to share a few takeaways. First of all, thank you again for joining over your lunch hour.
If there’s one thing I want you to take away from this presentation, it’s that the information you need to build great briefs and proposal decks – all that valuable knowledge – is available. It’s there in Dropbox.
Connecting the context that’s in Dropbox to ChatGPT can really help you not only save time but also put that information to work in a way that works better for you. Whether that’s preparing the first draft of a proposal or creating that future-self handover doc, there are so many opportunities to take advantage of the Dropbox plugin and ChatGPT.
So bring your knowledge and bring your judgement – AI is going to help make work easier for you. Again, thank you all for joining. Beth, back to you for questions.
Beth: Thanks so much. I know we’ve only got a few minutes, but let’s try to get some questions answered.
Thank you to everyone who has popped their questions in the chat. Let’s start with this question from Leo, who would like to know, ‘What happens when two of my source files disagree? The proposal says one start date and the customer notes say another.’
Rahul: That’s a great question – thank you for asking it. I want to go back to something Stephanie said earlier, which is to take an iterative approach.
You can use the Dropbox plugin and ChatGPT to compare these two files. My recommendation here, and I’d love to hear what Stephanie thinks about this, is to have ChatGPT create a table for you that explains the differences between these files.
Then you, as the human in the loop, so to speak, can review the results, take the right action and choose the right files. Stephanie, what do you think?
Stephanie: Yeah, I agree. Actually, the models are getting quite good.
Sometimes when information conflicts, ChatGPT will come back to me and say, ‘These two sources are conflicting. Which one should I pick?’ In that moment, I can say, ‘Use this as the source of truth when we’re working on this specific project.’
Beth: Fantastic. Let’s quickly do one more: once my one test works, how do I stop it being a one-off and make it part of how my business runs? Stephanie, do you want to take that?
Stephanie: The super quick answer is that I would create a skill. A skill is like a reusable workflow, and you can invoke it any time you like. And if you need to do that specific task on a specific day at a specific time, I would invoke that skill as part of a scheduled workflow so that you don’t have to think about running it.
Beth: Perfect, thank you so much. Sadly, that’s all we have time for today, but I hope you found the session useful.
I know there was so much information in it, and there are lots of links in the chat, so please do follow up on those. A massive thank you to our speakers, Stephanie and Rahul, for joining us today.
Thank you, and we’ll see you in the next one. Bye now.
Enterprise Nation has helped thousands of people start and grow their businesses. Enterprise Nation connects you to the resources and expertise to help you succeed.