Connor Heyward-Fox

before we built scout.io i was spending entire evenings researching companies and writing cold emails: - pull a list from apollo - open 50 tabs - check every website - figure out what to say - write something half-decent - repeat. by lead 30 i was cutting corners. scout does the part i couldn't keep up with; it researches every company properly, scores them on fit and timing, then writes outreach based on what it actually found. not templates with a name swapped in - emails that reference real intel about their business. that’s the gap we’re building for. we don’t think outbound is broken because people don’t try hard enough - it’s broken because doing it well doesn’t scale manually. scout makes 'doing it properly' the default, not the exception. we’re just getting started, and we’re shipping fast. if you’re doing outbound and want to see what this looks like in practice, give it a spin and let me know what you think - free tier is available now. (link to product in the comments below).

a 90-day test compared ai-only outreach against ai-plus-human hybrid. - ai-only booked 847 meetings. - hybrid booked 312. the hybrid generated 2.3x more revenue. fewer meetings. more money. this is the essence of why we built scout the way we did: every tool in the market is optimising for volume over quality. we have found quality research and personalisation outperforms raw volume across the board. three weeks into our own outreach and the pattern is already showing; the replies we're getting are from leads the research layer scored highest. he volume didn't reach them - the intelligence did.

@asaio87 Had to do some research into compliance and security myself to make sure I covered everything when auditing my own software - gotta cover your ass.

Most AI coding agents will not implement more than basic security when building an app or a website. It's your job as the professional using the coding agent to tell it, what to implement and how to implement properly. Thing is, most of these companies advertise the coding agents as software that can transform any regular person into a powerful developer. How are these people supposed to know what to ask the coding agent?

at this point non ai adopters are just waiting to be out sped, and replaced by other humans who are happy using it.

AI isn’t replacing developers. It’s replacing developers who don’t use AI. The gap between AI-native and non-AI devs will be 10x by 2027. Choose your side.

at least 1B tokens needed

@iamliamsheridan we're in this experimental process now; always trying new formats and approaches to see what returns the most results

"we tried cold email and it didn't work" no. you tried one email. to one list. with one angle. and gave up. that is not cold email failing. that is one test failing. the companies booking 30+ calls a month ran 50 tests before finding the one that works.

@leonabboud yes trying to remind myself this is a tech bubble and the average person doesn't know what claude is - my nan isn't going to launch a rival AI company anytime soon

Reminder that X is not a good representation of AI adoption. X has the super users of AI. I have entrepreneur friends who are just switching to Claude, 3 months after everyone here on X. Whenever you think "Oh everyone knows how to vibecode an app" no, 99% of people don't.

@signulll successful products are built on significant pain points

“you've got to start with the customer experience and work backwards to the technology. you can't start with the technology and try to figure out where you're going to try to sell it.” this is the fundamental problem with almost all of ai today. the founders who'll win are the ones who identify specific, painful, recurring workflows & make them vanish.

@ivanburazin LLMs are just the launchpad for the next phase of Intelligent technology

The entire world thinks LLMs are the final form of artificial intelligence. But they are more of a point-in-time solution. Powerful, but not the end state. We're already seeing the growth curve plateau. OpenAI raised $122B at $850ish billion, and the terms aren't great. There's a financial and physical endpoint to scaling LLMs. You can't keep 10x'ing compute costs forever. Society will be forced to find something else. And when forced financially, people find solutions. LLMs still have massive value to unlock. But they're not the internet-level standard. Something else is coming.

@TTrimoreau overthinking is a double edged sword for me

As a founder What drains your energy the most? -context switching -distractions -lack of results -too many ideas -overthinking 👇

@asaio87 it all depends on your prompting; a lot of models have the capabilities, it comes down to the prompting. The level of specificity is key for quality results.

Building a landing page is a trivial task now with AI. That is if you want to be very bad. Even the most basic thing you can do with AI requires expert knowledge if you want it done right. Nobody is going anywhere. But somebody is making a lot of money on this hype

@Jashanx_gill humans just went up in value

AI-generated content is everywhere. Authentic voice is now a "Luxury Good."

finishing up a linkedin optimisation this week in response to feedback from early users. built a chrome extension that scrapes linkedin profiles so scout can pull more data into its research layer and quick scrape new potential leads. click on a profile, it auto-scrapes the page, identifies an existing company and takes you straight there for further intel. feeds back into scout as a constant source of fresh leads to review, with the option to promote to a full pipeline lead if it's a good match. linkedin is tricky because they crack down hard on automation. the middle ground is you can capture information from pages you're already visiting but you can't automate the browsing itself. so the extension grabs data while you're naturally on the page. Secondly wiring in linkedin dm sending directly inside scout via heyreach and lemlist api integration so you can review and send messages without switching tabs. small thing but removes a lot of friction that users were having problems. early feedback was big on optimising the linkedin pipeline so this should make things much smoother for users.

