Boom Media

Boom Labs · Product Builds

Replyee vs Zoho SalesIQ — parity map & pick list

Every row below was checked against the Replyee codebase, not against the planning docs — those claimed Phase 2 complete and understated what actually shipped. Tick the boxes in “What’s missing” to choose the next build.

Verified 29 July 2026 · repo SaaS/Replyee · compared against zoho.com/salesiq

SalesIQ features already built

Partially built

Missing

Things Replyee does that SalesIQ can't

The report you asked for already exists

The Zoho “Daily Summary Report” — visitors on site, new vs returning, top pages, traffic sources — is already implemented in src/app/api/cron/daily-report/route.ts, and vercel.json already schedules it: "0 8 * * *", every morning at 08:00. It computes visitors total / new / returning, the top pages, traffic sources grouped by referrer hostname, conversations split by mode, and leads captured. That is the Zoho layout almost field for field.

So the only real question is whether it is firing. Replyee is on Vercel, so the cron is live only if the project has been deployed since vercel.json landed and the route's auth secret is set. Check Vercel → Project → Cron Jobs for the last run. If it has never run, this is a configuration task, not a build.

✅ Already wired — SalesIQ features Replyee has today

Built

Real-time visitor tracking dashboard

Live table of who is on the site right now — bot, current page, page views, visit count, device, last seen, and a one-click action to open a chat. Updates live via Supabase postgres_changes, no refresh.

src/app/dashboard/visitors/page.tsx

Built

Visitor heartbeat tracking

The widget pings every 30 seconds with current page, referrer, UTM parameters and device. This is the data layer everything else reads from.

POST /api/visitor-heartbeat · replyee_visitor_sessions

Built

Proactive chat triggers

SalesIQ's signature feature. Four rule types — time on page, URL contains, exit intent, and return visitor — evaluated client-side, auto-opening the widget with a custom message. Configured per bot in a Triggers tab.

PUT /api/chatbots/[id]/triggers · Triggers tab in bot settings

Built

Live chat with human takeover

Agent inbox with live transcripts. Replying flips the conversation from bot to human mode and the bot goes silent until released. Triggered by the visitor asking for a human, low bot confidence, or manual takeover.

src/app/dashboard/inbox · conversations.mode

Built

Daily summary report email

Visitors total / new / returning, top pages, traffic sources by referrer hostname, conversations by mode, leads captured. Scheduled 08:00 daily. This is the screenshot you sent.

src/app/api/cron/daily-report/route.ts · vercel.json crons

Built

Canned responses / agent shortcuts

Type / in the reply box for an autocomplete dropdown of saved shortcuts with body preview. Managed per user.

src/app/dashboard/canned-replies · GET/POST/DELETE /api/canned-replies

Built

Typing indicators, both directions

The agent sees “Visitor is typing…” live in the inbox, and the visitor sees the agent typing. Realtime broadcast events.

Built

Agent presence — “We're online”

The widget shows online status when an agent has the inbox open, and falls back to lead capture when nobody is there.

Built

Browser notifications for agents

Desktop alerts on new messages so the inbox does not have to stay in the foreground.

Built

Persistent visitor identity

A UUID in localStorage survives sessions, so return visitors are recognised — which is what makes the return-visitor trigger and the new-vs-returning split work.

Built

AI chatbot

Client documents (PDF, URL, text) chunked and embedded into pgvector, answered by Claude Haiku over vector search. Meaningfully better than SalesIQ's flow-builder bot for factual questions about a specific business.

/api/chat · knowledge_chunks · match_chunks RPC

Built

Lead capture

Captures the email when the bot cannot answer, notifies the client via Resend, and now also fires a per-bot webhook into the client's own portal.

replyee_leads · /api/leads · /api/handoff

Built

Analytics dashboard

Total conversations, active bots, leads captured, this-month volume, a 7-day activity chart, most-common visitor words, and a per-bot breakdown.

src/app/dashboard/analytics/page.tsx

🟡 Partially built

Partial

Lead scoring

A score field already exists on the visitor row and renders in the dashboard, but there are no scoring rules and the list is not sorted by it. SalesIQ scores on traffic source, time spent, last active, past visits, current page and location, then sorts hot leads to the top. The plumbing is there; the rules engine is not.

src/app/dashboard/visitors/page.tsx — score: number

Partial

Conversation analytics

Volume and lead counts exist. The operational metrics SalesIQ reports do not: bot deflection rate, first-response time, average handle time, and missed-chat rate.

🔴 What's missing — pick what to build

Ordered by my read of value per unit of work. Effort figures assume the existing visitor and conversation tables are reused. Tick the boxes for the ones you want.

Visitor geolocation — country, region, city

The single most visible gap. SalesIQ's dashboard is built around country flags and location; Replyee stores no geo data at all. The heartbeat already reaches the server, so this is an IP lookup on ingest plus three columns and a flag in the table.

Effort: ~1 day · unlocks location-based scoring and triggers

Lead scoring rules + sort hot leads first

Finish what the score field started: a configurable rule set over traffic source, time on site, return visits, pages viewed and (once geo lands) country — then sort the live view by score so the hot visitor is always at the top.

Effort: ~1–2 days · the feature you screenshotted second

Visitor action history / past chats timeline

Click a visitor and see their full path through the site, every prior visit and every past conversation. Replyee already stores the raw pings; this is a detail panel over data that exists.

