01
Department AI Discovery Survey
Early signals.
Clear priorities.
A decision-ready view of what TGP’s first three ambassadors need, where time is being lost, and what the workshops must resolve.
Live interim report
Results will develop as more departments respond.
3responses received
3departments represented
01 Sepresponse deadline
Snapshot
The opportunity is real. The operating model needs equal attention.
The first responses point to meaningful time savings across reporting, outreach, and market research. They also show that success depends on cleaner source data, clear ownership, agreed quality tests, and protected human judgment.
The workshops should not begin with tool selection. They should begin by defining the workflow, the source of truth, and the approval path.
Who has responded
02
Consumer Marketing Director
Marketing Advisory03
Brand Development Director
Brand Development & Graphic Design01 · Emerging workshop tracks
Three distinct priorities,
one shared constraint.
Each response names a different first workflow. All three depend on reliable inputs and a clear test for quality.
01
Reporting & data
Trusted reporting from fragmented sources
Bring HubSpot, campaign platforms, spreadsheets, trackers, and team inputs into one controlled reporting flow.
Workshop focusAgree the source of truth, KPI definitions, ownership, exception handling, and the five-day review path.
2–3 daysreported monthly effort
02
Growth & outreach
Influencer discovery without manual prospecting
Identify relevant creators by region, audience, content fit, and engagement, then structure outreach and tracking.
Workshop focusDefine fit criteria, data sources, quality thresholds, outreach stages, and the human approval point.
1 / monthtypical campaign cycle
03
Research & strategy
Cited market studies that ground brand decisions
Accelerate market, competitor, customer, pricing, and trend research without weakening strategic judgment.
Workshop focusSet required sources, citation standards, output structure, recency rules, and how conclusions are validated.
Up to 4 weekscurrent project-start effort
02 · Current environment
Work crosses too many surfaces to automate in isolation.
Reporting alone may draw on CRM, campaign, analytics, spreadsheet, presentation, file, and human inputs.
HubSpot CRMGoogle AnalyticsGoogle AdsMeta Ads ManagerLinkedInExcelPowerPointSharePointCanvaAgency reports
What the responses tell us
Data quality is part of the product.
Automation cannot compensate for incomplete CRM records, conflicting trackers, late inputs, or unclear ownership. The workflow must surface exceptions early.
Outputs must stay editable and explainable.
Respondents want polished decks, reports, and studies, but calculations, conclusions, citations, and commercial interpretation must remain open to human review.
Speed cannot weaken brand judgment.
AI must follow current brand guidance, avoid generic language, use current data, and combine concise storytelling with relevant insight and considered design.
The best first build is testable.
Each pilot should be measured against an approved historical output before it enters a live cycle, using accuracy, time saved, manual handling, and correction count.
03 · Readiness check
Materials are taking shape. Review capacity is the early risk.
All respondents have a current template. Two have strong examples and brand guidance. Only one confirmed the five-business-day review commitment without discussion.
Important distinction
Two respondents selected “Needs discussion” for the five-day review window. Workshop scheduling should include an explicit capacity and approval conversation, not assume the turnaround is settled.
04 · Workshop design
Five decisions every session must produce.
The workshop output should be a build brief, not a broad discussion about AI.
01
Source of truth
Which system owns each input, and what happens when records disagree?
02
Definition of done
What exact output, accuracy level, format, and citation standard must each skill meet?
03
Human control
Which calculations, recommendations, and commercial judgments always require review?
04
Ownership and timing
Who supplies inputs, resolves exceptions, tests the output, and approves within five days?
05
Build order
Which dependency must be fixed before automation can save time instead of adding another layer?