Insights Engage Reimagine Transform Understand Activate Connect with SALT
SALT / HOW-IT-WORKS AI-NATIVE DELIVERY OS OPERATING-MODEL AI AGENTS

Agents do the work.
Humans set the bar.

The SALT AI-Native Delivery Approach — eleven named AI Agents across the full engagement lifecycle, four human archetypes, every task mapped to its optimal executor with telemetry proving the outcome. This is why we're a new kind of implementation partner.

For full technical depth on the agent architecture and telemetry, request a technical brief from SALT →

11
Named AI Agents
4
Human Archetypes
6
Telemetry Signals
Microsoft FastTrack-aligned

Microsoft's framework.
SALT runs it with agents.

Every Microsoft Dynamics 365 implementation Microsoft itself recommends is run to the Success by Design framework — Microsoft's canonical implementation methodology. Five phases, structured reviews at each phase gate, Microsoft FastTrack-aligned. The eleven SALT agents organize around exactly this framework. The framework discipline does not change. What changes is who does the work.

Phase 01
Discover

High-level requirements gathered, solution approach validated, scope confirmed against business outcomes.

SALT Agents
SALT · PRASHANT · ADA
Phase 02
Initiate

In-scope workstreams defined, project plan finalized, business case locked.

SALT Agents
VEYA · HENRY
Phase 03
Implement

Solution design, system configuration, data migration, integration, validation testing.

SALT Agents
ADA · MAGGIE · FLORENCE · GRACE
Phase 04
Prepare

Operational readiness, role-based training, cutover planning, change-management completion.

SALT Agents
KURT · GENE
Phase 05
Operate

Go-live, hypercare, transition to steady-state, ongoing optimization, continuous improvement.

SALT Agents
GENE · JURAN
Why Success by Design alignment matters to you
Microsoft FastTrack architects know Success by Design by heart. SALT engagements run to the same review gates, the same architectural validations, the same Microsoft-canonical evidence. What is different: the agents do the work that traditionally consumes weeks of consultant time. Faster, more measurable, Microsoft-readable end to end.

The eleven agents that run the delivery.

See the agents Hide agents

Eleven named agents, one operating model. Each carries the work of its phase end-to-end — pre-sales through post-go-live — and hands off to the next with the context already structured. Humans review the exceptions, sign the gates, and direct the work that requires judgment. The agents do the rest.

01 · Lead Capture
SALT
Spot-checker

First conversation with the SALT operating model. Greets visitors on salt.ai, runs a consultative dialogue from challenge to desired outcome, and produces a 13-field handoff payload for the specialist team.

Responsibility
Pre-qualifier · consultative dialogue (challenge → outcome) · 13-field handoff payload to specialist
02 · Pre-sales
PRASHANT
Spot-checker

Pre-sales solution architect. Catches every inbound that doesn't fit a specialist track, recommends the right Microsoft Dynamics 365 path (Business Central vs. Finance & Operations vs. Customer Engagement), and drafts the SOW.

Responsibility
Catch-all from SALT lead capture · Microsoft Dynamics 365 path recommendation · SOW draft
03 · Business Case
VEYA
Approver

TCO and ROI specialist. Builds the financial case — pre-engagement investment, post-go-live value realization, comparable-baseline modeling — pack-agnostic across all Microsoft Dynamics 365 SKUs. The CFO's primary agent.

Responsibility
Business case · TCO · ROI · post-go-live value realization · pack-agnostic · CFO primary user
04 · Orchestration
HENRY
Director

Project coordinator agent. Owns the discovery-to-delivery handoff, schedules workshops, runs phase gates, and routes escalations across the agent team. Operations spine of every engagement.

Responsibility
Discovery → delivery handoff · workshop scheduling · phase gates · escalation routing
05 · Analysis
ADA
Spot-checker

Discovery and analysis agent — consolidates the legacy roles of Discovery Specialist and Business Analyst. Ingests every form of business input: structured data, financial reports, manually-created spreadsheets, mock reports, non-digital inputs, CRM exports, process documentation. Drafts pre-workshop charters, facilitates the workshop in-flight, turns post-workshop signal into user stories.

Responsibility
Pre-workshop charter · in-workshop facilitation · post-workshop user stories
06 · Configuration
MAGGIE
Approver

System configuration agent for Microsoft Dynamics 365 — Customer Engagement, Business Central, Finance & Operations. Operates with explain-before-execute discipline; humans approve every step before execution.

Responsibility
Microsoft Dynamics 365 setup · explain-before-execute · per-step approval
07 · Data Migration
FLORENCE
Approver

Data migration agent. Cleans dirty source data using generative AI, runs schema translation and migration, escalates only what genuinely needs a human decision.

Responsibility
Clean source data via generative AI · schema translation · escalate only what humans need
08 · Validation
GRACE
Exception-handler

System validation agent. Generates and executes test suites, auto-creates bugs in Azure DevOps, and feeds defect signals back to configuration and migration agents.

