Introduction: From Video Agency to AI Platform

So yeah, thank you very much, Rian, for the introduction. My name's Andrew, and I'm the founder of a company called 60 Seconds, based in Bristol. We're a small but scrappy startup, and we've been running for 12 years.

Through that time, we've been doing video production, we've been doing lead generation, and ultimately we've been helping businesses go to market and get more awareness around their business and their brand.

Now, with the rise of AI, we've then pivoted to learn, well, take everything that we've learn and then build it into a platform where people can then tap into the knowledge that we've created over those 12 years.

So today I'm going to go through some of the learnings and how we've come up with this product and how it should save time. Now, also during this talk,

we're going to be talking about context engineering and the difference between normal prompt engineering and giving it more context. So hopefully I'm going to give you a real world use case of how this has helped us in our business and how it can help other people

The Bootstrap Reality: Wearing Every Hat

in their business too so you're running running for 12 years with next to no money i literally started the business at 21 years old i had three thousand pounds redundancy money and i spent it all on a google master's degree um so i literally had to make it work from day one or i would have

died we're 12 years on raised a little bit of money through clients and partners um but we've really had to make every penny count through that time i was doing most everything as the only employee but then i had to offset my hats to people that were better than me so i was able to do this successfully for the video creation side the campaign side the marketing side where

i wasn't able to do it successfully was the sales side and i always struggled to find someone who could confidently sell and do the follow -ups and do everything it takes to be successful in sales

Why Traditional Sales Teams Don’t Fit a Small Startup

Now I worked and looked at other organisations. How did they make it work with sales? So you typically need three people to make sales functions work.

You'd need someone that opens the doors. So this will be called a sales development representative, who is going to be messaging people on LinkedIn, sending cold emails.

And then you'd need someone to do the call itself. And this would be an account executive. Then you would need someone to manage the client when they come on board.

So as a small bootstrap business, I'm thinking, wow that's hundreds of thousands of pounds in salaries before I even get started so I didn't have that so I had to do it all myself and that meant that I learned every area of how to make it work because I didn't have 165 grand plus marketing spend I would literally have to use every penny

When Lead Gen Works Too Well: The Follow-Up Bottleneck

that we possibly could but I did figure out lead gen so we had one campaign very successful and in in the past year alone it generated a thousand calls now that sounds great right but a thousand calls comes with a lot of admin so of the thousand calls that got booked it was costing 60 pounds per call so we'd spent a lot we spent more than the salaries of people to come in and do it but those

calls are going into my calendar now that's eight to ten calls a day let's say and then you need to follow up you need to do multiple calls with each person you need to send proposals you need to update the crm so all of that stuff bogs me down with admin now when i say the math how do we math

that out i'm not very good at this side by side presentation uh but yeah each each follow -up each proposal let's say a good proposal would take two hours now i didn't have the time to do all of that so i was neglecting my pipeline so then i had a conversion problem not just well i had the lead

generation problem fix that now i've got a conversion problem because i've got too many calls to do so if you actually looked at what it would take per lead we worked out it was going to take 2025 hours over the course of a year for just the admin right i didn't want to work nine hours a day just on admin i loved the calls right i'm very good at the calls i enjoyed the calls but i sucked at the follow -up if you take the calls into account as well then we're talking 14 hours a day

So something's going to slip. It's either the cause or it's the admin.

If you had the traditional SDR model where you've got someone opening the door for you, it's typically a junior, right? And I've learned that in sales, people don't want to talk to the junior, they want to talk to the founder.

Using AI to Open Doors (Without Losing the Founder Voice)

So what we did was we actually cloned my profile and we used AI to go out and open the doors. And that's how we managed to solve the lead generation problem.

