Wizard Step 5: Outcomes
This step answers the operational question: “What do we do after each call?” It controls retries, follow-ups, and how you measure success.
At a glance
- Retries decide how persistent the campaign is.
- A/B testing helps you compare two or three agents fairly.
- Follow-ups decide how results turn into action.
- Goals give you a target so “good” isn’t vibes.
If your outcome settings create unexpected behavior (too many retries, weird voicemail loops), you should be able to pause confidently. Put pause criteria in your launch checklist.
Retries
Retry on voicemail
Recommended Keep this ON if voicemail is not a meaningful outcome for your program.
Keep it OFF when:
- voicemail is the intended touch (rare), or
- compliance constraints require one attempt only.
Max retries
Max retries controls the total attempts per contact. The UI supports 1–3.
Suggested defaults:
- Recommended 1 for very sensitive audiences or strict compliance.
- Recommended 2 for most pilots.
- Advanced 3 when you’ve proven your numbers and list quality.
Max voicemail attempts
If voicemail retries are enabled, this caps how many of the attempts may hit voicemail before the system gives up. If left blank, it defaults to the same value as Max retries.
If your list is low-consent or low-relevance, retries don’t improve results — they create spam risk. Fix list quality first, then tune retries.
A/B testing (2–3 agent variants)
Use A/B testing when you want to compare:
- two opening strategies,
- two objection handling approaches,
- two levels of “aggressive vs consultative” tone.
How it works (important)
- Contacts are split deterministically across variants (so results stay comparable).
- Variant weights must total 100.
- Variant A always uses the primary agent from Step 1 (Basics).
Variant rules (enforced)
- 2–3 variants only.
- Every variant needs a name.
- Every variant must use a distinct agent.
- Every weight must be positive.
- Total weight must equal 100.
If Variant B changes the opening, the pitch, and the qualification questions, you won’t know what caused the lift. Make the variants meaningfully different, but not chaotic.
Follow-ups & memory
Follow-ups decide how your system turns outcomes into action:
- create tasks,
- write notes,
- route leads,
- trigger downstream workflows.
Memory should be used deliberately:
- enable when you want consistent multi-touch behavior, and
- keep it OFF for short pilots unless you’ve validated data governance.
Goals (success targets)
The wizard supports these goal targets:
- Recommended Answer rate (%)
- Advanced Conversion rate (%)
- Advanced Average call score (%)
Goals don’t change behavior — they change how teams interpret results.
Validation rules (enforced)
- Goal targets must be numeric.
- Values must be within valid ranges (0–100 for percentage goals).
If goals are invalid, the wizard blocks Next (so you don’t launch with broken targets).
Final checklist
- Retries match your audience sensitivity.
- Voicemail behavior is intentional.
- If A/B is enabled: weights total 100 and agents are distinct.
- Follow-up behavior is clear to the team.
- Goals reflect the metrics you actually care about.