Enterprise Ai

AI and Salesforce transformation

By David Stott, MBA · November 2, 2026 · 6 min read

Do not choose a Salesforce configuration, AI product, or automation roadmap before diagnosing the operating friction the change must remove. Start with an evidence-first friction inventory, tie every proposed platform decision to a named accountable outcome, and treat AI as an amplifier of redesigned, owned decisions—not as a shortcut to cure unclear work or poor data. Sequence three parallel workstreams: (1) an Executive Friction Report to expose root causes, (2) an Enterprise AI Maturity assessment for any AI use case, and (3) a Strategy-to-Systems alignment that maps required capabilities, integrations, and ownership. Each recommended action must name an accountable decision owner, an observable measure of reduced friction, and a 30/60/90-day action that can be reviewed by the executive team.

Recommendation

Stop selecting Salesforce features, integrations, or AI toolkits as the first step. Diagnose the operating friction that prevents the business outcome; then choose the platform response that removes that friction with least added complexity. Treat Salesforce as the context for decisions, data, and workflows—not the solution without an operating model and accountable owners.

Why this matters

Enterprises move faster toward platform and AI procurement than they do toward clarifying who decides, which data matters, and which work should change. That sequence converts platforms into additional governance, technical debt, and user resistance. AI only amplifies human judgment when the work is owned, simplified, and measured. Otherwise AI accelerates the same problems.

A practical decision sequence

Define the business outcome and evidence of value

State the one business decision or customer outcome the program must improve. Describe where value is currently lost and the observable evidence you will use to judge improvement.
Assign one senior accountable owner for that outcome. The owner must have authority to make the tradeoffs required to change process, data, and platform.

Run an Executive Friction Report before platform or AI choices

Inventory recurring friction across decision, process, data, technology, organization, trust, and customer experience dimensions.
Prioritize by impact and urgency. For each top item name the decision required to remove it and the single owner who will make that decision.
Sequence 30/60/90-day interventions that reduce the friction most likely to block platform success.

Use an Enterprise AI Maturity lens for AI proposals

Classify each AI use case by business decision, risk tier, and who remains accountable when AI recommends or acts.
Evaluate data readiness and decision architecture before approving any model into a production workflow.
Treat AI as conditional: approve bounded pilots that mature governance, grounding, monitoring, and human-review before scaling.

Apply Strategy-to-Systems alignment for Salesforce architecture choices

Translate the outcome into the business capabilities Salesforce must enable: specific workflows, data constructs, integrations, and user experiences.
Expose misalignment between desired capability and current org structure, data ownership, or integrations. Do not let a desired feature displace unresolved operating gaps.
Use an Architecture Decision Framework to compare options against business value, operating complexity, integration impact, data implications, and reversibility.

Design and operating rules for Salesforce + AI

Redesign before automating: Simplify and own the work before you automate it. Automation should reduce handoffs, exceptions, and duplicate effort—not multiply them.
Clarify decision rights: For every automated or AI-supported action, name the accountable human, the evidence required to act, and the escalation path.
Ground AI outputs: Require documented grounding sources, confidence signals, and a human override for material decisions. Monitor drift and behavior after launch.
Limit scope to reduce coupling: Start with bounded processes and a minimal set of integrations; add complexity only after value is proven and governance demonstrated.
Preserve reversibility: Prefer approaches that can be reversed or constrained quickly when operating evidence contradicts assumptions.

Measures that matter

Measure reduced friction not activity. Track changes in decision cycle time, repeat escalations, rework, customer effort, and the specific business outcome that justified the investment.
Make measures visible in executive forums and tie release decisions to them. Approve scale only when evidence shows reduced friction and improved decision quality.

30/60/90-day pragmatic agenda (example structure)

0–30 days: Executive Friction Report: inventory top 5 recurring friction points; assign accountable decision owners; agree measures and the one business outcome.
30–60 days: Narrow scope: select a single Salesforce workstream and one bounded AI use case; run Enterprise AI Maturity and architecture decision reviews; implement short technical spikes to validate integration assumptions.
60–90 days: Pilot and measure: run the pilot under practical governance, collect outcome evidence, document incidents and overrides, and prepare a proceed/prepare/pause recommendation for scale.

Common anti-patterns to avoid

Buying features before resolving ownership and data: This creates islands of capability that require constant tribal knowledge and escalate rather than solve the problem.
Treating AI as a shortcut to replace work design: Automation amplifies ambiguity if the underlying work is inconsistent or contested.
Averaging readiness scores: Use the lowest material dimension to determine the appropriate action—don’t hide a critical blocker inside an average.
Overcentralizing governance without decision rights: Committees slow action; named owners accelerate it.

Decision checklist before committing to Salesforce or AI scale

Is the primary business outcome and its evidence explicit and agreed by accountable executives?
Has the Executive Friction Report identified the root friction the program must remove and named decision owners?
Are the required data elements governed, accessible, and trusted at the point where the decision will be made?
Does the Architecture Decision record show integration, maintenance, and reversibility consequences that the business owner accepts?
For AI: Is the use case classified, risk-tiered, and assigned to a named accountable human who can pause the system?
Are stop-or-adjust measures and a review cadence defined and accepted by executives?

Next steps for leaders

Commission an Executive Friction Report as the gating artifact before vendor selection, platform design, or AI procurement.
Require an Enterprise AI Maturity assessment for every AI use case that will influence decisions or take action.
Require architecture decision records that include business tradeoffs, operating consequences, and a review trigger.

Claims that require verification

Any assertion about the suitability of a specific Salesforce edition, packaged product, or integration approach for your environment requires an inventory of your org’s custom objects, integrations, and technical debt.
Any projection of AI impact or cost savings requires business-case evidence tied to the defined outcome and a data-readiness review.

David’s Perspective

Start with the problem you expect technology to solve. Platforms and models look attractive because they are concrete; that is precisely why leaders must force the opposite discipline: make the operating problem explicit, name the decision owner, and require measurable evidence of reduced friction before platform or AI scale. Do this and the platform becomes a tool for capability, not an additional operating burden.

Back to Insights