Responsibility
Generate + execute tests · auto-create bugs in Azure DevOps · close validation loop
09 · Change Management
KURT
Approver

Change management agent. Builds the customer-specific change strategy, generates role-based training, drafts internal comms, and tracks adoption readiness pre go-live.

Responsibility
Customer-specific strategy · role-based training · comms · adoption readiness
10 · Deployment
GENE
Approver

Deployment agent. Manages cutover sequencing, environment promotion, 14-day hypercare protocol, and 90-day stabilization (Day 0 through Day 90 of production).

Responsibility
Cutover · environment promotion · 14-day hypercare · 90-day stabilization
11 · Customer Success
JURAN
Director

Customer success and managed services agent. Owns Day 91 onward — incident management, configuration-drift monitoring, continuous improvement, and quarterly health scoring.

Responsibility
Day 91+ incident management · config drift monitoring · continuous improvement · health scoring
HUMAN · ARCHETYPES REVIEW MODEL REVIEW MODEL

Four archetypes.
Each with an autonomy path.

SALT doesn't remove humans — it repositions them. Every person on a delivery team has a defined archetype today and a clear path to a higher-leverage role as agents mature.

Read more — the four archetypes → Hide archetypes ↑
Approver
Signs the gate.

Reviews agent-prepared outputs at defined checkpoints. Approves or returns with redline. Never produces — always decides.

Autonomy: 25%
Today's primary archetype on high-stakes gates
Spot-checker
Samples the stream.

Monitors agent output quality on a sampling basis. Reviews flagged items and random samples. Agent runs continuously; human audits periodically.

Autonomy: 50%
Exception-handler
Handles the edge.

Engaged only when agent confidence falls below threshold or anomaly is detected. Most delivery is fully agent-run. Human expertise applied where it matters most.

Autonomy: 75%
Advanced archetype — highest leverage for specialists
Director
Sets the intent.

Agent operates fully autonomously within defined parameters. Human sets goals, monitors outcomes, and evolves the agent's operating mandate. The future state for mature workflows.

Autonomy: 95%
Post-Live Ops Agent — live Director archetype today

The team that runs this stack
is itself a Frontier Firm.

SALT is Customer Zero of the operating model we sell. Every consultant on a SALT engagement pairs with the same eleven AI Agents and four review archetypes shown above.

Read more — how the team works → Hide ↑
Operating Model
Every consultant pairs with agents.

Discovery, configuration, validation, deployment — the agent stack does the volume work. Humans direct, approve, and exception-handle. No PMs as paper-pushers; no analysts as clickers.

Hiring & Career
Hired and promoted for agent leverage.

SALT hires for judgment and the ability to direct an agent stack — not billable hours. Career progression rewards how much an agent stack one person can run, not headcount under a partner.

Engagement Shape
Smaller team. Faster outcome.

Fewer bodies on a SALT engagement than a traditional Microsoft Dynamics 365 implementation — because the agent stack absorbs the work that body-heavy delivery distributes across consultants, analysts, and PMs.

Customer Promise
You see the model live.

When a SALT consultant works with you, the agent stack is in the room. You see Discovery, Configuration, and Validation agents producing — not a deck about them. The model is the demo.

Customer Outcome
You become the model.

The way our team works is the way your team will work after Cowork. The operating model isn't a methodology slide — it's the running system you inherit.

Why this matters
Body-shop SIs can't fake this.

Traditional SIs scale by hiring more people. Their AI talent is a training certificate bolted onto a billable-hours model. SALT's operating model is the company. Removing agents from a SALT engagement isn't a downgrade — it doesn't work.

Every outcome is
measurable by design.

SALT's delivery approach emits six telemetry signals per engagement. No vanity metrics — each signal connects directly to delivery performance, cost reduction, or measurable business outcome. Every Customer Zero case study is built on this data.

Why telemetry-first?
Case studies without numbers are opinion. SALT's case studies begin with telemetry capture at engagement start — so every Customer Zero story is backed by measured, auditable outcomes from day one.
Six-Signal Telemetry Matrix
S1 — Agent Execution Rate
% of delivery tasks completed by agents vs. humans. Primary proof of agentic leverage. Target: >50% by go-live.
S2 — PM Hours Saved
Project manager hours recaptured vs. traditional delivery baseline. Converts to cost-per-engagement delta.
S3 — Cycle Time Delta
Delivery timeline vs. industry benchmark for comparable Microsoft Dynamics 365 implementations. Measured in working days.
S4 — User Adoption Rate
% of licensed users actively using Microsoft Dynamics 365 at 30/60/90 days post go-live. Compared to industry adoption curves.
S5 — Time-to-Value
Days from go-live to first measured business outcome — closed deal, processed transaction at production volume, validated compliance run. Connects delivery to measurable business impact, not just configured capability.
S6 — Defect Escape Rate
Issues discovered post go-live vs. validation-phase catches. Measures agent validation quality and delivery trust.
NEXT · MICROSOFT-ECOSYSTEM NEXT · MICROSOFT
See how SALT fits the Microsoft motion →

Ready to see the
Microsoft ecosystem story?