How this reading was taken
Every figure below comes from self-reported answers, not from an AI-audited paste of real chat history — no scoring here should be read as more precise than that.
- Level bands (Novice / Experimenter / Practitioner / Expert) are provisional — not yet calibrated against a real submission pool. Treat them as a rough sort, not a grade.
- Any aggregate shown for fewer than 4 respondents is suppressed below, per our privacy floor — a small group's answers shouldn't be individually identifiable.
What you told us before anyone answered a single question
- ~1,200 fixed internal staff, ~700 in Monterrey, ~300 across points of sale, plus a large external commission-based salesforce.
- Gemini enabled company-wide via Google Workspace; Claude limited to a small experimental group; Google Drive is not connected to Gemini (security-team decision).
- Two training goals named on the call: personal productivity for the leadership team, and strategic literacy to sponsor and prioritize AI initiatives across the company.
- Decision-maker for any training commitment: the director general.
Where the group sits across three capability dimensions
Collaboration quality — how deliberately people set limits, check output, and hand off work — lags noticeably behind raw usage and knowledge. This group uses AI more than it manages the process of using it.
The one gap almost everyone shares
Five of eight respondents are weakest specifically at collaboration quality — setting boundaries, checking output, handing work through to completion. That's a single, shared, addressable gap, not eight different problems.
Who's confident, and whether that confidence is earned
2 of 8 respondents rate themselves highly but score below the practitioner line — the group most worth a direct, low-ego conversation before training starts, since they're least likely to volunteer that they need it.
Feature familiarity, tool-agnostic
Low familiarity with connectors and agent mode reflects access, not skill — both rows are currently blocked company-wide. Everything else on this table is a genuine literacy gap, and it's a wide one.
Access and how people feel about it
Half the group doesn't have a clear path to the tools they're supposed to be using. Independently, this is the largest single correlate of capability we've seen in comparable groups — bigger than a formal AI strategy, bigger than a dedicated AI lead.
- The Google Drive–Gemini block is a named, deliberate security decision — worth revisiting alongside training, not instead of it.
- An enablement fix (clear access, clear policy) is cheaper and faster than a skills fix, and it compounds with training rather than competing with it.
The recurring work eating the most hours
- "The Monday sales-region rollup — six spreadsheets I stitch together by hand every week."
- "Reading every incoming credit application to flag the ones missing documents."
- "Chasing branch managers for their weekly numbers before I can even start my own report."
What that time is worth, addressable
| category | h/wk (n) | exposure | addressable hrs/yr | addressable $/yr |
|---|---|---|---|---|
| Drafting/editing content | 22h (n=7) | 0.70 | 739 | $18,480 |
| Searching & summarizing | 18h (n=6) | 0.70 | 605 | $15,120 |
| Recurring reports by hand | 16h (n=5) | 0.60 | 461 | $11,520 |
| Data entry/reconciliation | 13h (n=5) | 0.60 | 374 | $9,360 |
| Triage & classification | 9h (n=4) | 0.50 | 216 | $5,400 |
| Total (top 5 categories) | — | — | 2,395 | $59,880 |
Addressable, not saved — no pilot has yet measured a real time reduction at DMX. This is a ranked opportunity list, priced at a placeholder rate, not a committed return.
Which track fits, and for whom
At n=8, below the minimum-5-per-track cohort threshold for a firm group call — but the 5-person Literacy majority is a strong lean. Two respondents show a real Cowork signal, gated only on licence: enabling Cowork on the Enterprise account would unlock that as a bundle rather than a bolt-on.