AI as a structured accelerator within a trusted workflow
A complex configuration workflow slowed staff and demanded institutional knowledge. We refined the workflow first, then introduced AI as a guided accelerator — improving speed and usability while preserving precision, transparency, and control.
This was a commercial engagement delivered by the team behind This Waay. It is presented as relevant experience for public-systems work — not as a government contract reference.
Context
JAMS is a mature automation platform built to manage complex, mission-critical workflows. Its job-creation experience was powerful and highly configurable, designed for technical staff who valued control and precision.
Over time, the workflow accumulated complexity. Experienced administrators could navigate it efficiently, but new or infrequent staff faced high cognitive load and a steep learning curve.
The opportunity was twofold: improve the clarity of the workflow itself, and explore how AI could responsibly reduce friction without disrupting trusted system behavior.
We partnered with product and engineering leadership to define where AI could meaningfully support existing workflows, shape the experience strategy, and design AI-assisted interfaces grounded in real operational constraints.
Methodologies
- Workflow analysis and task decomposition
- Information architecture evaluation
- Heuristic analysis
- System data and error-state mapping
- Concept modeling and iterative prototyping
- Cross-functional stakeholder alignment
Deliverables
- Refined job-configuration experience
- Explainable guidance and input patterns
- Natural-language-to-structured-configuration concept
- Interaction models for AI-supported workflows
- Design specifications for engineering implementation
The issue was not capability. It was friction within a trusted workflow.
Creating a job required navigating multiple configuration steps, selecting from extensive option sets, and understanding dependencies that were not always visible. Key challenges included:
- A multi-step configuration process that slowed setup
- Limited guidance on which inputs were required for a given intent
- Risk of misconfiguration for less experienced staff
- Heavy reliance on documentation and institutional knowledge
A full replacement with a conversational AI interface risked disrupting trusted workflows and reducing transparency in a system built for precision.
Clarify and strengthen the workflow first — then decide where AI belongs.
We streamlined the highest-friction steps, clarified required inputs, improved sequencing, and reduced unnecessary cognitive load — preserving the canonical configuration model while making it more legible. Only after the structural improvements were defined did we determine where AI could assist. The approach emphasized:
- Strengthening the underlying workflow before augmentation
- Preserving the configuration model staff already understood and trusted
- Introducing AI within clearly defined boundaries
- Maintaining full visibility into system-generated outputs
- Requiring explicit review and confirmation for every suggestion
AI was positioned as a co-pilot within a clarified system, working inside the structure staff already trusted.
Natural language in, structured configuration out — with a person confirming every value.
An AI-assisted creation flow was introduced alongside the refined configuration experience. Staff could describe their intended automation in natural language; the system translated it into structured configuration fields, pre-filling relevant parameters while preserving visibility into each setting.
The improved workflow provided the validation layer for AI suggestions — every generated value mapped directly to structured fields and required confirmation. This enabled:
- Faster setup for common use cases
- Reduced manual configuration effort
- Greater clarity in complex dependencies
- Continued access to advanced controls for experienced staff
Reduced friction without compromising operational trust.
- Reduced time to configure standard jobs
- Increased confidence among new and infrequent staff
- Maintained trust among experienced administrators
- Improved consistency in configuration
- A scalable pattern for future AI-assisted capability
Because AI operated within a validated configuration model, expansion could occur without eroding transparency or governance. Structural clarity made intelligent augmentation sustainable.
Most government systems aren't broken. They're trusted, but inefficient.
Replacing trusted workflows with opaque AI-driven experiences introduces risk, reduces transparency, and erodes the confidence of the staff who run them — the fastest way to stall an AI initiative in the public sector.
Introducing AI within the workflow — every output mapped to structured fields, every suggestion visible and confirmed by staff — improves speed without sacrificing the governance and auditability public systems require. Efficiency and trust are not a trade-off when the work is sequenced correctly.
More engagements that show how we work.
Let's find the signal.
Is there a workflow your staff trust but everyone knows is slow? Let's find where AI can responsibly accelerate it.
Start a conversation →