eniks.Talk to Louis
← All three examples

Example workflow · Scattered exports → Sent client report

Recurring reporting

The client review is at 09:00. The report is still in pieces.

6 steps by hand → 1 approval

Before 6 steps, all by hand

  1. RaviGathers exports and screenshots
  2. RaviSpots one file is for the wrong month
  3. RaviRequests a refreshed export
  4. RaviCounts and reconciles the records
  5. RaviWrites the commentary
  6. RaviReviews and sends

After 1 step for a person

  1. AIChecks each figure’s month
  2. AIRequests the right export
  3. AICounts and reconciles the records
  4. AIDrafts factual commentary
  5. RaviRavi checks it and approves — the report is sent

Watch it happen · Recurring reporting

Source periods → refreshed June records → reconciled counts and factual commentary → final owner approval → sent client report. Source connections appear at the final review. Invented business and sample data; playback sends nothing.
Film transcript and visual description

The client review is approaching

The client review starts at nine. Is the monthly report ready? June is still in pieces: exports, screenshots, and another round of copying. Before you can explain the figures, you have to check which ones belong together. Campaign spend covers June. The website enquiries export covers May.

On screen: Alder Repairs has a monthly review at 09:00. Exports and screenshots still need assembling. Campaign spend is for June, while the website enquiries file is for May.

Align the metrics and period

An AI workflow can carry the assembly and checking forward. It reads each export and identifies its metric and reporting period. This report is for June. Every included figure must match that period. June campaign spend is ready: one thousand two hundred pounds. The May enquiries stay outside the June total. The missing figure remains unconfirmed.

On screen: AI reads each metric and period. June spend of £1,200 enters the June report; May enquiries stay outside it. The June enquiries figure remains unconfirmed.

Refresh, count and explain

The AI emails the source owner for the June enquiries export. That figure waits while the workflow prepares the rest of the report. Later, the reply arrives with the refreshed file. The AI matches the reply to this report and reads the new export. It checks that the records cover the first through the thirtieth of June. It counts thirty-one website enquiries and reconciles the count with the export. Then it groups the records by week, making the pattern easier to check. It drafts a factual summary: the busiest full week had ten enquiries.

On screen: AI requests the June enquiries export, matches the reply and checks 1–30 June. Thirty-one records are counted and reconciled, then grouped as 6, 7, 10, 6 and 2. The last two days are a partial week. The factual summary identifies 15–21 June as the busiest full week.

Review figures and commentary

Only now does Ravi step in to check the finished report. Campaign spend links to its June export. One thousand two hundred pounds. The June enquiry records support the count and the weekly grouping behind the summary. The original file, refresh request, reply, and period checks are available beside the report. The summary keeps spend and enquiries distinct. These figures alone do not establish advertising attribution.

On screen: Ravi reviews the finished report. Spend connects to four £300 source entries; enquiry records connect to the count and weekly grouping. The refresh history and checks remain available. Commentary keeps spend and enquiries distinct and makes no unsupported advertising attribution.

Approve and send the report

Ravi checks the client recipient, then approves the report or returns a correction. After approval, the workflow sends the report and records the outcome.

On screen: Ravi checks the client recipient and can approve or return a correction. Only after approval does the report travel to the client inbox, with the send outcome recorded.

A review that can focus on decisions

Now the review can start with checked figures and focus on what to do next. Bring your reporting workflow to Eniks.ai.

On screen: The opening monthly review returns beside a sent June report: checked figures, a supported weekly summary and reviewed sources. The invitation is eniks.ai/free-poc.

What happens at each step

AI prepares and checks the work. Ravi reviews the final output, its sources and where it’s going before anything is sent.

  1. 1. Input

    June report, mismatched source periods

    Alder Repairs has a monthly review at 09:00. Campaign spend covers June, while the website enquiries file covers May. Its June enquiry total is still unconfirmed.

    Client
    Alder Repairs
    Reporting period
    1–30 June 2026
    Campaign spend
    £1,200 · June
    Enquiries source
    May · excluded from June
    June enquiries
    Unconfirmed
    Client review
    09:00

    Separate metrics can appear together only with compatible periods and clear definitions.

  2. 2. Processing

    Refreshed records and a checked summary

    AI asks the source owner for 1–30 June, matches the reply and reads the attachment. The workflow validates the dates, counts and reconciles the records, groups them by week and drafts a factual summary.

    Refresh request
    Website enquiries export · 1–30 June
    Reply
    Refreshed June enquiry records attached
    Period check
    First date 1 June; last date 30 June
    Count
    31 rows = 31 website form submissions
    Weekly counts
    1–7: 6; 8–14: 7; 15–21: 10; 22–28: 6; 29–30: 2
    Factual summary
    Busiest full week: 15–21 June · 10 enquiries

    29–30 June is a partial week. Spend and website enquiries remain separate metrics.

  3. 3. Human review

    Review the figures and explanation

    Ravi checks the completed report against its exports, period checks, count, weekly grouping and request history. He reviews the commentary and client recipient, then approves or returns a correction.

    Spend evidence
    Four June entries of £300 = £1,200
    Enquiry evidence
    31 June records; count and weekly grouping reconciled
    History
    Original May file, refresh request, reply and checks retained
    Proposed note
    Advertising attribution is not established by these figures.
    Recipient
    Alder Repairs · client contact
    Decision
    Approve & send, or return for correction

    Source connections appear at the final review. Commentary receives the same scrutiny as the totals.

  4. 4. Output

    Sent June client report

    After Ravi approves, the workflow sends the report and records the outcome. The review can begin with checked figures and a supported explanation of the weekly pattern.

    Period
    1–30 June 2026
    Campaign spend
    £1,200
    Website enquiries
    31
    Factual summary
    Busiest full week: 15–21 June · 10 enquiries
    Limitation
    Advertising attribution is not established by these figures.
    Delivery
    Approved by Ravi; sent to Alder Repairs client contact

    The report retains its source evidence, commentary and approval history.

What would your proof measure?

Time to a correct, approved and sent monthly report, including period checks, reconciliation and review of commentary. Track source-refresh waiting separately from active handling time.

This example uses an invented business and reports no measured saving. Your call would agree your own baseline and the one result that matters.

Recognise this? Tell Louis about your version.

Thirty minutes with Louis and Svilen. Bring one example of the work — no slides needed.

Check whether your workflow fits$0 working proof no commitment

Two more examples