Data and AI Consultancy Desk
Data, analytics and AI consultancies
Data and AI engagements from discovery to handover, with deliverables, evals and margin.
Tech & ITSoftware and ITdata consultingai consultancyanalyticsmachine learningengagementsconsultants
Use this templateData and AI Consultancy Desk
Start Monday on the Dashboard: deliverables due this week, anything in client review or rework, and delivery margin by offering. Each engagement moves Discovery, Proposal sent, Data audit, Pilot, Go or no-go, Production, Handed over. Log time from the engagement so its margin stays true.
Start here: how this workspace runs
Dashboard
From discovery to a go or no-go
Evaluations and handover
Engagements
Every client engagement from discovery to handover: offering, fee, data access and delivery margin.
| Engagement | Stage | Client | Offering | Engagement lead | Fee type | Contract value (USD) |
|---|---|---|---|---|---|---|
| Demand forecast for 40 stores | Handed over | Harbor and Hearth Home Stores | Omar Haddad | $68,000 | ||
| Data readiness audit, claims data | Handed over | Bluestem Health Plan | Lena Kristoffersen | $24,000 | ||
| Policy assistant pilot for the service desk | Pilot | Bluestem Health Plan | Priya Natarajan | $55,000 | ||
| Route cost dashboard | Production | Copperline Freight Systems | Callum Reid | $36,000 | ||
| Warehouse and pipelines rebuild | Data audit | Northvale Credit Union | Omar Haddad | $120,000 | ||
| Scrap-rate prediction, go or no-go | Go or no-go | Ironleaf Fabrication | Omar Haddad | $42,000 | ||
| Monthly analytics advisory | Production | Harbor and Hearth Home Stores | Lena Kristoffersen | $72,000 | ||
| Grant reporting dashboards | Proposal sent | Riverbend County Health Department | Callum Reid | $28,000 |
Clients
The organizations you work for, their industry and data maturity, and the work won with them.
| Client | Status | Industry | Data maturity | Executive sponsor | Sponsor email | Start an engagement |
|---|---|---|---|---|---|---|
| Harbor and Hearth Home Stores | Delia Fontaine, COO | dfontaine@harborhearth.net | — | |||
| Bluestem Health Plan | Raj Malhotra, VP Member Services | rmalhotra@bluestemhealth.net | — | |||
| Copperline Freight Systems | Aoife Brennan, CFO | abrennan@copperlinefreight.net | — | |||
| Northvale Credit Union | Marcus Webb, CEO | mwebb@northvalecu.net | — | |||
| Ironleaf Fabrication | Greta Solberg, Plant manager | gsolberg@ironleaffab.net | — | |||
| Riverbend County Health Department | Dr. Nia Coleman, Director | ncoleman@riverbendcounty.net | — |
Deliverables
What each engagement hands the client: audits, gap maps, dashboards, models, evaluation sets and runbooks.
| Deliverable | Status | Engagement | Kind | Due | Accepted on | Evaluation score (%) |
|---|---|---|---|---|---|---|
| Store-week forecast model and backtest | Accepted | Demand forecast for 40 stores | Aug 12, 2026 | Aug 19, 2026 | 91% | |
| Claims data quality report | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — | |
| Use-case gap map | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — | |
| Policy question evaluation set, 200 questions | Client review | Policy assistant pilot for the service desk | Oct 19, 2026 | — | 84% | |
| Pilot assistant for 15 agents | In progress | Policy assistant pilot for the service desk | Nov 4, 2026 | — | 78% | |
| Route cost dashboard, v1 | Rework | Route cost dashboard | Oct 21, 2026 | — | — | |
| Source system inventory | In progress | Warehouse and pipelines rebuild | Oct 25, 2026 | — | — | |
| Scrap-rate pilot results and recommendation | Not started | Scrap-rate prediction, go or no-go | Oct 27, 2026 | — | — |
Time entries
Hours by consultant and engagement at cost, so every engagement shows its true delivery margin.
| Work done | Billing | Engagement | Consultant | Logged on | Hours | Cost per hour (USD) |
|---|---|---|---|---|---|---|
| Feature engineering and backtests | Demand forecast for 40 stores | Omar Haddad | Jul 21, 2026 | 96 | $95 | |
| Store manager interviews | Demand forecast for 40 stores | Lena Kristoffersen | Jun 21, 2026 | 22 | $120 | |
| Profiling claims tables | Data readiness audit, claims data | Lena Kristoffersen | Aug 12, 2026 | 58 | $120 | |
| Policy library cleanup and chunking | Policy assistant pilot for the service desk | Priya Natarajan | Sep 30, 2026 | 64 | $105 | |
| Evaluation set drafting with service leads | Policy assistant pilot for the service desk | Callum Reid | Oct 14, 2026 | 30 | $85 | |
| Dashboard build and finance walkthrough | Route cost dashboard | Callum Reid | Sep 16, 2026 | 120 | $85 | |
| Reconciliation fixes after review | Route cost dashboard | Callum Reid | Oct 18, 2026 | 18 | $85 | |
| Source system interviews | Warehouse and pipelines rebuild | Omar Haddad | Oct 15, 2026 | 26 | $95 |
Data and AI projects fail in predictable places: the data was not what anyone thought, nobody agreed what good looks like, or the client cannot run the thing after you leave. This workspace is built around those three risks.