Context Engineering: Moving Beyond Prompt Engineering

So AI is great, but without the context it's not very good at helping you so for example if i put into a prompt write a proposal for the call i've just had and i put that into something like a chat gpt or a claude it would ask me for the context that i'm missing right so who is it and then i'd be spending the

The Missing Ingredient: Call Transcripts as High-Value Context

time giving it the context to get to the next stage so this was the problem that i set out to solve how do we give it the context i think something that most businesses are currently sleeping on is the context of the call transcript so most everyone has like a meeting recorder that

joins every sales call and by having that transcript that's the context that you can feed in to power up like the clots and the chat gpts so once you give it the context everything changes so the meeting recording comes back in and then instead of feeding the transcript directly in we've created a way to just generate a proposal so what this process is now doing is going out

Pulling Context From Everywhere: CRM, Emails, Calendar, and Past Proposals

and finding the context from the call it's looking at past calls it's looking at the crm it's looking at the emails the calendar it's getting everything that a sales rep probably knows and it's learning it and then it's going to go and find all of our previous proposals

And then with those proposals, it's going to come up with a tailored proposal or first draft, let's say, that we can then send to the prospect. So I think it's very important just to spend the time looking at what context do I need to give AI and where is that context sitting?

So who here uses a meeting transcriber? Amazing. So is that working with your CRM or do you have to copy it over? like

This is the problem with live demos Great.

From Transcript to Proposal: A Practical Workflow

So we've got the proposal back. So this proposal here is taking all the context. So we've got what we've discussed

And it breaks down the challenges the the pricing everything that I've verbally said on the call and it's looked at the Package that we've presented and then it's come up with a full proposal that we can send over to the client or modify

So this was a big step up and this is where I really saved a lot of time is doing the calls back to back but knowing that the context was there so it's very quick for me to then take the next action then I thought how can we make it

The Next Phase: AI as a Proactive Teammate

one step further instead of me having to go in and prompt or to generate the proposals could it be that we actively look at what's happening or AI is watching and it's doing it in the background and it's then helping you as a proactive teammate and I think that's going to be the real unlock and the next phase of the evolution of AI is it's the proactive teammate not the reactive

waiting for you teammate right so we've set it up with a connection into slack it's looking at the calls it's waking up in the night and it's looking at the CRM and it's saying hey Andrew what we need to do today is this right so my brief is given to me during the day I can then know what actions are going to move things forward and on top of that AI's already done some of the work and I've I've got a human -in -the -loop approach where I can then approve it and then take it to the next level.

Human-in-the-Loop Follow-Ups That Keep the Pipeline Moving

So I can approve this, and then I get given a follow -up message. Now, I really am terrible at the follow -up process, so having AI actively helping me has been an absolute game changer. So being able to just focus on the calls and then AI helping has absolutely wiped off how much time I spend doing admin.

And that admin isn't enjoyable, and it's something that should and can can be automated. And I think this is where AI can be super helpful in pushing the pipeline forward and really helping you be more productive.

Scaling AI Across the Team

In my business right now, we've given AI to most every team member. And they're finding that they're 5x -ing their output purely through looking at the processes that they've got in place

and then coming up with skills. And that's more context that is being given to every single team member.

“Skills” as Reusable Commands for Repeatable Work

So when I say a skill, it's typically like a forward slash command. and this is a big pre -prompt and it could be write a proposal like we've seen and that proposal prompt is then pasted in forward slash proposal and then you give it the context right so it's

that's the starting point that i really think people are should be using more and more is the skills and if you haven't heard of skills i think it's just watching a youtube video or trying it it yourself and if you then say what do i do over and over again what is the repeatable thing in my business that you might not enjoy but would be a big time saver which means that you can

10x your output or 5x your output so have a look at everything you're doing put a forward slash and then say create me a skill for this and i think as soon as you start using that you'll you'll start to see the big level up in what AI can do for you. Then you can go into these processes that I'm showing you now to make it more proactive and learning.

So hands up in the room, who's actually used the skill so far? Okay, who's going to look up skills after this? Fantastic.

Conclusion and Next Steps

I really do think, yes, skills will be the next step up for what you're trying to do.

And if anyone wants to find out how to use skills or wants to find out more about this, then feel free to connect with me on LinkedIn and I'll be more than happy to help you.