Checklist: is your business ready for an AI messaging funnel?
A 12-point readiness checklist for SMBs considering WhatsApp or Telegram AI automation: channel, offer, content, process, metrics.
Every week we get "we want an AI bot, where do we start?" The honest answer — first check whether your business is ready for automation, not just wants to buy one. Automating on a raw process breaks things and creates a bad experience faster than manual work ever would.
Here are 12 points to audit readiness in about 15 minutes. Each one is either a check or a miss. At the end — what to do with the result.
Channel (points 1-3)
1. You have an active WhatsApp or Telegram number for the business.
Not the owner's personal number — a dedicated work number or business account. Customers already message it, and it gets checked during the day. If "technically we have one but nobody watches" — miss.
2. You know how many inbound messages you get per day.
At least roughly: 10, 50, 200 per day. Without volume, you can't size the ROI. If you can't ballpark it within ±50% — miss.
3. You know where customers come from into the messenger.
Website, Instagram, Google Business, referrals, paid ads — each channel brings different context and different expectations. Without this, the AI will default to "hi, how can I help?" where the customer expected a reply about a specific promo they clicked on.
Offer (points 4-5)
4. You have a written service list with durations and prices.
Not "starting from $45," but specific line items: "Women's cut — 60 min — $45." If the price list lives in the receptionist's head and changes weekly — miss.
5. You can name 3-5 daily repeat scenarios.
Book a service, ask about hours, cancel or reschedule, ask price, ask location — the top five. If you can list them in descending frequency — check. If every conversation is unique (rare for SMBs) — miss, automation won't pay off.
Content (points 6-8)
6. Your FAQ answers are written down in one place.
Even in a Google Doc, even just 10-20 common questions with answers. This is the training base for the assistant. Without it you'll be writing answers live during setup, which drags launch out by 2-3 weeks.
7. You have a written policy on cancellations, reschedules, deposits.
"Cancel 24h ahead — free; 2h ahead — we keep 50%." If your policy is "we'll see" and depends on the receptionist's mood — miss. AI doesn't do "we'll see."
8. You know your tone of voice.
Formal or casual? Emojis or no emojis? For a salon and a law office this is fundamentally different. If you can't articulate it, look at the last 10 conversations your staff handled — the pattern is usually obvious.
Process (points 9-10)
9. Someone on your team can monitor the bot during the first 2-4 weeks.
Not necessarily a marketer — a solid front-desk person or you yourself. Someone needs to spend 30-60 minutes a day reviewing logs, adding scenarios, correcting mistakes. Without that, the bot plateaus at 60% quality and stops growing.
10. You have a human handoff process.
When the AI can't handle something, a team member gets a notification and replies. If "the bot exists so nobody has to reply" — miss. You'll lose 15-20% of complex customers — the ones who matter most.
Metrics (points 11-12)
11. You can state your current "inbound message → customer" conversion rate.
At least approximately: out of 100 inbounds, how many convert? 10? 30? 60? Without this number you can't tell whether automation helps or hurts. If the answer is "I don't know" — miss. Start with a manual count for one week.
12. You know customer lifetime value (average ticket × repeat visits).
For a salon: $45 × 6 visits/year = $270 LTV. For an auto shop: $120 × 3 visits = $360. If you have "average ticket is around X and repeat is whatever" — miss. Without LTV, you can't calculate payback on automation within 1-2 months.
What to do with the score
Tally your checks out of 12:
- 10-12 — launch now. Your process is mature; automation will show results in 2-4 weeks.
- 8-9 — launch, but close the weak points in parallel during the first month.
- 5-7 — clean up the process first (1-2 weeks of work), then automate. Saves a lot of pain.
- Under 5 — automation right now will have negative ROI. Start with basics — channel, pricing, FAQ — and come back in a month.
At ZiFlow we often talk customers out of connecting immediately. Losing a week on prep beats losing three weeks on "why is the bot dumb" after launching on a raw process.