Practice 03
Ventures
For owners, corporates and partners with something new to build.
In short
HAM Agency is a venture accelerator. It takes an idea from nothing to a working product with customers, tests it cheaply before building it properly, and stays on while it finds its feet. AI does the heavy lifting and a named person is accountable for every decision.
Who it's for
Owners who want a new line that does not exist yet. Corporates who need it built outside the machine that would slow it down. Partners with the distribution and nothing to put through it.
Also teams buried in work a machine should be doing.
The problems we fix
- An idea that has been talked about for two years and never had a week of real work spent on it.
- A corporate innovation process that produces decks and no products.
- A build with nobody accountable. A development shop that delivers what was asked for rather than what was needed.
- Distribution with nothing to sell, or a product with no route to a buyer.
- AI bought as a subscription and never wired into anything that matters.
- Manual work repeated daily by people who were hired to do something else.
What we build
- The concept: the proposition, the buyer, the economics and the honest reason it might not work.
- The validation: the cheapest test that would tell you to stop, run before the expensive build.
- The product or the platform, built to work rather than to demo.
- Automations and AI wired into the actual operation, with a person accountable at every decision point.
- Dashboards that show the thing running, not a screenshot of last month.
- The handover or the stay-on, named in the scope at the start.
The validation method
For consumer goods and services businesses we use our own validation method. It was developed in consulting, has been tested over seven years and has been run with 350+ start-ups.
It answers one question before the money is spent: will people buy this.
- Find the gap. We start with retailers and buyers and ask what they cannot currently get.
- Test it. The concept goes in front of real customers under a stand-in brand. The pass mark is agreed before the test starts. A test takes between two weeks and two months.
- Stop or go. If it fails, it stops there and the cost has been a few weeks. If it passes, there is evidence to act on.
- Take the proof to market. A concept that passes goes to buyers, investors or launch with the results behind it.
How it runs
A scoping call, then a written scope with numbered deliverables, exclusions, two rounds of amends and a fixed price. Work starts when both sides have signed it.
We validate before we build, and we will tell you to stop if the test says stop. That conversation is cheaper at week two than at month six.
Equity, where it applies, is on top of a fee and papered before a line is written.
Questions
What do you mean by ventures?
New products and new businesses. We take an idea from nothing to a working product with customers, test it cheaply before building it properly, and stay on while it finds its feet.
Are you a venture studio?
No. We do not run a fund or a stable of start-ups. We work as a venture accelerator: we build and prove new ventures with the owners, corporates and partners who will run them.
Do you take equity?
Sometimes, and only on top of a fee, never instead of one. The terms are written down before anything is built. Where we back something ourselves, we say so plainly wherever it appears.
What does human in the loop mean in practice?
AI does the heavy lift: the reading, the drafting, the pattern finding, the repetitive build. A named person is accountable for every decision that reaches you, and signs off every claim. Where AI wrote something, it has been read by a person before you see it.
Where should a small business start with automation?
With the task that is done most often and the same way every time, usually in admin rather than in the thing you sell. Automate one of those end to end, measure the hours it returns, then do the next. Starting with the most complicated process is the common mistake.
Do you hand it over or stay on?
Either, named in the scope. A build that is handed over at launch usually stops working within a quarter, so we price a stay-on period by default and you can decline it.