Why we automated sales first (and why you probably shouldn't)

If you've been running a profitable business for years, you don't need another person telling you to "automate your sales funnel." Your customers already found you. They keep coming back. Sales isn't your problem.

Your problem is that quoting takes 45 minutes when it should take 5. Or that your margins on delivery are not where they need to be. Or invoices go out three days after delivery. Or every decision routes through you because nobody else has the context. Or you're hiring your fourth admin person this year just to keep up with paperwork.

So when someone like me shows up talking about AI automation, and all my blog posts are about sales pipelines and outreach workflows, I understand the skepticism. Sounds like I'm selling a hammer and calling everything a nail.

Here's the thing: we automated sales first because of where WE are as a business, not because it's the right answer for everyone. And understanding that difference is the entire point of what we actually do.

In this article:

• The real question isn't "what can AI do?"
• How we approach it: goals first, tools second
• A freight company example
• So why did we start with sales?
• Different stage, different answer

The real question isn't "what can AI do?"

Every week there's a new tool that can generate content, summarize meetings, or chat with your documents. The technology is real. But the question most businesses skip is simpler and more important: where in MY business will this have the biggest impact?

Most AI consultants start from what they know how to build. Chatbot specialists recommend chatbots. CRM consultants recommend CRM automation. Content agencies recommend AI-generated content. The answer always seems to be whatever the person in front of you is selling.

We think the starting point should be your business, not anyone's toolkit. And that means looking at the whole picture before deciding where to focus.

How we approach it: goals first, tools second

When we work with an established business, we don't walk in with a pre-built solution. We follow a process that starts with your goals and works toward the right answer, whatever that turns out to be.

Step 1: What are you trying to achieve?

Before we look at any process or workflow, we ask: what are your business goals for the next 12 months? Growing revenue by 30%? Reducing costs? Entering a new market? Improving customer retention? Handling more volume without hiring?

The goal determines where we look. A company trying to scale without adding headcount has different priorities than one trying to break into a new geography.

Step 2: Map your business

A key step is making sure you look everywhere in your business before deciding where to focus. Most owners have a gut feeling about where the biggest pain is. They're usually right. But sometimes the real bottleneck is somewhere they stopped noticing because it's been "the way we do things" for so long.

We keep it simple and re-use the Business Model Canvas as an intuitive visual tool to force looking in every place at least one.

Step 3: Measure how each area is actually performing

For each area of the business, we ask: what are the jobs that need to happen here, and how well are they being done today?

Not "what software do you use?" but "what are you actually trying to accomplish?" Your accounts person's job isn't "enter data into QuickBooks." The job is "make sure we get paid accurately within 30 days of delivery." The data entry is just one step in that job, and probably the most automatable one.

We score each job on two dimensions:

High importance + low satisfaction = where AI will have the biggest impact. Not the flashiest use case. The one that actually changes how your business operates day to day.

This approach draws on Clayton Christensen's Jobs to Be Done thinking, which is one of the best tools I've found for cutting through the noise of "what technology should we use?" and getting to "what outcome actually matters?"

Step 4: Identify the opportunities and build a roadmap

Once you have the gaps scored, the next step is prioritizing. Not every gap is worth closing with automation. Some are better solved by hiring, or by changing a process, or by doing nothing yet.

We look at each opportunity through the lens of the business goal from Step 1. If the goal is "handle 50% more volume without adding staff," then the roadmap focuses on the high-volume, repetitive jobs with the biggest gaps. If the goal is "reduce our collection cycle from 47 days to under 20," the roadmap starts with invoicing and payment follow-ups.

The deliverable isn't a strategy deck. It's a prioritized list of what to automate, in what order, with rough estimates of effort and impact, measured by the metrics that matter to the business.

Step 5: Build, measure, iterate

Then we build. Not everything at once, but starting with the highest-impact opportunity and working in short cycles. Build something, measure whether it's working, adjust, move to the next one.

This is the part most assessment-focused consultancies skip. They hand you a roadmap and wish you luck. We stick around to actually implement it, because the roadmap is only as good as the automation that comes out of it.

A freight company example

Let me make this concrete. Say you run a freight forwarding company. 30 people. Based in Dubai. Profitable for 12 years. You handle 50 shipments a day. You're thinking about AI because everyone's talking about it, but you're not sure where it fits.

Your goal: handle more volume without hiring more staff.

We map the business and score the jobs:

Look at the last row. Sales and marketing scores a zero gap. This company doesn't need a chatbot or an outreach sequence. Their customers come through referrals and relationships built over a decade.

But invoicing, status updates, and quoting? Three gaps of 3 each. That's where the hours are. That's where the errors happen. That's where automation turns a three-day invoice cycle into same-day and cuts collection time from 47 days to under 20.

The roadmap for this business starts with invoicing automation (highest impact, most structured data, quickest to implement), then moves to automated status notifications, then to quote generation. Sales automation doesn't even make the list.

If someone had walked in pitching a sales chatbot, they'd have been solving the wrong problem.

So why did we start with sales?

Because we're a startup. And for startups, the calculus is completely different.

When you're pre-revenue, you have exactly one question: does anyone actually want what you're building? Not "does it sound good?", does someone want it enough to pay for it?

Steve Blank wrote about this decades ago: stop building in a vacuum, get out and talk to customers. Rob Snyder, who writes The Physics of Startups, takes it further. He argues you don't interview customers first and sell later. You sell from day one, because selling is the research.

His key insight: real demand shows up when someone has a specific project, a deadline, options that aren't working, and they're actively looking for something better. Not polite interest. Not "sounds cool, let's stay in touch." Actual pull. As he puts it: "Complaining isn't switching."

For us at Black Hills Labs, we had zero revenue and a thesis that startups and SMEs in the Gulf needed automation help. The only way to test that was to talk to a lot of people, fast.

So we ran our own process. Our business goal: validate demand. The biggest gap on our own canvas: we had no way to reach potential customers at scale. So we built a sales machine. 12 automated workflows on a single server. Lead discovery scanning news sources for funding and expansion signals. Contact enrichment. AI-drafted outreach with a quality gate. Reply tracking. Follow-up scheduling.

The automation wasn't our product. It was our instrument for learning. Every reply, every open, every silence taught us something. A health tech founder replied within hours. A freight company never opened the email. A fintech founder wanted something we didn't offer yet. Each data point reshaped what we understood about our market.

We didn't automate sales because that's what we sell. We automated sales because, for a startup, that's where the biggest gap was between what mattered and how well we were doing it.

Different stage, different answer

The question "where should AI go first?" has exactly one honest answer: wherever the biggest gap is between what matters and how well it's working today.

For a startup with no customers, that gap is almost always sales and discovery. You need to find out if your business works, and the fastest way is to try selling it.

For an established business, the gap is usually somewhere in operations. Quoting. Invoicing. Status updates. Compliance documentation. The process that eats 20 hours a week and everyone just accepts as normal.

The canvas and the scoring aren't complicated. You could sketch this on a napkin over coffee. The hard part is resisting the temptation to jump straight to whatever tool is trending this month.

We start with your goals. We map your business. We measure what's working and what's not. Then we figure out, together, where automation will have the biggest impact on the metrics you actually care about. And then we build it.

If you're curious about where AI fits in your business, reach out. I'm based in Dubai and happy to walk through this over a call or a coffee.