@KevinSzabo14 if you can make it through the silent film days eventually you'll have surround sound

Nobody buys from strangers. Post consistently for 90 days and watch how differently people respond to you. Trust is the only sales strategy that compounds. What’s your biggest challenge with staying consistent?

@troyaitken_ this is exactly it. we spent months on the research and scoring layer before we even touched email writing. most tools start with the message and work backwards. if you don't know why someone is worth contacting the writing doesn't matter.

Every B2B founder using AI AGENTS for cold email is still struggling with inconsistent pipeline because ai is being used on the wrong layer writing better emails doesn’t fix: - bad data - weak targeting - poor deliverability it just scales the problem faster what actually works is upstream, most people aren’t even looking there

@srishticodes tried it - ended up with an online marketplace app for renting superhero costumes for cats. it must be onto something.

Has anyone tried this one killer prompt?

@johnennis we hit the same thing building our scoring system. hard numeric thresholds produced weird results but describing what a good lead looks like and letting the agent weigh signals on its own got it closer.

Here is a really interesting thing that I have learned working with AI agents It is very important to leverage the intelligence of the agent and let it make its own qualitative decisions If you give a quantitative metric to an agent, you will get extremely bizarre behavior as the agent learns to game the metric It’s kind of strange, but you get it better results by just letting go and letting the agent make up its own mind after you have described qualitatively what you are trying to do At the end, of course you still need to step in and be the final decision-maker before something goes into production But you will be much closer if you learned to trust the agent It’s kind of similar to managing people like that

@johnennis I feel the same way when i get lost in a deep middle earth lore hole

I think one of the biggest challenges when it comes to going hard into using AI is loneliness I am learning all these awesome things and becoming super capable But the set of people that I can really talk to about it is very small Is anyone else having this experience?

@atmoio this is the gap atm, builders are assuming a great product = large interest. Quality marketing is 100% the most painful learning curve on the vibe-coding path and one that you need to learn manually.

there’s a sort of permanent valley of despair i’m seeing a lot of non-technical people who are building software fall into. they have a wonderful time vibecoding the thing. but when they reach the point where everything is functionally correct, they have no idea what to do next. they have no idea how to go from having a product to actually getting people to use or pay for it. it’s not something an agent can help with because it’s the part of the job that requires you to go soak yourself in present, dynamic reality and learn how to change yourself and your product to satisfy a market need. it’s a very painful and slow process many builders get out of the way in their twenties. and now you have a bunch of middle aged men with kids and a mortgage trying to navigate this, and it’s not really working out too well for them. it’s why everyone is talking about what they’re building but no one is shipping.

@levie the scheduling example is great; we built an ai system for a recruitment client that automated candidate screening. didn't replace anyone, they just took on more clients. the bottleneck moved downstream.

There are far more categories where AI agents making things more efficient will induce demand for that skill than spaces where agents eliminate the work. This is why the AI jobs predictions will not play out as advertised. AI making it easy to produce more code will mean we start to apply code to far more parts of our businesses. We will build automation and software for things that wouldn’t have made sense before. Marketing automation, client onboarding, modernizing old systems, doing far more research on existing data, and more. More engineers. Far more software will mean vastly more security risks. This will mean far more people thinking through system security, compliance, and governance. This used to be primarily manual and only large companies could afford this work. AI will make it so more companies care about this (and maybe can do something about it), causing more security roles. AI will also lower the cost of a bunch of previously relatively niche or harder to access categories of work. Companies will now be doing 10X more with video and graphics, and will need people to manage that work. More media. We’re going to have a near unlimited set of legal challenges in a world of AI as AI helps write even more bespoke and complicated legal docs. More lawyers. Then there’s the impact of AI efficiency on non-office worker jobs. Talked to a customer that said they’re going to make scheduling medical appointments and getting referrals so efficient the next problem will be there will be no booking time slots available. More healthcare. Many industries will have this same dynamic play out. The examples are endless once you start to think through second order effects of agents making work more efficient.