The four tables
Engagements: one row per piece of client work. Stages follow how this work really runs: Discovery, Proposal sent, Data audit, Pilot, a Go or no-go decision, Production, then Handed over. Each shows fee type, contract value, whether data access is granted and whether personal data is in scope, plus hours, cost and delivery margin rolled up from time entries.
Clients: industry, data maturity and the executive sponsor, with engagements won and contracted value.
Deliverables: audit reports, gap maps, dashboards, models, evaluation sets, runbooks and training. Each has Accepted when written in the client's words and, for models and assistants, an Evaluation score.
Time entries: hours by consultant at their cost per hour, unbilled, invoiced or non-billable.
A normal week
Monday: the managing partner reads Live engagements and the Waiting on the client list. Anything in review over five days gets a call to the sponsor.
Each consultant logs time daily from the engagement with Log time.
Every model or assistant deliverable has its evaluation run before Client review, with the score recorded.
At the Go or no-go stage the sponsor decides with the pilot numbers in front of them; write the decision in the engagement page.
Friday: unbilled time is checked against fee type and invoiced.
Who owns what
Personal data
When Personal data in scope is ticked, data stays in the client's environment, access is named per person, and the handover includes how to remove our access.
| Engagement | Stage | Client | Offering | Engagement lead | Fee type | Contract value (USD) |
|---|---|---|---|---|---|---|
| Demand forecast for 40 stores | Handed over | Harbor and Hearth Home Stores | Omar Haddad | $68,000 | ||
| Data readiness audit, claims data | Handed over | Bluestem Health Plan | Lena Kristoffersen | $24,000 | ||
| Policy assistant pilot for the service desk | Pilot | Bluestem Health Plan | Priya Natarajan | $55,000 |
| Engagement | Stage | Client | Offering | Engagement lead | Fee type | Contract value (USD) |
|---|---|---|---|---|---|---|
| Demand forecast for 40 stores | Handed over | Harbor and Hearth Home Stores | Omar Haddad | $68,000 | ||
| Data readiness audit, claims data | Handed over | Bluestem Health Plan | Lena Kristoffersen | $24,000 | ||
| Policy assistant pilot for the service desk | Pilot | Bluestem Health Plan | Priya Natarajan | $55,000 |
| Deliverable | Status | Engagement | Kind | Due | Accepted on | Evaluation score (%) |
|---|---|---|---|---|---|---|
| Store-week forecast model and backtest | Accepted | Demand forecast for 40 stores | Aug 12, 2026 | Aug 19, 2026 | 91% | |
| Claims data quality report | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — | |
| Use-case gap map | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — |
| Work done | Billing | Engagement | Consultant | Logged on | Hours | Cost per hour (USD) |
|---|---|---|---|---|---|---|
| Feature engineering and backtests | Demand forecast for 40 stores | Omar Haddad | Jul 21, 2026 | 96 | $95 | |
| Store manager interviews | Demand forecast for 40 stores | Lena Kristoffersen | Jun 21, 2026 | 22 | $120 | |
| Profiling claims tables | Data readiness audit, claims data | Lena Kristoffersen | Aug 12, 2026 | 58 | $120 |
| Client | Status | Industry | Data maturity | Executive sponsor | Sponsor email | Start an engagement |
|---|---|---|---|---|---|---|
| Harbor and Hearth Home Stores | Delia Fontaine, COO | dfontaine@harborhearth.net | — | |||
| Bluestem Health Plan | Raj Malhotra, VP Member Services | rmalhotra@bluestemhealth.net | — | |||
| Copperline Freight Systems | Aoife Brennan, CFO | abrennan@copperlinefreight.net | — |
| Engagement | Stage | Client | Offering | Engagement lead | Fee type | Contract value (USD) |
|---|---|---|---|---|---|---|
| Demand forecast for 40 stores | Handed over | Harbor and Hearth Home Stores | Omar Haddad | $68,000 | ||
| Data readiness audit, claims data | Handed over | Bluestem Health Plan | Lena Kristoffersen | $24,000 | ||
| Policy assistant pilot for the service desk | Pilot | Bluestem Health Plan | Priya Natarajan | $55,000 |
Waiting on the client
| Deliverable | Status | Engagement | Kind | Due | Accepted on | Evaluation score (%) |
|---|---|---|---|---|---|---|
| Store-week forecast model and backtest | Accepted | Demand forecast for 40 stores | Aug 12, 2026 | Aug 19, 2026 | 91% | |
| Claims data quality report | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — | |
| Use-case gap map | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — |
Back for rework
| Deliverable | Status | Engagement | Kind | Due | Accepted on | Evaluation score (%) |
|---|---|---|---|---|---|---|
