Use Cases#
This page groups practical Admin UI scenarios and common use cases supported by OptScale AI. The examples below illustrate how the platform fits into different workflows and supports everyday operational and development tasks.
Dashboards#
Use Home → Dashboards for at-a-glance monitoring without opening Analytics → Usage and Cost or Analytics → Traces for every check. Start from the shared DEFAULT tab for organization-wide coverage, or + CREATE DASHBOARD for role-specific layouts. For panel types and the DEFAULT layout, see Core Services — Home / Dashboards and Dashboard panel types.
| Use case | Typical panels | Dashboard pattern |
|---|---|---|
| Daily platform health check | Total requests, Success rate, Avg latency | Open DEFAULT; set time range to 24h; use REFRESH or auto-refresh during active rollout |
| FinOps and spend review | Total spend, Cost breakdown, Models breakdown, Users breakdown | DEFAULT or a pinned custom tab; use 7d or 30d to review trends before budget meetings |
| Find top cost drivers | Models breakdown (Spend), Users breakdown (Spend) | Custom dashboard for finance or platform owners; rank by Spend to target model or user optimization |
| Track token growth | Total tokens, Total input tokens, Total output tokens | Compare cards and the Total tokens chart over 7d / 30d for capacity and quota planning |
| Governance and compliance monitoring | Top guardrail violations, Success rate | After policies and guardrails are enabled; use 7d to spot recurring violation types |
| Validate a new model or provider | Model reliability hotspots, Success rate, Avg latency | Custom or DEFAULT tab with 24h after rollout; drill into failures in Traces when hotspots appear |
| Incident or outage triage | Success rate, Total requests, Model reliability hotspots | Short window (1h); REFRESH frequently; pair with Traces for per-request detail |
| Evaluate prompt caching | Cache efficiency | Add to a custom FinOps dashboard when repeated prompts or shared context are in use; measure dollar savings on Optimizations |
| Executive or org-wide overview | DEFAULT summary cards and charts | PIN DASHBOARD on DEFAULT (or a curated custom tab) so operators land on the same view |
| Personal operator workspace | Mix of cards and charts for your responsibilities | + CREATE DASHBOARD on a MY tab; EDIT layout and Panel type per tile |
| ML training and experiment oversight | Recent tasks, Recent runs, Runs activity, Recent models, Model versions | Custom dashboard for ML leads; combine with Experiment Tracking for run detail |
| User adoption and activity | Users breakdown (Total tokens), Total requests | 7d / 30d on DEFAULT or a team-facing custom dashboard to spot heavy usage or idle accounts |
Allow time for new traffic to appear after you change providers, routing, or policies before expecting panels to update. Use Usage and Cost for tabular drill-down and Traces when a dashboard metric needs request-level evidence.
Usage and Cost#
Use Analytics → Usage and Cost for detailed FinOps analysis by tab—organization-wide trends, cost drivers, and scoped activity. For page layout and chart descriptions, see Core Services — Usage and Cost.
Organization totals#
Use the TOTALS tab as the starting point for AI usage analysis. For chart descriptions, see Organization totals.
| Use case | What to review | What to compare / look for |
|---|---|---|
| Monitor AI adoption | Total Requests, Total Tokens | Whether request volume and token consumption are increasing, decreasing, or stable over the selected period |
| Analyze token usage | Total Tokens | Input vs output vs total vs cache read tokens—whether prompts or responses grow, and whether cache reads reduce consumption |
| Monitor request trends | Total Requests | Successful, failed, and total requests—demand spikes or changes in usage patterns |
| Detect platform issues | Total Requests | Rising failed requests or falling successful requests (provider outages, config problems, application errors) |
| Verify optimization effectiveness | Total Requests and Total Tokens | Stable request volume with decreasing token consumption (caching or prompt compression impact) |
| Identify unusual activity | Both charts | Unexpected jumps in requests, tokens, or cache reads (new workloads, behavior changes, inefficient prompts) |
| Establish a performance baseline | TOTALS, then drill down | After spotting a trend, continue in Cost or an activity tab to find which models, users, agents, or teams drive it |
Cost breakdown#
Use the COST tab to analyze spending trends and cost drivers. For chart descriptions, see Cost breakdown.
