Culture & Values

Results, not busywork

We judge success by what actually changes for the client, not by hours worked or tickets closed. Did support volumes drop. Did processes get faster. Did the project pay for itself.

In practice this means clear goals for every piece of work, tracking against real business measures, being upfront when something is not delivering value, and trusting people to organise their time sensibly as long as the work gets done.

Engineering before fashion

AI changes quickly. The basics still matter more.

We care about how systems really behave in the wild: architecture, trade-offs, limits, and how things fail. Not just wiring APIs together and hoping for the best.

That means thinking properly before choosing tools, reviewing code for long-term reliability, treating monitoring and testing as part of the job, and learning from things that go wrong instead of pretending they did not.

Practical beats perfect

Shipping something useful is better than polishing something that never leaves the building.

We start small, test with real users, and improve in steps. We design for sensible automation with human oversight, not fantasy systems that claim to handle everything.

So we release early, expect mistakes, build systems that fail safely, focus on the cases that matter most, and avoid over-engineering for rare edge cases.

Straight talking

No politics, no theatre.

We are direct with each other and with clients. Prices are clear. Problems are raised early. If something is not working, we say so.

That shows up in honest estimates that include risk, blunt feedback in reviews, transparent pricing, and being willing to admit when we were wrong or do not yet know the answer.

Real ownership

Projects do not get passed around.

The people who design them help build them. The people who build them help run them. Everyone is accountable for what ships and how it behaves afterwards.

Engineers speak to clients. Designers understand technical limits. Everyone pays attention to production systems. Responsibility does not stop at “done”.

Keep learning

No one stays current in AI by pretending they already know everything.

We value people who are curious, who test ideas, and who share what they learn. Saying “I don’t know, I’ll find out” is a strength, not a weakness.

That means time and budget for learning, regular sharing inside the team, trying new approaches on our own projects first, and treating failures as information, not embarrassment.

AI and machine learning are not diverse fields, and small consultancies feel that more than most. We do not pretend to have solved this.

What we are doing:

We remove names and universities from CVs at the first review.

We offer alternatives to live coding interviews, because they suit some people and unfairly exclude others.

We work remote-first, so good candidates are not limited to London.

Why this matters:

Teams with different backgrounds build better systems. You get better questions, fewer blind spots, and more robust decisions. Homogenous teams miss things.

If you think we should be doing more, tell us.

careers@thebotforge.io

Sustainable Pace This is a marathon, not a sprint. Occasional crunches happen (production incidents, critical deadlines), but they're exceptions, not the norm. 40-45 hours/week is typical.

No "Always-On" Culture Slack notifications off outside your working hours. Email doesn't need instant replies. If it's urgent, we'll call. Respect for personal time.

Automate where we can Production systems need monitoring, admin needs doing, records need keeping, code needs documenting. But we will always try to automate where its sensible to do so.