We sold our first AI agent to a legacy industry–now we're stuck. Help us Advice? | Hacker News Reader

We sold our first AI agent to a legacy industry–now we're stuck. Help us Advice?

13 pointsgoldmanX19 comments

My co-founder and I recently launched an AI startup building custom agents for manufacturers in old-school industries (think: industrial suppliers, factory ops, regional distributors, etc.).

The good news: we shipped a real solution, got our first paying customer, and even sent out our first invoice! The bad news: getting our second and third customers has been way harder than we expected.

We're running into what feels like a canal-product fit issue — our product works, and solves real pain, but the people we’re targeting don't check emails, barely pick up their phones, and often aren’t even on LinkedIn. It’s like trying to sell software to 1997.

We’re hustling hard — cold calling, showing up at trade shows, chasing introductions — but it's slow and uncertain.

If you've navigated early traction in a legacy industry, how did you break through the wall after the first win? Would love to hear any advice, strategies, or similar war stories.

Thanks in advance

You don't sell to your customers in those industries - you sell to the industry associations from whom your customers get operational/industry advice. Partner with them to promote your product to the customers.

Agree with this. Most of your customers are going to pay consulting firms that specialize in finding and adapting software solutions for them. Find a couple of those to partner with and they will help you sell your product.

Of course, they'll want their own slice of the pie, but that's the nature of the business you've gotten into.

The key is to remember that you're an Outsider, an unknown quantity, and so you present an inherent risk from the outset (especially in the genAI field, which brings a lot of additional suspicion all on its own). What you need is an advocate who is already established and respected in the industry you're targeting. That gives you a bit of "social proof".

You have one customer... assuming that they are happy with your offerings, perhaps you can get them to be your advocate? Offer them a great deal in exchange for being able to use their name in marketing materials, to get them to write up an article about your product in a trade journal, to talk you up at trade shows, etc.

With good word of mouth, one customer becomes ten.

not when the next customer is the competitor of the first one.

Please re-read the word "Good" and apply it in context to the remainder of the sentence

What you said may be valid for other sectors, but our customers are in fierce competition, so they do not recommend us to each other. Another problem is that when we go to new customers, they think we are 'agents' of the rival company, in short, they have difficulty trusting us.

"good word of mouth"

I might be giving my age away, but it’s like we’re back in the days of fax machines and Rolodexes. Most of the industrial suppliers and ops folks we talk to aren’t hanging out on LinkedIn or tech forums. They trust face-to-face chats, trade magazines, and the vendors they’ve known for years. Cold emails go nowhere, old-school phone calls or even showing up in person actually work better. These buyers move slower, want to see you’re serious, and often need help understanding what the tech even does. So we’ve had to meet them where they are, usually offline.

Congrats on the first sale,that’s huge! Legacy industries are brutal because you’re not just selling tech; you’re selling a behavior change. A few things that worked for me in similar spaces:

Trade magazines & physical mailers (seriously). Decision-makers in these industries often ignore digital noise but read industry pubs. A well-placed ad or "How Similar Company Cut Costs by 30%" case study can land meetings.

Embed yourself in their language. One client doubled replies by swapping "AI agent" with "24/7 digital foreman" in outreach terms that resonate with their mental models.

Leverage your first customer. Film a 90-second testimonial (phone video is fine) of them saying "I was skeptical, but…"use it as your #1 outreach asset.

If you’d like, I’d be happy to audit your messaging for free. I’ve helped other hardware/AI hybrids reframe their pitch to cut through the skepticism (e.g., one company replaced their ‘dashboard’ with a ‘command center’ and saw a 40% bump in demos). No strings just hate seeing solid tech get stuck in positioning purgatory!

Thank you for these useful advices, I will share them with my team. I may want to ask you about this in the future. Can you leave your LinkedIn link?

My 0.02c

- carry out some deep market research and analysis, work out the best niches, garner insights into buyer behavior

- content marketing still works but if you use AI effectively becomes extremely efficient and powerful. For example, you could build an MCP server that gives you superpowers with keyword research and SERP analysis tools. There's a growing body of knowledge in this area, sometimes called vibe marketing

Not even remotely an expert and this advice is from second hand knowledge.

Hold in there. Sales cycles are long. Figure out conferences where people go to. Go there. Drop into offices and ask for their time. Find out people who know people.

What did you do different, from getting your first customer (which was successful) to getting the second now? Either way, keep creating value for that first customer, that'll be much insightful

Find people willing to help on the sales team of your first customer. They will have way more experience and established networks into the rest of the ecosystem.

How did you find this niche?

My rule of thumb is a problem presents its own solution. But I’m not sure how it applies to you without knowing more about your product.

Have you looked into tradeshows? My company (large industrial equipment) still heavily does them. My company is not sold on AI, so stick to a quick demo of it's utility and stay away from 'buzz words' when talking to them. When I bring up 'Agent Mode' people roll their eyes and what we do is mostly software these days! Target the young people in those companies who are frustrated but haven't bailed yet. If it works and people can see it they'll buy in if it saves them time. Especially in the world of 'we can cut 10% of a workforce and still ship products'. Best of luck! Open invitation to reply with a link or repo!

Ask your first customer for referrals.

Build another product.

Microsoft didn't stop with basic.

Diversify.

There is no rulebook.

Good luck.

Reply on news.ycombinator.com