Age I
Probabilistic engineering
From producing a single "correct drawing" by deterministic means, toward generating and evaluating a distribution of design candidates — selecting probabilistically rather than authoring one answer.
Vision & Philosophy
Where multiple intelligences grow together.
Tektome is not a place to use AI — it is a place where human and machine intelligence become wiser, side by side. The platform itself holds no intelligence; intelligence lives in the people and agents who meet on its ground, and it compounds with every project.
The work of a designer in one day becomes the wisdom of an organisation for one hundred years.
Tektome is where multiple intelligences grow together. The more the platform is used, the more these intelligences are woven into one larger organisational intelligence.
01
How we read this moment.
AEC has always trailed the software world by roughly a decade — changes there reach us, almost unchanged, a few years later. So before describing the future of design, we look hard at what is already reshaping software-led intellectual work. We organise that shift into three Ages, each about to arrive in AEC.
Age II
Vibe coding
The "build anything in natural language" phase converges, under the weight of maintainability and accountability, onto a durable shape: a strong deterministic core surrounded by a flexible space for AI-driven adjustment.
Age III
AI organisation
Organisations move from "using AI as a tool" to "AI collaborating with AI under human direction" — forcing new answers on knowledge, role design, and the question of AI persona.
Our role is to be the translator.
We carry these three Ages into AEC first. Whoever translates an Age late inherits a market already owned by someone else — so translating first, ahead of the industry, is itself the moat.
02
The category we define: Multi Intelligence Platform.
Tektome is a Multi Intelligence Platform for the AEC industry — a field where multiple, qualitatively different intelligences meet, collaborate, and grow one another. It is not a faster chatbot or a lone autonomous agent; it belongs to a fourth generation of AI systems.
Gen 1 · 2022
Conversational AIA single AI responds — ChatGPT, Claude.
Gen 2 · 2024
Agentic AIA single AI acts autonomously — Devin, Cursor.
Gen 3 · 2025
Multiagent AIMultiple AIs collaborate — AutoGen, Crew.
Gen 4 · Tektome
Multi IntelligenceHuman and AI intelligence collaborate as equals and grow one another. Tektome defines this layer — leaping past Gen 3 to claim it first in AEC.
"Tektome itself holds no intelligence. Intelligence lives in those who meet on its ground."
The intelligence of designers, the judgment of veterans, the collective wisdom of organisations, specialised AI agents, and shared industry knowledge meet on a single field, collaborate, and grow one another.
We call this "Field, not Brain." It is the line that separates us from AI-centric platforms: where "Multiagent" can only describe collaboration between AIs, "Multi Intelligence" treats human and machine intelligence as first-class equals. Human intelligence is the protagonist — never the supporting role.
03
The six intelligences.
"Multiple intelligences" is meant literally. Six qualitatively different kinds coexist on the field that is Tektome — and treating them as equals is the essence of the platform.
Human · Individual
Individual intelligence
Designers, architects, regulatory staff — tacit knowledge, experience, and intuition.
Human · Inherited
Inherited individual intelligence
A veteran's judgment axis, continued as a Twin so it outlasts any single tenure.
Organisation
Organisational collective intelligence
The distilled organisational "why" — the collective-intelligence layer of KnowledgeBuilder.
AI · Specialised
AI agent intelligence
Role agents — PM, Regulation, and others — carrying specialised, expertise-focused judgment.
AI · Foundation
AI foundation intelligence
General-purpose language understanding — Anthropic's Claude and its peers.
Industry
Industry-shared intelligence
The public, collective knowledge of the AEC industry itself.
04
Our core asset: the Intelligence Graph.
As an organisation works, it grows a living structure we call the Intelligence Graph. If Google's Knowledge Graph structured the world's facts, the Intelligence Graph defines the next chapter — it structures intelligence itself: a graph of agents, twins, skills, and knowledge, and the connections between them that strengthen with use.
| Aspect | Knowledge Graph (2012) | Intelligence Graph (Tektome) |
|---|---|---|
| What it structures | Facts | Intelligence — judgment and connections |
| Time | Static, reference-type | Dynamic, evolving |
| Owner | Central administrator | The customer organisation |
| Connections | Relationships | Transmission paths, with strength |
| How it updates | Manual / batch | Through natural business activity |
Knowledge Graph changed how machines understand the world. Intelligence Graph changes how organisations grow their own intelligence.
The brand stack that follows from this — from category to the components that grow the graph:
BrandTektome
CategoryMulti Intelligence Platform
Core assetIntelligence Graph — the next generation of the Knowledge Graph
NodesAgent · Twin · Skill · Plugin · Knowledge · Guideline
Edgesdelegates_to · uses_skill · has_knowledge · provided_by — operating as synaptic connections
ProductsSmartCheck · ReqManager · KnowledgeBuilder · App Builder — the touchpoints that grow the graph
Every organisation owns its own Intelligence Graph. A graph database is a place to store; Tektome is a place to cultivate.
05
What we believe, what we give up, how we win.
A strategy is defined as much by what it refuses as by what it pursues. We hold these three to be a single, inseparable set — and we make the trade-offs explicit.
What we believe
- AEC's productivity problem can only be solved by overwhelming AI-driven automation.
- Fully automated design is technically reachable, and will reshape the industry's structure.
- Data and knowledge accumulate inside an organised AI — not in scattered files.
What we give up
- Being a general-purpose tool for everyone.
- Short-term revenue from bespoke contract customisation.
- The urge to build every layer ourselves.
How we win
- A domain moat from accumulated AEC use cases.
- Positioning that general-purpose LLMs cannot absorb.
- Deliberate use of open vs. closed to move an ecosystem.
When these three fall out of alignment, an organisation loses the axis by which it makes decisions. Holding them together is the discipline.
06
Why we are not a general-purpose AI.
We respect general-purpose models — we build on them. But we do not believe AEC is merely "one more domain." The design process itself is a form of intelligence, and it lives outside any single general model. That requires an organised group of intelligences, not one ever-larger brain.
| Axis | General-purpose LLM | Tektome |
|---|---|---|
| Unit of intelligence | A single, ever-larger model | An organisation of many specialised intelligences |
| Domain | Bolted on afterward (fine-tune, tools) | The structure itself |
| Data | Text-centric | Design artefacts, regulation, physics |
| Source of value | Quality of the conversation | Correctness and autonomy of the design |
07
Our philosophy of data.
Intelligence is grown from data, so how we treat data is a matter of principle, not policy. We design ownership in layers.
Layer · Shared
Knowledge becomes a shared asset
Standards and typical solutions are cultivated as community assets, lowering the cost of intelligence for the whole industry.
Layer · Sovereign
Customer data stays sovereign
Project- and company-specific data remains strictly the customer's. We learn from and reuse it only within explicitly granted permission.
This layered model is how we hold ecosystem growth and customer trust together at the same time — rather than trading one away for the other.
Growing intelligence together.
Tektome is the foundation on which an organisation grows its own Intelligence Graph by itself, for itself.