A Day at Kumi SACCO: From Manual Calls to AI Voice Outreach

Morning before Wetu Voice – the scramble

6:00 am – The branch manager, Grace, opens the loan‑collection spreadsheet.

• 120 members are past due. Their phone numbers sit in a static Excel file.
• Grace assigns three junior officers to call each member manually, noting outcomes on paper.
• By 9 am the office is a chorus of ringing phones, missed calls and scribbled notes.
• Members without a data bundle receive only a missed‑call reminder; many never see the message.
• The team spends hours reconciling who was spoken to, who promised payment, and who asked for a callback. Errors creep in, and the daily collection rate hovers around 30 %.

The pain points are clear:

• Time‑consuming manual dialing.
• No guarantee the member actually receives the reminder.
• No record of what was said or promised.
• No way to segment members by preferred contact time.

First contact with Wetu Voice – setting the stage

Grace schedules a scoping call with FTA. The conversation focuses on three goals:

1. Reach every member, even those without mobile data.
2. Automate follow‑up so each conversation picks up where the last left off.
3. Gain real‑time reporting on delivery, replies and cost.

FTA outlines how a single recorded message – "Your loan repayment is due on Friday. Please press 1 to confirm payment, 2 to request a new schedule" – can be sent as:

• An outbound voice call for members with voice minutes.
• An SMS for those preferring text.
• A WhatsApp message for members with data.

The system also creates a smart audience: members who usually answer after 6 pm are scheduled for evening calls, while those who answer early get a morning message. Consent checks ensure quiet‑hour rules are respected and opt‑outs are logged automatically.

Day 1 after deployment – automated outreach begins

1. Recording the message

Grace records a concise voice note (30 seconds). The AI agent indexes the script and links it to the SACCO’s knowledge base – repayment dates, contact numbers, and FAQs about loan terms.

2. Audience segmentation

Using the member database, the Wetu Voice platform builds three groups:

Segment	Preferred channel	Best contact window
Urban members (data)	WhatsApp	9 am‑12 pm
Rural members (no data)	Voice call	6 pm‑9 pm
Young members (SMS)	SMS	12 pm‑3 pm

3. Launching the campaign

At 8 am the system dials the first 40 rural members. Each call begins with the AI voice agent greeting the member by name, then playing Grace’s recorded reminder. If the member presses 1, the agent confirms the payment plan and logs the interaction. If the member presses 2, the agent offers to schedule a callback and records the request.

4. Real‑time reporting

Grace opens the dashboard and sees:

• Delivery rate: 98 % of calls reached a live person.
• Responses: 45 % pressed 1, 20 % pressed 2, 35 % did not respond.
• Cost: Total spend displayed per channel, with voice calls costing marginally more than SMS but yielding higher response.

Mid‑day – shifting from manual to AI‑assisted follow‑up

By 12 pm the junior officers no longer need to dial. Their tasks change to:

• Reviewing the dashboard for members who requested callbacks.
• Using the AI agent’s memory to continue the conversation. For example, Member A, who asked for a new schedule, receives a personalized follow‑up call the same day: "Hi A, you asked to move your repayment to next Friday. Does that work?"
• Updating the loan ledger only after a confirmed payment, eliminating duplicate entries.

The AI’s memory ensures each interaction feels personal, even though the same recorded script is used. The system recalls that Member B previously asked about early repayment penalties and automatically includes that answer in the next call.

Afternoon – measuring impact and refining strategy

Grace runs a quick analysis:

• Collection rate rose to 55 % within the first half‑day, compared with the usual 30 %.
• Agent time saved: 3 officers reclaimed roughly 4 hours each, allowing them to focus on high‑value member service.
• Member satisfaction: A brief post‑call survey (delivered via SMS) shows an 85 % satisfaction score, with comments praising the clear reminder and easy response options.

Based on the data, Grace adjusts the audience timing:

• Moves the rural‑call window to 7 pm‑10 pm for better answer rates.
• Adds a second reminder for members who didn’t respond within 24 hours, using the same AI voice agent but a different script offering a callback link.

End of day – the new routine

At 5 pm Grace reviews the daily report:

• Total members reached: 120 (100 %).
• Confirmed payments: 68.
• Callback requests: 22, now queued for personal follow‑up.
• Opt‑outs: 2, automatically excluded from future campaigns.

The manual phone‑tree is replaced by a predictable, data‑driven workflow:

1. Upload or sync the member list.
2. Record one message.
3. Define smart audiences and timing.
4. Launch – the AI agent handles delivery and interaction.
5. Review the dashboard for outcomes and cost.
6. Act on callbacks or payment confirmations.

Grace’s team now spends the afternoon on relationship‑building activities rather than dialing numbers. The SACCO’s cash flow improves, and members feel heard because each call references their previous conversation.

Practical steps for any bank or SACCO

1. Map your outreach need – loan reminders, fee notices, product offers.
2. Gather a clean member list – include name, preferred channel, and best contact window if known.
3. Create a concise script – 30‑seconds, clear call‑to‑action.
4. Set up smart audiences – let Wetu Voice decide timing per segment.
5. Enable consent checks – protect members with quiet‑hour and opt‑out rules.
6. Launch a pilot – start with a small group, review delivery and response.
7. Iterate – tweak timing, scripts, or channels based on the dashboard.
8. Scale – roll out to the full member base, continuously monitoring cost and ROI.

By following this roadmap, banks and SACCOs can shift from labor‑intensive manual calls to an automated, AI‑powered voice outreach that respects member preferences, saves staff time, and boosts collection rates.

Explore Wetu Voice →