Field notes  ·  AI marketing

Choosing an AI implementation partner: the questions that separate real from theatre.

An AI label does not establish value. DigitalOS recommends asking how a provider measures outcomes, assigns responsibility and explains the limits of its work.

Choosing an AI Implementation Partner: What to Ask

Seven questions

  • Which outcomes will you report? Define the primary metric, supporting indicators, costs, period and attribution limits. Booked and collected revenue differ.
  • Who is accountable? Agree on responsibility and review milestones.
  • What do we retain when we leave? Put ownership, licenses, access, exports and transition support in writing. Third-party software may not transfer.
  • Do you receive vendor commissions? Ask for material relationships to be disclosed and recommendations justified against alternatives.
  • Can we reconcile media costs? Require access and itemized spend, fees, taxes and credits. Direct billing alone does not prove transparency.
  • What is within your control? Separate delivery commitments from forecasts and third-party outcomes. Ask about exclusions and remedies.
  • What supports each claim? Ask for the original source, method, scope and limitations.

Evaluate each AI application

Potential uses include triage, follow-up, scheduling, reporting and content assistance. Compare accuracy, cost and impact with a baseline. Use appropriate privacy controls and human oversight. Not every inquiry can be answered or booked automatically.

The reporting test

Request a redacted or authorized example explaining objectives, costs, outcomes, data quality and attribution limits. Awareness work may need supporting indicators before reliable cost-per-outcome figures are available.

These are DigitalOS evaluation recommendations, not claims that a particular tool or provider guarantees growth.

Sources and limitations

Reviewed October 29, 2026. External findings are not DigitalOS experiments or guarantees. Practical recommendations are DigitalOS guidance; research, products and policies can change.

  1. U.S. advertising-substantiation policy · November 23, 1984

    Objective advertising claims require a reasonable basis before publication. This policy does not prove that a particular service improves revenue.

  2. U.S. endorsement and testimonial guidance

    Material relationships and endorsements require appropriate disclosure where applicable; private procurement and other jurisdictions can differ.

  3. U.S. healthcare privacy guidance for marketing

    Guidance for covered entities and relevant activities. Authorization, security and vendor obligations require a case-specific review.

Alongside Your Marketing Team

Explore Alongside Your Marketing Team →

A quick, no-cost conversation about your goals, where AI can help first, and the next steps we'd recommend.