Tour of the Insights dashboard

The Insights dashboard (/insights) is your analytics home base. It pulls together inquiry volume, deal outcomes, revenue, quotes, referrals, and customer sentiment into one page so you can see what the business is doing without building a spreadsheet. This article walks through every block in the order it appears on the page and explains what to do with the numbers.

1. AI weekly narrative

At the top of the page Yesoma streams a short written summary of the week — volume trend, most-common category, and any open quotes that need a follow-up. It renders while the rest of the page loads, so you might see a skeleton card for a second before the text appears.

Think of it as the exec summary before you read the details below.

2. At-risk customers

Directly below the narrative, a gold-tinted card lists customers Yesoma has flagged as needing a check-in, based on:

• Open complaints
• Overdue follow-ups
• Silence patterns — cases that have been quiet longer than expected

Each row links directly to the customer record so you can act immediately. This card only appears when someone is on the list — if everything looks healthy you will not see it at all.

3. Hero stat band — this week at a glance

The large violet card at the top of the data area shows:

• Inquiry count for the last 7 days — with a direction indicator (up, flat, or down vs the previous 7 days)
• 14-day sparkline — so you can see whether this week is a spike or part of a longer trend
• Peak day label — the single busiest day in that window

Use this band for your Monday morning check-in: is inbound volume growing, shrinking, or steady? If it shifted sharply, scroll down to the category and source breakdowns to understand why.

4. Compact KPI row

Four metric tiles sit side by side just below the hero band. Each shows a delta chip comparing the current value to the previous 30-day window.

• Avg response time — minutes from a customer's first message to your first reply. Lower is better.
• Follow-up rate — completed follow-ups divided by total follow-ups scheduled. If this is below 80%, check the Follow-Ups tab for overdue items.
• Complaints (30d) — new complaint-category inquiries this month, plus how many are still open. A rising open count means complaints are not being resolved.
• Opportunities — open quote or booking requests that have no follow-up scheduled. These are the cases most likely to go cold.

5. Case status funnel

This section shows how many conversations currently sit in each case state — Needs reply, Awaiting customer, Follow-up due, and so on — plus the average dwell time: how long cases typically stay in that state before moving on.

Active states appear at the top. Terminal states (Won, Lost, Closed) are separated by a divider at the bottom for context.

Long dwell times in "Needs reply" or "Follow-up due" are the actionable signal. Those cases are waiting on you, not on the customer.

6. Outcomes

The Outcomes section answers "are we winning or losing?" It contains two cards and an optional third one.

Wins, losses, and waiting scoreboard

Won and Lost counts for the last 30 days, with all-time totals underneath, plus the number of cases currently in a "waiting" state (you replied, no response yet). A win-rate percentage sits in the top-right corner; the delta chip compares it to the previous 30 days. Each tile is clickable and routes to the matching inbox filter.

Top close reasons

The most common reasons you typed when closing cases. If "price" appears repeatedly, that is a signal to revisit your pricing conversation or package presentation. If "timing" dominates, a seasonal follow-up strategy may help.

Win rate by channel (appears once you have outcomes on multiple channels)

Which inbound source — email, WhatsApp, website form — actually converts. A channel with a lot of lost outcomes may attract lower-intent traffic and is worth re-evaluating before spending more on it.

When to look: weekly review, or any time you want to answer "where are we leaking deals?"

7. Revenue + customer hub

This section only appears once your workspace has sent at least one quote, received a payment, logged a referral, or recorded a portal interaction. It contains several cards.

Invoices paid

Last-30-day revenue from paid invoices, an all-time total, and a vs-previous-30-days delta shown as an amount (for example, "+$420 vs prev 30 days"). A 14-day sparkline shows the daily shape.

The currency shown is the predominant currency of your recent payments — a UK-only workspace sees GBP, a Ghana-based one sees GHS. Revenue is summed from paid invoice subtotals minus refunds. For the full mechanics behind this number, see Understand Revenue Impact.

