Client and project health
A rating per client and project, based on current activity and open items.
Case study ยท Business intelligence & human oversight
Making Operational Intelligence Visible and Actionable
An inside look at the human-centered dashboard our team has built to support business decisions. It shows where clients need attention and leaves the decision with the person accountable for the account.
Client Health
Q3 review Client Health Deal Insights Proposals CFO Reports Deal Progression
Healthy
14
Watch
5
Open commitments
23
Client Owner Signal Health
- Northwind Freight J. Alvarez Two overdue commitments Watch
- Helix Dental Group M. Okafor Upsell signal: reporting module Healthy
- Corvane Logistics S. Patel No activity in 21 days Attention
- Bluepeak Software R. Lindqvist Proposal awaiting response Healthy
Pending for manager review
Suggested follow-up: Northwind Freight
Source: two open commitments past due. Draft message prepared, not sent.
Approve Edit Dismiss
Illustrative reconstruction. Companies are fictional and the figures are not real results.
The challenge
Client Health Cannot Be Read From One System
Client health and project status are rarely visible in a single place. Customer records sit in the CRM, project activity sits in task tools, commitments live in meeting notes, and manager concerns live in email.
Without a consolidated view, a risk is usually noticed late, and an upsell opportunity is missed because nobody connected the signals.
The solution
An Operations Dashboard Built for Managers
The dashboard is already built and in internal use. It is organized into views for different questions a manager asks each week.
- Client Health
- CFO Reports
- Proposals
- Deal Insights
- Deal Progression
Watch and risk signals
Accounts that have gone quiet, slipped on delivery, or have overdue commitments.
Open commitments
Promises made in meetings and emails, tracked until someone closes them.
Activity history
A timeline of interactions so context is visible before a conversation.
Ownership
Every account has a named person responsible for the next step.
Management attention areas
The short list of items that need a decision this week.
Potential upsell signals
Usage or request patterns that suggest a relevant service conversation.
The image above is an anonymized reconstruction with fictional companies, not a screenshot of internal data.
The broader architecture
Signals Become Recommendations. People Decide.
The intended operating model runs from signals in business systems through analysis and recommendations, to a human decision, and only then to an action. Each stage is logged so the decision can be reviewed later.
Individual AI agents and end-to-end automation steps are at different stages of development. The dashboard is built; some of the agents and automation steps around it are still being developed.
- Signals from business systems Business system
- Analysis of activity and records AI-assisted
- Recommendations with supporting evidence AI-assisted
- Human review by the account owner Human decision
- Approved actions run through n8n Deterministic logic
- Monitoring and feedback Business system
- Business system
- Deterministic logic
- AI-assisted
- Human decision
What this demonstrates
Useful AI Needs Context and Accountability
Business AI is most useful when it gives a manager context, supports an operational decision, and works within clear accountability. A recommendation is only valuable if someone knows who acts on it and why.
We have not measured business outcomes from this dashboard, so this case study describes the design and the operating model rather than claiming results.
Related
See How Human-Controlled AI Could Work for Your Business.
We can map the signals your managers rely on today and design the review steps that should sit between analysis and action.