Recently, I’ve shared my observations of an AI assistant that is slowly coalescing around me and the work I do. If I borrow from a more science-fiction trope, then ATLAS — the name I’ve given my collaborator — doesn’t have corporeal form yet. It exists more in spirit: a higher form of capability or intelligence. But as the technology advances closer to my vision, it is manifesting in several distinct ways.
Let me take you on that journey — how ATLAS currently supports me, what consistent value it delivers, and what’s still missing in its evolution from tool to true thinking partner.
1. Capture → Organise
Capture and Organise Thoughts
Notes apps are great for quickly logging an idea, whether typed out or dictated via voice. But AI tools can now go further — not only capturing input, but also interpreting, tagging, and transforming it. Depending on how you configure the system prompt, your assistant can do anything from simply logging your thought to diving into research or sketching a plan.
With voice-to-text features, I find this especially helpful when inspiration strikes in less convenient moments — on a mountain bike ride or when I wake in the night. I capture the week’s worth of thoughts in a single chat under the project title Thought Machine, and by week’s end, the LLM summarises my ideas, offering patterns and reflections I might have missed.
2. Collaborate → Develop
Researcher
Once an idea is captured, the next step is often exploration. AI models now have access to real-time search integrations and tools that allow them to move beyond their training cut-off — making them viable replacements for traditional search engines in many cases. They help me scan emerging topics, surface niche references, or generate briefings in a fraction of the time.
Ideation and Concept Development
Engaging an LLM in conversation around a half-baked idea is one of the best ways I’ve found to break creative blocks. Their strength in associative thinking makes them particularly good at connecting dots — spotting weak spots, amplifying strong signals, and proposing novel directions.
I often use Anthropic’s Claude, where I create “projects” — chat collections with shared memory. This enables continuity: a new thread can access previous context and decisions. Many of these apps also include a canvas feature — a side panel where documents evolve alongside the conversation. These artefacts can be saved, referenced, and versioned as the project matures.
Need a clean summary for a pitch or review? A single prompt pulls it together. The assistant not only helps develop the ideas — it builds the structure that holds them.
💬 Want to dive deeper into why AI is so effective at ideation?
I explore the concept of AI as a “Third Brain” — a unique cognitive partner that complements both our conscious thinking and external knowledge systems — in my upcoming piece: Why AI Excels at Ideation in Knowledge Work.
3. Challenge → Reflect
Sounding Board & Inspiration
Ideation in a vacuum rarely yields the best outcomes. A good assistant doesn’t just agree — it challenges. As ideas take shape, I often ask my assistant to explore various “what if” scenarios, take on different personas, or critique my assumptions. I regularly use a trio of roles to interrogate an idea:
The Optimist – spots the potential
The Pessimist – surfaces risks
The Challenger – forces deeper thinking
This dynamic helps me stress-test concepts, avoid blind spots, and strengthen the overall case. When I do need validation, it’s great at articulating why something works — often in neat, tabular form — which boosts confidence and helps guide prioritisation.
Here's a recent example exploring how I could bring novelty to my writing, including AI. I had the rough concept for ATLAS already, so I was challenging its fit to the vision I have:
4. Construct → Execute
Manage Projects and Actions
Like many, I sometimes struggle with moving from concept to execution. AI helps me bridge that gap — breaking a broad initiative into logical, achievable next steps. The assistant lays out the sequence of work in plain terms, showing whether the plan aligns with my capacity and constraints.
I recently used this approach to plan a year of monthly family events — combining ideation, logistics, and execution in one ongoing dialogue. The assistant nudged me forward in quiet moments, helping make progress without the pressure of a full-on planning session.
💬 Here’s how part of my chat with Claude went:
“Ah yes, that makes more sense! September will likely have better weather for London exploring, and a cosy cinema night works well for December. Here’s the updated calendar:”
January – Board Game Café ✓
February – Sushi Making
March – Eden Project Family Day
April – Highclere Castle ✓
May – Watercress Festival ✓
June – Comedy Club Night
July – Concert ✓
August – Garden Tour & Tea
September – London Day Out (Tate/NHM)
October – Devil Wears Prada Show ✓
November – Nice meal out (Birthday)
December – Cinema Evening
Next steps:
Research comedy clubs
Look into garden venues
Plan London day trip
Shortlist indie cinemas for December
Strategist
Between ideation and execution, there’s often a need for a strategic intermediary — someone (or something) to shape your intent into a guiding plan. AI is increasingly capable here too.
Whether I’m mapping a digital presence or framing seasonal goals, I often prompt my assistant to help construct a framework. For example:
Digital Strategy: Balancing presence across LinkedIn, Substack, and Bluesky — with distinct tones and topic clusters.
Seasonal Planning: Integrating personal goals with work commitments, constraints, and broader life rhythms.
With the right prompts (and context from previous outputs), the assistant can generate strategic overviews like this:
Big Theme:
“Building environments, systems, or personal practices for flourishing in a modern world shaped (and sometimes warped) by technology.”
What ties it all together?
Across these roles — from ideator to organiser, strategist to editor — a few consistent patterns emerge in how I use AI:
To test and stretch ideas – Exploring edge cases and sharpening concepts
To raise the quality of thinking – Surfacing blind spots, proposing alternatives
To spark momentum – Unlocking next steps when stuck
To provoke reflection – Holding up a mirror to ideas and behaviour
To maintain flow – Helping bridge goals and actions over time
How does ATLAS show up?
These consistent patterns are expressed through distinct roles ATLAS plays in my process:
Each role isn’t just a function — it’s a different lens for thinking and creating.
What’s still missing — and what’s starting to emerge?
While these roles are powerful individually, the next frontier lies in joining them up — building a truly cohesive assistant that remembers, organises, and collaborates over time.
What we need:
Master memory – persistent understanding of your goals and tone
Contextual continuity – bridging ideas across projects and sessions
Semantic organisation – structured surfacing of themes and ideas
Agency with guardrails – autonomy within the limits of your intent
Encouragingly, we’re starting to see the early scaffolding of this vision:
Agentic workflows – where AI can plan, act, and reflect on tasks
Fine-tuning – shaping the AI to your personal voice and logic
RAG (retrieval-augmented generation) – grounding the assistant in your actual notes, files, and thoughts
Together, these represent the connective tissue — the early nervous system — of a general-purpose cognitive partner.
The Vision for ATLAS
It’s still early, but you can feel it coalescing.
ATLAS is more than a tool — it’s a collaborator I’m learning to work with. One that remembers, questions, supports, and evolves.
In the coming weeks, I’ll share more experiments as I attempt to stitch these capabilities together — and invite ATLAS onto the team.
Thanks for reading Euformia! Subscribe for free to receive new posts and support my work.