| Store-week forecast model and backtest | Accepted | Demand forecast for 40 stores | Aug 12, 2026 | Aug 19, 2026 | 91% | |
| Claims data quality report | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — | |
| Use-case gap map | Accepted | Data readiness audit, claims data | Aug 26, 2026 | Sep 2, 2026 | — |
Live engagements
| Engagement | Stage | Client | Offering | Engagement lead | Fee type | Contract value (USD) |
|---|---|---|---|---|---|---|
| Demand forecast for 40 stores | Handed over | Harbor and Hearth Home Stores | Omar Haddad | $68,000 | ||
| Data readiness audit, claims data | Handed over | Bluestem Health Plan | Lena Kristoffersen | $24,000 | ||
| Policy assistant pilot for the service desk | Pilot | Bluestem Health Plan | Priya Natarajan | $55,000 |
How to read the numbers
Most failed AI projects were never set up to succeed: the question was vague, the data was thinner than promised, or nobody said what result would be good enough. This page is the path every engagement takes before real money is spent.
Discovery (one to two weeks)
Write the business question in one sentence the sponsor agrees with, for example "Cut stock-outs in the top 200 products by a third."
Name the number that will move and how it is measured today.
List the systems, tables and documents involved, who owns them, and whether personal data is in them.
Data audit (two to three weeks)
The audit ends with a gap map: each gap, its effect on the question, and what fixing it costs. Sometimes the right answer is to stop here and fix the data first.
Pilot (four to eight weeks, fixed fee)
One use case, on real data, with the people who will use it.
Agree Accepted when before building: for a model, the error on held-out data; for an assistant, the share of an agreed question set answered correctly with a source.
Build the evaluation set with the client's experts, not from our own guesses.
Report the score honestly, including the cases it gets wrong.
The go or no-go meeting
Score against the agreed bar
Cost to run in production, per month
Who on the client side will own it
What it would take to reach the bar if it fell short
A no-go is a good outcome when the evidence says so. Write it down and close the engagement cleanly.
A model or assistant is only finished when the client's team can check it and run it without us. This page covers how we evaluate before review and what a complete handover contains.
Evaluation before every client review
Run the agreed evaluation set on the exact version being delivered.
Record the score in Evaluation score on the deliverable.
Read every failure, not just the total. Group them: missing source document, wrong number, out-of-date policy, misunderstood question.
Fix the biggest group first and re-run. Never tune on the evaluation set itself; keep a held-back slice nobody looks at until the end.
Move the deliverable to Client review only when the score meets Accepted when.
What we tell the client
Handover checklist
Runbook: how to run, monitor, re-train or update the documents
Evaluation set and how to re-run it
Dashboards or alerts for drift and errors
Training session with the people who will own it
Our access removed, listed person by person
Sponsor sign-off, then press Mark handed over
After handover
Offer a short advisory retainer for the first months. The usual questions are about drift: the business changed, the data changed, and the score quietly fell. A monthly re-run of the evaluation set catches it early.
Worked example
The service desk assistant pilot scored 78 percent against an 85 percent bar. Grouping failures showed most came from three outdated policy pages; once the client replaced them, the score rose to 84 and the next round targets the remaining phrasing gaps.
What's inside
5 tables · 14 views · 1 form · 4 row actions · 1 agent5 pages
- Data and AI Consultancy Desk
- Start here: how this workspace runs
- Dashboard
- From discovery to a go or no-go
- Evaluations and handover
4 tables
- Engagements3 views · 9 sample rows
- Clients2 views · 6 sample rows
- Deliverables7 views · 9 sample rows
- Time entries2 views · 12 sample rows
Related workspaces
App Development Agency StudioApp agency: projects, sprints, change requests, store releases, retainers and bugs.
SaaS Startup Operating SystemSaaS operating system: MRR by plan, churn by signup cohort, support SLAs and roadmap.
Security Consultancy Engagement DeskSecurity consultancy: engagements, scope, findings by severity, fixes and retests.
Employee DirectoryTeams of any size: find anyone by team, manager, skills and time zone.
IT Managed Services DeskFor MSP owners and techs: clients, devices, patching, tickets on SLA and agreements.
Performance Review CycleHR and managers: run review rounds with goals, feedback and sign-off.