System Atlas
Atlas
Constructs/Process

Standard architecture: Artificial intelligence

Workflows that drive an outcome from step to step. Process steps that hand work to each other.

Standard architectureThe parts at least half of the organizations in Artificial intelligence have, as many as 20 of them run all together. Counted from public, sourced parts only; each organization below links to the parts it was counted on.

Create a free account to save Create a free account to follow
Organizations running it
20
Of their industry
29%
The pattern
What every instance has, in order
  1. Decisions
    Market and positioning
    in 100% of instances
    then
  2. Process
    Lead generation
    in 100% of instances
    then
  3. Process
    Product or service design
    in 100% of instances
    work
  4. Process
    Production or delivery
    in 100% of instances
    then
  5. People
    Org structure
    in 100% of instances
    then
  6. Process
    Hiring
    in 100% of instances
    then
  7. Decisions
    Strategy and goals
    in 100% of instances

Coloured arrows: a flow of that type seen between the two at instance organizations. Grey arrows: order only.

Structure
GroupsLayersEverything
Element
Share of instances
Companies
Invariant 7
At least 90% have it
People1
Org structure
100%
20
Process4
Lead generation
100%
20
Product or service design
100%
20
Production or delivery
100%
20
Hiring
100%
20
Decisions2
Market and positioning
100%
20
Strategy and goals
100%
20
Variant 32
20 to 80%: where they differ · Members
People4
Sales
75%
15
Suppliers and partners
65%
13
Customer service and support
60%
12
Account management
25%
5
Process9
Customer onboarding
70%
14
Evaluation and red-teaming
50%
10
Logistics and fulfilment
40%
8
Pretraining and fine-tuning runs
35%
7
Uptime and incidents
35%
7
Onboarding and training
35%
7
Learning and improvement
35%
7
Quality control
30%
6
Billing and invoicing
20%
4
Software2
Inference serving
55%
11
Integrations
45%
9
Data5
Training data acquisition and licensing
30%
6
Compute capacity
25%
5
Data management
25%
5
Inventory and assets
20%
4
Reporting and dashboards
20%
4
Decisions12
Policies and culture
80%
16
Pricing
65%
13
Data protection and security assurance
65%
13
Security and access
60%
12
Regulatory compliance
60%
12
Usage pricing and metering
55%
11
Model release and versioning
45%
9
Performance and pay
40%
8
Health, safety and environment
35%
7
Contracts
30%
6
Governance and decision rights
30%
6
Usage policies and model documentation
25%
5
Variant · where they differ
Member
Variants, couplings and instances

Every level, number, source and dimension: Member, $300 a year.

See Member
Couplings · what comes together
Member
Variants, couplings and instances

Every level, number, source and dimension: Member, $300 a year.

See Member
Entropy
Where it runs rough

Operating-reality findings at instance companies: workarounds, manual hand-offs, bottlenecks.

  • Sales12 findings
    Manual hand-off 7 · Dependency 3 · Bottleneck 1 · Informal channel 1 · at 11 companies
  • Suppliers and partners6 findings
    Dependency 6 · at 6 companies
  • Customer onboarding5 findings
    Dependency 5 · at 5 companies
  • Production or delivery5 findings
    Dependency 3 · Bottleneck 1 · Manual hand-off 1 · at 5 companies
  • Training data acquisition and licensing4 findings
    Dependency 2 · Manual hand-off 1 · Workaround 1 · at 4 companies
  • Lead generation2 findings
    Dependency 1 · Informal channel 1 · at 2 companies
  • Product or service design2 findings
    Bottleneck 1 · Informal channel 1 · at 2 companies
  • Pricing2 findings
    Tacit rule 2 · at 2 companies
Seen across
Where it holds
Size
  • 201-500 people15%
  • 51-200 people10%
  • 501-1000 people5%
Region
  • United States35%
  • North America10%
  • Canada5%
  • France5%
  • Japan5%
  • New York (US); also offices in London and San Francisco5%
Instances
The 20 companies it was seen at
Member
Variants, couplings and instances

Every level, number, source and dimension: Member, $300 a year.

See Member