Quotes funnel

Four proportional bars: Sent, Viewed, Accepted, Paid. The headline is the overall conversion rate (Paid / Sent). Use it to diagnose where quotes stall:

• Low view rate: customers are not opening the quote link. Check the subject line and preview text you send alongside it.
• Viewed but not accepted: the quote is being read but not signed. Price, clarity, or urgency may need attention.
• Accepted but not paid: the deposit step is creating friction. A payment reminder follow-up usually unblocks this.

Referrals

Sent / pending / converted counts. Below those, the top three referrers are listed with their converted/total ratio. Use this to spot your best advocates and decide whether it is worth seeding the referral program more aggressively. This card only appears when you have sent at least one referral. See Run a referral program for setup.

Customer Hub engagement

Three counts: portal messages sent through the customer portal, documents shared with customers, and customers who have saved communication preferences. These numbers tell you whether the portal is being used. If portal messages are near zero, customers may not know the portal link exists — share it more proactively after onboarding them.

8. Categories and inquiry sources

A two-column row sits below the Revenue section.

Most-requested categories

A ranked bar chart of inquiry types — pricing, scheduling, booking request, complaint, and so on — classified automatically by AI. Use it to spot:

• A single category dominating (a saved template or FAQ page for it would save reply time)
• A complaint spike relative to other weeks
• A new category growing that you have not built a process around yet

Inquiry sources donut

A donut chart breaking inbound down by channel — email, WhatsApp, website form, manual, and any others. The center shows your all-time total. Use this when deciding where to focus marketing spend. If most inquiries come from WhatsApp but you have been investing in a website form, the data is showing you where customers actually prefer to reach out.

9. Opportunity alert, customer sentiment, and regulars

Three cards sit side by side lower on the page.

Opportunity alert

Re-surfaces the count of open quotes going cold with a direct link to schedule follow-ups. If this card says zero, every quote and booking request has been replied to or has a follow-up scheduled.

Customer sentiment

A stacked bar showing the AI-classified sentiment breakdown across all inquiries — positive, neutral, negative, urgent. A large negative segment is worth investigating via the inbox; look for common themes in those cases.

Your regulars

The five customers who have sent the most inquiries, each linking directly to their customer record. Use this to understand who your highest-touch customers are and whether they have open cases that deserve a personal check-in.

10. Suggested actions

The final card on the page is a short, auto-generated checklist Yesoma builds from the data above:

• Overdue follow-ups that need to be closed out
• Open complaints still waiting for resolution
• A dominant inquiry category that would benefit from a template
• Open opportunities without a follow-up scheduled

If everything is on track, the card shows "Nothing urgent — you're running a tight ship." instead of a to-do list.

11. Empty-state handling

If your workspace has no inquiry data yet, the entire Insights page shows a single "Not enough data yet" card. Once inquiries arrive, the page populates automatically.

Individual blocks throughout the page also self-hide when they have no relevant data. If you do not see the Revenue section, the Referrals card, or the At-risk list, it means either those features have not been used yet or there is nothing actionable to surface. Nothing is broken. Each block reappears as soon as the underlying data exists.

12. How fresh is the data?

Insights revalidates on every page load, so what you see reflects the state of your workspace at the moment you opened the page. Hard-refresh to pull the latest numbers.

A few accuracy notes:

Outcomes depend on your close discipline. Win rate, Won count, and Lost count are only as accurate as the dispositions you pick when closing cases. Cases closed without a Won or Lost disposition do not appear in the scoreboard. See Mark inquiries as won or lost.

Revenue depends on marking invoices paid. If you collected cash or a bank transfer and did not mark the invoice paid in Yesoma, that amount is not counted. The revenue tile is only as accurate as your "Mark as paid" hygiene.

Categories and sentiment are AI-assigned. They are accurate most of the time but can be wrong. You can correct a category by editing the case — corrections feed back into the totals immediately.

Time range. Most blocks default to a last-30-days window (or last-7-days for the hero band). All-time totals are shown alongside the 30-day counts where relevant. A date-range picker for full historical slicing across all blocks is in progress.