Effort: ~1–2 days

Chat ratings and CSAT report

Thumbs up/down plus an optional comment at the end of a conversation, aggregated into a satisfaction report. Nothing exists today. This is also the cheapest proof to show a client that the chat is working.

Effort: ~1 day

Operational metrics — deflection, first-response, missed-chat rate

The numbers that justify the subscription at renewal: what share of chats the bot handled without a human, how fast agents responded, and how many chats went unanswered.

Effort: ~1–2 days · strong retention argument

Saved views and filters on the visitor list

SalesIQ's “Select a view” — filter and sort by last activity, time spent, past chats, visits, and hide visitors already being handled by another agent. Presentation-layer work over data Replyee already has.

Effort: ~1 day

Multi-agent — team members, departments, routing, transfer

Today the inbox assumes one operator. No department, assignment, or transfer logic exists anywhere in the codebase. This is the big one for selling to any client with more than one person answering.

Effort: ~1 week · needed before multi-seat pricing works

Per-client branded daily/weekly report

The daily report exists but goes to the account owner as plain summary mail. Make it a Boom-branded client-facing report — matching the transactional email shell — with a weekly option and per-bot breakdown, so it doubles as an account-management artifact.

Effort: ~1 day · builds on what already ships

Mobile agent experience

Responsive inbox first, PWA with push after. SalesIQ has native apps; a good responsive inbox covers most of the real need for a small business owner.

Effort: ~2–3 days for responsive, longer for PWA push

Omnichannel — WhatsApp, Messenger, Instagram

SalesIQ's instant-messaging channels. Nothing exists today. Worth noting this overlaps the SMS add-on plan and needs a boundary ruling before it is built.

Effort: ~1–2 weeks per channel incl. platform approval

Privacy-first tracking — consent banner, IP masking, chat privacy disclosure

SalesIQ leads with this: a cookie/consent opt-out banner before tracking starts, optional IP masking, and a privacy disclosure with consent opt-out when a chat begins. Replyee tracks visitors with no consent gate at all today. Given Compliee is already in the suite and you sell to regulated clients, this is a credibility item as much as a legal one.

Effort: ~2 days · GDPR/CCPA posture, pairs with Compliee

Audience criteria — multi-condition trigger rules

Replyee's four trigger types each fire on one condition. SalesIQ composes rules with and/or: “page is pricing and country is USA and visitor type is returning”. Upgrading the existing trigger JSON to a condition tree is a contained change on top of what already works.

Effort: ~2 days · needs geo for the location conditions

Returning-visitor personalization panel

When a known visitor returns, show the agent their chat history, previous browsing paths, and whether they are active or idle — so the greeting can reference what they were looking at last time. Replyee has the persistent visitor ID and the message history; this is the panel that joins them.

Effort: ~2 days · overlaps the action-history item above

Automated mail list building from visitor behaviour

Auto-add visitors to a mailing list based on pages visited, source, and lead score. For Boom this should almost certainly push into SureContact rather than a list inside Replyee — the lead webhook pattern already exists, so it is mostly rule configuration.

Effort: ~2 days · check app boundaries first (SureContact owns CRM)

Screen share and audio call

SalesIQ upsells both. Lowest priority here — high build cost, and rarely the reason a local business buys chat software.

Effort: high · recommend skipping

⚡ Where Replyee already beats SalesIQ

Edge

RAG over the client's own documents

Upload the menu, the policy PDF, the service pages — the bot answers from them. SalesIQ's bot is a flow builder; matching this needs their separate Zobot/AI tier.

Edge

Live order status from Boom Online Ordering

A visitor can ask “where is my order” and the bot answers from real BOO data. No generic chat vendor can do this for your restaurant clients.

/api/internal/order-status · /api/webhooks/boo-order

Edge

Leads pushed into the client's own portal

Per-bot webhooks deliver leads straight into the client's dashboard rather than trapping them in a vendor inbox — the pattern behind the BMBB portal integration.

Edge

White-label economics

SalesIQ is priced per operator per month, and the client sees Zoho branding. Replyee is yours, embedded in the client's site as your product, at Boom pricing.

🧭 Recommended order

#	Item	Why this order	Effort
0	Confirm the daily report cron is firing	The feature is already built. Verify before building anything.	Minutes
1	Visitor geolocation	Most visible gap, and it unblocks location-based scoring and triggers.	~1 day
2	Lead scoring rules + sorting	Half-built already; completes the flagship SalesIQ view.	~1–2 days
3	Chat ratings / CSAT	Cheapest client-visible proof the chat is working.	~1 day
4	Operational metrics	The renewal argument — deflection rate and response time.	~1–2 days
5	Branded client report	Turns an existing cron into an account-management artifact.	~1 day
6	Multi-agent + departments	Biggest build; needed before multi-seat pricing.	~1 week

Verification note. Every “built” row was confirmed by reading the route or page in SaaS/Replyee. Every “missing” row was confirmed by searching the whole src/ tree and finding no implementation — geolocation, ratings, departments, assignment and the messaging channels each return zero matches. The planning doc .md/SALESIQ_UPGRADE_PLAN.md is behind reality: it lists the daily digest as an unstarted Phase 3 item when it is built and scheduled.