“Deep Research” is one of the breakout agentic use cases of 2025.

So far this year, I’ve run well over a hundred deep research tasks and built three, custom Deep Research agents for clients.

I’ve now got a pretty good feel for the capabilities of the leading labs’ agents and compared them to my own (which are designed for narrow and specific workflows) in practise.

My conclusion: in the right circumstances, “rolling your own” agent - rather than using an generalist research agent - is the superior choice.

There’s an interesting dilemma in the space sciences known as the Incessant Obsolescence Postulate (IOP). It concerns interstellar travel. Essentially, the IOP says something like: any spacecraft you launch today risks being overtaken by a future one if technology improves fast enough. Hence the dilemma: launch now or wait?

Artificial Intelligence is no longer a sci-fi dream — it’s woven into the fabric of modern business. But here’s the catch: as many as 95% of generative AI projects fail, especially in sales and marketing.

So how do you ensure yours lands in the 5% that succeed? The secret lies in one deceptively simple question: Have you chosen the right problem to solve?

In this blog post, we take you through our three criteria to select a problem suitable for AI to solve, and give you the confidence to start successful project.

In this essay I’ll compare three real-world agentic AI platforms - n8n, Lindy and CrewAI. They neatly represent different points on the “agentic” spectrum: the agent-on-rails (n8n), specialised (“assistant-like”) agents (Lindy) and the high autonomy, multi-agent toolkit (CrewAI). 

I’ll focus on when to use (and when not to use) each of them.

In this post, we introduce a simple, 2x2 matrix that we use to triage and organise AI automation opportunities. It will help you to determine which ideas to pursue and how to get going.

Over Christmas, Nathan Labenz and our Data Science Director Will Hardman spent 4 hours doing an extended deep dive into Vision Language Models (VLMs) on Nathan’s awesome podcast, The Cognitive Revolution. The podcast can be found on The Cognitive Revolution website, or wherever you get your podcasts.

At its heart, a good AI strategy needs to answer four questions:

1. What does the journey look like? What should it do and what should it not do?

2. How should the organisation change to mitigate future risks and seize future opportunities?

3. What is the gap between the organisation’s capabilities today and the capabilities it will require to do this?

4. What impact will AI likely have upon the organisation, its stakeholders and its market ecosystem in the short, medium and long term?