Discovery phases and a lot of AI chat – notes from 3rd December 2025
Posted by Paul Silver
On the 3rd December, 11 freelancers met in the Battle of Trafalgar pub to talk tech, projects and handling clients. This is some of what we talked about:
- Using discovery phases to push back against fixed price projects
- Building apps that weren’t affordable in the pre-AI/LLM era
- Cajoling AI tools into doing what you want
- Using Perplexity to help accelerate learning a new framework
- Using AI to make tools to fix your own business problems
- Lovable (AI tool) for rapid prototyping and Cicero – an alternative to Lovable with good design output
- Getting back into Laravel (and being delighted about it)
- Ebay making it free to list items
- Handling tenet and role based auth when using AI tools to make projects
- Artisan Makers Fair
- Upcoming AI conference in London
- Silicon Peel
- Ghostdance by Chris Limb
- Working out what you want to do next
- Monorail!
- Pacing yourself as you get older
Discovery phases
This is particularly relevant for programmers, but I’m sure it happens for people doing other kinds of complex work too. You have a client who insists on having a fixed price for a project, but there are a lot of unknowns in the project and you can’t predict how long they will take.
One way to handle this is to quote a high amount to cover your potential high workload. Another is selling the client on a “discovery phase.”
A discovery phase is time to explore the problem you are trying to solve. It might be letting you put together a set of code to talk to an API you are not familiar with, or to get some experience with a type of database you haven’t worked with before, or have a proper look through the codebase that the client already has.
The discovery phase is independent from the main work. You do it so you can make a more accurate quote for the client, which is what they want, without completely stuffing yourself if it turns out their problem will take much longer to tackle than you initially thought. It is not there to completely solve their problem, it is to help you get a handle on the depth of the task of fixing it.
When talking to a client who wants a fixed price or small range for their project where there is a bunch of unknowns, you say to get a handle on the project you need a discovery phase of x time and/or y cost in order for you to be able to give them a proper estimate later.
Your work involves:
- Research into whatever the unknown parts are
- An estimate for the project they want
- It may involve: a write up of your findings which they can use if they want to hire someone else
The discovery phase is for you as well as them.
If it turns out you’re going to hate the project now you know more, be honest that you do not want to do it. This is where a write up can be useful – you can pass onto them knowledge that another programmer will need, and the client has got something useful for the money they’ve paid for the discovery phase.