Why Fund the Landing — Not Only the Model?

AI investments are not neutral. Soft landings design a more human future: machines absorb fragmentation and low-judgment transaction so people get larger blocks for insight, empathy, decision making, direction, problem definition, creativity, and collaboration. Hard landings buy speed, takeout, and denser leftovers. Efficiency alone on the dashboard is not a strategy.

I define that fork in The AI Soft Landing. These ten tests turn “we’re investing in AI” into a fundable landing — or a deliberate no — before the vendor demo becomes destiny. If your organization is already buying the wrong landing, see 8 Signals You’re Preparing for the Wrong Future of Work.

If you cannot pass these ten on one page, you are not funding AI. You are funding a hard landing.

1. What Is the Named Soft Landing Test?

Test: Can we describe the more human future this investment creates — in human terms — not only the vendor roadmap?

Pass: A written soft landing: what machines absorb, what humans keep, what depth and dignity grow.

Fail: “AI to stay competitive and reduce cost” with no landing paragraph.

Ask before you fund: Which landing are we buying — soft or hard — in one paragraph? Boards that ask this early stay ahead of the spend — see 7 Questions Smart Boards Ask About the Next Five Years.

2. What Is the Cognitive-Labor Split Test?

Test: Is there an explicit split between glue and transaction work for AI and named human endeavors that must grow?

Pass: An offload list plus a protect list — insight, empathy, judgment, creativity, collaboration, teaching, repair, and more.

Fail: Vague “higher-value work” with no calendar or role changes.

Ask before you fund: What contiguous human work expands if this works? For the protect catalog, see 11 Human Endeavors AI Should Free (Not Replace).

3. What Is the Time-Dividend Policy Test?

Test: Is there a rule for reclaimed time — depth versus denser busyness?

Pass: Explicit policy — for example, a share of saved time funds deep work, coaching, and recovery — not only more tickets.

Fail: Utilization stays the religion; calendars refill automatically.

Ask before you fund: What happens to the first 100 hours this AI saves?

4. What Is the Human Accountability Test When AI Acts?

Test: When the system drafts, routes, decides, or acts — who is askable, with undo and escalation?

Pass: Named accountable role; decision rights; appeal path; “the model decided” banned as an answer.

Fail: Autonomy without ownership; humans as rubber stamps.

Ask before you fund: Who owns the outcome when the AI is wrong? Pair with 5 Scenarios for Agentic Organizations and 6 Trust Pillars for Agentic Customer Experience when agents act for customers.

5. What Is the Work-Redesign-Funded-With-the-Bet Test?

Test: Are job design, incentives, enablement, and old-path kill funded with the AI spend — not after?

Pass: Redesign budget and owners equal to the tech workstream.

Fail: Copilot bolted onto broken process; denser leftovers called transformation.

Ask before you fund: What work redesign ships in the same release train?

6. What Is the Trust Contract Test for Affected Humans?

Test: Do customers and/or employees get a trust contract — disclosure, control, consent to scope, recovery — matching stakes?

Pass: Written trust requirements for the use case; powered make-right.

Fail: Containment, surveillance, or “helpful” scope creep without consent.

Ask before you fund: What would betrayal look like — and how do we prevent it?

7. What Is the Human-Success Scoreboard Test?

Test: Will we measure human success and adopted behavior — not only FTE takeout, automation rate, or demos?

Pass: Dual scorecard: efficiency and time-to-confidence, completion, trust, relapse, depth time protected.

Fail: Business case opens and closes on headcount math.

Ask before you fund: Which human-success metrics can veto a “green” efficiency story?

8. What Is the Change-Capacity Honesty Test?

Test: Can the organization absorb this change without stacking another “and also”?

Pass: Visible stop/start list; portfolio load named; permission to refuse.

Fail: Another AI epic on top of an overloaded human system.

Ask before you fund: What initiative dies so this landing can live?

9. What Is the Behavior-Evidence-Before-Scale Test?

Test: Is there a falsifiable human behavior, a cheap evidence plan, and a decision date before enterprise scale?

Pass: Named behavior + kill criteria + date; scale gated on evidence.

Fail: FOMO scale from a demo; “we’re past pilot” without adopted-behavior proof.

Ask before you fund: What must humans do differently — and by when must we know? Before any pilot check clears, use 11 Questions Before Funding Any Innovation Pilot.

10. What Is the Dignity and Honest Winners/Losers Test?

Test: Have we named who gains, who loses, and how dignity is protected in transition?

Pass: Honest impact map; transition paths; no silent extraction.

Fail: “Win-win for everyone” while politics and fear go underground.

Ask before you fund: Whose agency shrinks if this works — and what do we owe them?

What Is the Go/No-Go Checklist Before the Next AI Investment Review?

Ten checks on one page:

1. Landing named?
2. Labor split written?
3. Time-dividend policy?
4. Askable human?
5. Work redesign funded?
6. Trust contract?
7. Human-success metrics?
8. Capacity / stop list?
9. Behavior + decision date?
10. Dignity map?

Fund only if the landing is soft by design. Pause if the demo is strong and the humans are vague. Kill if efficiency is the only value on the dashboard.

Mantra: Don’t buy a model. Buy a more human future — or don’t write the check.

FAQ: More Human Future Tests for AI Investment