Hey friends,

Quick one this week. I am writing this with a bag half packed by the door, because in a few hours I am on a very long flight to Geneva. The United Nations is calling. I will tell you the whole story at the bottom.

But before I go, I want to leave you with something worth more than the trip. It is the story of three businesses, and how one fairly simple change to the way they quote and propose grew each of them. One of them lifted its conversion rate by fourteen percent, which does not sound like much until you see the volume it sits on top of.

Here is the thread that connects all three.

Most people look at automation through a single lens. Cost. Fewer hours, fewer people, lower overhead. And when that is the only lens you use, you end up measuring a change like this in hours saved, which badly undersells what it can actually do to a business.

The cost saving is real. I am not going to pretend it isn't. But it is the least interesting thing that happens here. And if saving money is all you are looking for, you will aim too low and stop too early.

Here is what actually transforms a business.

The first is speed of revenue. A faster quote is a shorter sales cycle. When a customer gets a response while they are still thinking about the problem, instead of three days later when they have already moved on, you close more of them. You are not just spending less, you are making more, and making it sooner.

The second is harder to see, and it matters more. When the work of servicing a sale is manual, that manual process quietly becomes the ceiling on how much you can grow. You can win all the demand you like, but if every quote needs a person to build it by hand, your growth is capped by how many quotes those people can physically produce. Take that constraint away and the ceiling lifts.

That is the shift. From "how much can we cut" to "how much faster can we sell, and how much bigger can we get." Three businesses showed me exactly what that looks like.

I want to walk through all three, because together they show something important. This is not one solution you drop on top of every business. The right level of automation depends entirely on how your process actually works. We built at three different levels, and each one was correct for the business it was built for.

Level One: The Machine Runs the Whole Quote

The first was a manufacturer. They sell products to other businesses, and the pattern was always the same. A quote request would come in by email. A salesperson would pick it up, pull the product list, find the customer, create an entry in the CRM, work out the right pricing for that particular customer, go to the freight provider to compare shipping, add the margin, calculate the final cost, build the quote, and send it out. Then wait for it to be signed and returned.

Every step was doable. Together they took about three hours of real work, and in practice the quote often went out two or three days later.

Because the product data already existed and the freight providers had APIs we could connect to, we were able to hand the whole thing to an agent. Now a request comes in, in whatever format the customer sends it, and the agent reads it, finds the customer in the CRM, pulls their address, prices the product, adds the margin, works out the freight, and produces the quote. Three hours became three minutes. The customer often gets a response almost immediately.

Their conversion rate went up by fourteen percent. That sounds like a small number until you apply it to their volume, at which point it is not small at all. But the part that stayed with me came later. When they launched a new product line into a new market, they scaled it up quickly using the same agent, instead of standing up a whole new team to handle the extra work. The system they already had simply absorbed it.

Level Two: A Short Voice Note, and the Rest Builds Itself

The second business could not be fully automated, and that was the point. They are in the trades. Their quotes depend on sending someone out to inspect the job, because you cannot price the work properly without a human eye on it first. There is genuine judgement in the loop.

So we did not try to remove the human. We removed everything around them.

Now, when a request comes in, the agent gets everything ready. The estimator looks at the job and makes the call that only they can make. This service, this many hours, these are the specifics. Then they record a short voice note, ten or twenty seconds, and that is enough. The agent takes it from there, assembles the full quote, and sends it back to the estimator to review before it goes out.

What used to be a slow, effortful job, where some things were considered properly and some things got missed, now runs cleanly. The human still owns the judgement. They just stopped doing the assembly.

Level Three: The Human Stays at the Wheel

The third was a professional services firm, and their situation is one a lot of you will recognise. Their quoting is not a product with a price. It is a proposal. It depends on understanding what the client actually needs, and that changes every time.

Their process was to receive a requirements document, meet the client, understand and write up the requirements, do some research, and then build a proposal that laid out their services and their fees. Start to finish, at least a couple of days of work.

Here we did not build an agent at all. We built a Claude skill, connected to their previous proposals, so it understood their service offerings and how they price. Now a client enquiry comes in and the skill tells the salesperson what to ask. They have the conversation, the skill picks up the transcript, and then it works with them to produce the proposal.

This one needs more from the human, not less. Someone still has to sit with it for fifteen or twenty minutes. But at the end of those twenty minutes they have a finished proposal, where before they had two days of work ahead of them.

And honestly, the time was not even the main win. The bigger change was quality and consistency. The things that used to get missed, the questions and objections that should have been handled up front, are now built into the process. The proposals come out sharper, and the back and forth afterward drops right off.

Three businesses. Different sales processes. Three completely different levels of automation. And that is exactly the point.

The mistake I see people make, the real Dumb Monkey move, is buying the word before they understand their own process. They hear "agent" and picture a machine that does everything, then get frustrated when their business does not fit that picture. But "agent" is not one thing. Full automation, a human in the loop, a human led by a skill, these are all valid. The right one depends entirely on where the real judgement in your process sits.

You do not start with the tool. You start by understanding what your process actually is and what it actually needs. Then you pick the level.

If you do one thing this week, map out how a quote or a proposal actually moves through your business, from the moment a request lands to the moment it is signed. Mark every step where a person is doing something a machine could do, and every step where real human judgement is genuinely required. That single map tells you which of the three levels you are looking at, and it is almost always the fastest place in a business to see a return.

If you have not heard, Agents for Humanity won the national Innovation Factory competition. I am still not sure it has fully landed for me yet.

And it came with something I genuinely did not see coming. That is why the bag is by the door. I am heading to Geneva to represent Australia on the main stage at the UN's AI for Good Global Summit. I get to stand up in a room full of people who are using this technology to do real, important things, and talk about the work.

So my next note to you will come from there.

Start Up Daily Article: https://www.startupdaily.net/topic/artificial-intelligence-machine-learning/enterprise-money-wins-un-ai-for-good-awar/

UN AI For Good Summit: https://aiforgood.itu.int/how-enterprise-monkey-is-rethinking-ai-for-development-through-agents-for-humanity/

It is easy to get pulled into the language of all this. Agents, automation, skills, the constant pressure to be seen doing something with AI. But the businesses in these three stories did not win because they adopted a buzzword, and they did not win by counting the hours they saved. They won because they looked hard at one slow, painful, revenue-affecting process and fixed it at the right level.

Quoting and proposals might be the most obvious place to start. It is the exact moment a customer is ready to hand you money, and in most businesses it is still slow, manual, and inconsistent. That is a strange thing to leave untouched.

You do not need to automate everything. You need to find the one process that is quietly costing you deals, and start there.

See you next week,

— Aamir

🎧 Listen to the Podcast Episode on: Spotify | Apple Podcasts | YouTube

📘 Book: The CEO Who Mocked AI (Until It Made Him Millions) by Aamir Qutub

🚀 Dumb Monkey AI Academy

📱 Dumb Monkey AI Academy App: Apple | Android