Alignment Integration

Alignment Integration / for organizations / sample diagnosis

demonstration on fictional data

Sample diagnosis, on fictional data

You map out the full process, whatever it is that we're automating or improving, and through that you can identify in each step what's required and how the whole workflow will be done.

offer
Diagnosis and intervention. Two weeks, $3,500 CAD flat.
setting
A fictional 40-staff community mental health and addictions agency in Ontario.
process
The quarterly report to a funder: indicators pulled from records and spreadsheets.
reader
Executive directors and finance officers.

01

The process map, before and after

People understood the mess because they lived in it.

Before: as it runs today

demonstration on fictional data

A B C D email 01 Quarter ends. The funder template arrives by email to finance records system 02 Data lead exports the client list from the records system spreadsheet 03 Exports service contacts separately and joins the two in a spreadsheet spreadsheet, paper 04 Four supervisors fill caseload sheets from memory and paper tallies email 05 Supervisors email their sheets to the data lead spreadsheet 06 Data lead reconciles sheets against the export. The totals do not match email 07 Lists clients with missing fields and emails each worker records system, email 08 Workers fix some records, answer the rest by email. Data lead exports again exception 09 Clients in two programs counted twice. De-duplicated by hand exception, paper 10 Walk-in and phone contacts sit in a paper binder. Reception tallies them funder template 11 Data lead types the indicators into the funder template spreadsheet 12 Finance checks service units against budget targets. They do not line up email 13 Executive director reviews and questions a number shared drive 14 Director signs, finance submits. Sheets saved to the shared drive
  • A06 to 04: totals do not match, supervisors recount
  • B08 to 02: records changed, export again
  • C12 to 06: units off target, reconcile again
  • D13 to 09: a number questioned, de-duplicate again
14 steps 4 loops 2 exception cases the same fact in 5 places
Demonstration on fictional data

After: the same report, redesigned

demonstration on fictional data

E records system 01 In the quarter: intake and discharge fields are required pick lists records system 02 In the quarter: reception logs walk-in and phone contacts as they happen email 03 Quarter ends. The funder template arrives by email to finance records system 04 Data lead runs the indicator report. One client, one count records system 05 The report lists records with a missing required field records system 06 Supervisors clear them in the records system within two days funder template 07 Data lead copies the indicators into the funder template 08 Finance checks units against targets. Director signs, finance submits
  • E06 to 04: exceptions cleared, run the report again

Deleted outright

  • 04Supervisor caseload sheets
  • 07Emailing each worker
  • 09De-duplicating by hand
  • 10The paper binder
  • 14Copies on the shared drive
8 steps 1 loop, bounded 0 exception cases each fact in 1 place
Demonstration on fictional data

Sometimes you don't even need AI or automation; it's just simplifying, because you may have multiple steps in the process that don't need to be there, multiple sources, too many exception cases. Once you can map it all out you can really optimize it.

We will build within your systems or on top of your systems, and not have you change systems.

02

The numbers, before and after

Only certain people would be able to get that data, or it would take programs and teams full days of pulling data manually.

fictional figures. before: measured in week one. after: the plan's estimate
measurebeforeafter, as planned
TimeStaff hours per quarterly report, and working days from quarter end to submission. 62 hours
14 working days		 18 hours
4 working days
Data accuracyHow many things are not documented properly or are documented in different ways, so they're not captured when running reports. 71 of 100 records
complete and consistent enough to count		 94 of 100 records
complete and consistent enough to count
CostStaff cost per report at a blended $42 an hour, against the one-time cost of the fix. $2,604 per report
$10,416 a year		 $756 per report
$3,024 a year, after a $5,180 fix

Hours, before fictional

02 to 03 exports and join
8 h
04 four supervisors, 4 h each
16 h
06 reconcile the totals
6 h
07 to 08 chasing missing fields
10 h
09 de-duplicate by hand
4 h
10 tally the paper binder
3 h
11 type the template
3 h
12 finance check
4 h
13 director review and rework
6 h
14 sign, submit, file
2 h
per report
62 h

Hours, after fictional, as planned

04 run the indicator report
2 h
05 to 06 four supervisors, 1.5 h each
6 h
04 again, 07 copy into template
2 h
08 finance check
3 h
08 director review
3 h
08 sign, submit, file
2 h
per report
18 h

Cost fictional, $42 an hour

before: 62 h × $42
$2,604
after: 18 h × $42
$756
saved per report
$1,848
saved a year, 4 reports
$7,392
the fix: diagnosis and intervention
$3,500
the fix: 40 h of the agency's administrator × $42
$1,680
the fix, one time
$5,180
pays back after
2.8 reports

If you can set up something properly and it's less than what you're currently paying, then it's worth it.

03

What you receive

Once you can map all that out, it becomes easy to automate, to improve, to optimize.

A process map

  • Before: every step, loop, exception case, and place the same fact lives.
  • After: the redesigned process, indicators captured once in the records system.

The initial plan

  • What to delete, what to change in the records system, in what order.
  • The three measures, before and as planned: time, data accuracy, cost.

The two weeks

  1. Leadership meeting

    A leadership meeting to ensure the vision and alignment.

  2. Week one

    The first week would be collecting info, having meetings with people, checking the whole landscape, gathering documents and materials.

    At the end of the first week, we develop the initial plan and process map.

  3. Review

    Review that for feedback with leadership, and then each stakeholder that's responsible for certain parts will review those.

  4. Final product

    We take all of that feedback, adapt, and then have a final product with any minor adjustments.

AI is not a miracle solution that will do everything you want.

Start the conversation For organizations

The initial consultation is free.