Booking Q3 2026 · 2 retainer slots open · Direct or via SI Paris ·Seoul
Sébastien Tang SALESFORCE SOLUTION ARCHITECT
EN / FR Book a call
Foundation Method · self-assessment

Is your Salesforce org ready for the AI era?

Nine questions about your data, your architecture, and your AI surface. I ask them one at a time. You answer. At the end you get a verdict, not a brochure.

See the 9 questions
~2 min9 questions, 4 answers each
No signupAnswers stay in your browser
Instant verdictAnd a 3-step next move
Fig.02 · The conversation

Nine questions. One at a time.

Score 0–100 · instant
ST
Sébastien Asking 9 questions
0 of 9
Your answers stay in your browser. No tracking on this section. Restart
Fig.03 · The 9 questions

The full assessment, in one page.

Every question, the canonical "what good looks like" answer, and a short essay per pillar. Use it as a checklist before your next architecture review.

Data foundation

Can your data answer an agent's questions?

Agentforce is only as good as the data it reaches. Three indicators decide whether an agent gets one trusted answer or three contradictory ones: identity resolution, where the 360 actually lives, and how fresh the calculated insights are. Most stalled pilots fail here, not in the prompt.

Q1 · D1 Do you have a single customer ID resolved across clouds? What good looks likeIdentity Resolution rulesets are deployed and trusted by Sales, Service, and Marketing. One contact, one ID, one source of truth.
Q2 · D2 Where does your customer 360 actually live? What good looks likeData 360 in production, with Data Streams from each cloud and a Data Graph that an agent can grounded-query.
Q3 · D3 How fresh are your calculated insights? What good looks likeRefreshed in minutes, not days. Drift between insight and reality is the #2 reason agents give wrong answers in production.
Architecture maturity

Can your integration layer serve agents fast enough?

Agents amplify whatever architecture you have. If Flow and MuleSoft duplicate each other's logic, the agent picks the wrong one and you debug it for a quarter. Ownership, governance, and integration debt are the three architecture-maturity indicators that matter most before AI.

Q4 · A1 Who owns orchestration between Flow and MuleSoft? What good looks likeOne named architect. A documented decision record (ADR) for which layer owns which orchestration pattern.
Q5 · A2 Is there a governance model for architectural changes? What good looks likeADRs reviewed monthly. Change request board with a clear escalation path. Decisions are recorded, not remembered.
Q6 · A3 When did you last map your integrations end-to-end? What good looks likeWithin the last six months. The map is current enough that a new architect can be productive in a week, not a quarter.
AI readiness

Can you safely let an agent act?

An agent that gives wrong answers is annoying. An agent that takes wrong action is a liability. The three AI readiness indicators are about the contract between the agent and the rest of the org: what it may know, what it may decide, what it may do.

Q7 · AI1 Is there a written grounding policy? What good looks likeA one-page document, signed by Security, listing every Data Graph and field an agent may use, with the legal basis.
Q8 · AI2 Are you using the Atlas Reasoning Engine in production? What good looks likeYes, with Prompt Builder templates versioned in Git. Pilot is not production. Production has rollback.
Q9 · AI3 Do you have guardrails for what an agent may do? What good looks likeApproval flows for irreversible actions. Audit log on every agent decision. A kill switch the CTO can pull.
Fig.04 · Methodology

Why these nine. Why this weighting.

Nine indicators selected from the patterns I've seen across 200+ Salesforce engagements between 2014 and 2026. Each indicator has a weight, set by how often it appears in pilots that stall in months four through eight.

The scoring is deterministic. Same answers, same score, every time. No black-box model. If you want to see the source of any weighting, ask. I send the receipts.

01
Three pillars, three questions eachData Foundation, Architecture Maturity, AI Readiness. Each pillar scores out of 100. Overall score is the weighted average.
02
Weights by failure-mode frequencyAn indicator that appears in 60% of stalled pilots is weighted higher than one that appears in 20%. Source: my engagement log.
03
I don't know still countsIt scores as zero on that indicator, because in production 'I don't know' and 'no' produce the same outage.
04
Privacy-firstYour answers never leave your browser. If you ask to email yourself the report, only your address is processed. No tracking on this page.
Fig.05 · Book the call

We'll know in 30 minutes whether I can help.

No slides. No pitch deck. Bring the architecture diagram or describe the problem in your own words. I tell you whether I'm the right fit and what the next step costs, before you've finished your coffee.

  1. 01.Replies within 24 hours, always
  2. 02.If I'm not the right fit, I'll point you at someone who is
  3. 03.No follow-up emails unless you ask
Book a Discovery Call → hello@sebastientang.com
Book a Discovery Call →