| Use case | What to review | What to compare / look for |
|---|---|---|
| Identify the most expensive AI models | Models Breakdown | Models that dominate spend or tokens; whether premium models are used where cheaper alternatives may work |
| Identify high-cost users | Users Breakdown | Power users and unusually high spend or token use; support for chargeback and adoption analysis |
| Monitor spending trends | Cost Breakdown | Rising or falling spend, spikes, and links to usage, model choice, or policy changes |
| Compare cost with usage | Spend vs Total tokens | Whether spend rises from premium models or from higher request/token volume |
| Validate optimization initiatives | Cost trends over time | Spend before/after caching, prompt compression, routing, model switches, or new assistants |
| Support budgeting and chargeback | Models and users breakdowns | Cost allocation by model and user; estimates for future budgets |
| Detect unexpected spending | Cost trends and breakdowns | Sudden spend or token growth, or a few users/models with disproportionate cost |
Activity tabs#
Use an activity tab when you need detail for one model, user, agent, or team. For page layout and filters, see Core Services — Activity tabs.
| Tab | Use when you need to |
|---|---|
| Model Activity | Compare cost, reliability, and token patterns for a selected model; validate routing or prompt changes for that model |
| User Activity | Investigate power users, preferred models, policy compliance, or training needs for an individual |
| Agent Activity | Measure cost and reliability of an AI application; see which models an agent uses; troubleshoot agent-specific failures |
| Team Activity | Allocate departmental spend, compare team adoption, and measure impact of AI initiatives at team scope |
Model Activity#
| Use case | What to review | What to compare / look for |
|---|---|---|
| Evaluate model adoption | Request volume, token consumption over time | Whether usage is rising or falling and whether the model still fits organizational needs |
| Measure model cost | Cost Breakdown | Expensive models, cost spikes, impact of introducing or replacing a model |
| Analyze token consumption | Total Tokens (input, output, total, cache read) | Large prompts or responses, cache efficiency, prompt-optimization opportunities |
| Monitor model reliability | Successful/failed requests, success rate, provider latency | Outages, rising errors, or performance degradation |
| Compare models before standardization | Usage, cost, token efficiency, latency, reliability | Best balance of performance, quality, and cost by workload |
| Validate optimization changes | Metrics before and after prompt, routing, provider, or optimization changes | Lower cost, better token efficiency, lower latency, or higher success rate |
| Investigate anomalies | Spend, tokens, volume, latency, failures | Unexpected model-level behavior before drilling into users, agents, or teams |
User Activity#
| Use case | What to review | What to compare / look for |
|---|---|---|
| Analyze individual AI usage | Requests, tokens, spending | Personal adoption trends and resource intensity |
| Understand model preferences | Model Usage | Which models the user uses most; alignment with approved catalog |
| Monitor user spending | Cost for the selected user | Power users and unusually high AI costs |
| Analyze token consumption | Total Tokens | Resource efficiency and optimization opportunities for that user |
| Verify user experience | Success rate, failed requests, provider latency | Reliability issues affecting an individual |
| Investigate unusual usage | Jumps in requests, tokens, or spend | Unexpected behavior or inefficient prompt patterns |
| Support governance and adoption | Usage and model preference patterns | Policy compliance and users who may need training |
Agent Activity#
| Use case | What to review | What to compare / look for |
|---|---|---|
| Monitor agent usage | Requests, tokens, spend over time | Adoption trends and operational impact of the application |
| Understand model utilization | Model Usage | Model mix and whether premium models are used appropriately |
| Track agent costs | Cost Breakdown | Financial impact of features, adoption growth, or workflow changes |
| Analyze token consumption | Total Tokens | Growing prompt/response sizes and cache efficiency |
| Verify agent reliability | Success rate, failed requests, provider latency | Provider or application issues affecting end users |
| Evaluate optimization changes | Metrics before/after prompt, workflow, routing, or model updates | Efficiency gains without harming performance |
| Troubleshoot agent-specific issues | Failures, latency, tokens, costs | Problems scoped to one application rather than org-wide |
| Compare AI applications | Usage, cost, model mix, efficiency, reliability | Scale, improve, or retire decisions across agents |
Team Activity#
| Use case | What to review | What to compare / look for |
|---|---|---|
| Analyze AI adoption | Requests, tokens, spend over time | Growth within the team and across departments |
| Understand model usage | Model Usage | Alignment with recommended models; optimization opportunities |
| Monitor team spending | Cost Breakdown | Budget trends, spikes, departmental AI cost evolution |
| Analyze token consumption | Total Tokens | Efficiency, prompt/response size changes, caching impact |
| Monitor operational health | Success rate, failed requests, provider latency | Issues that may reduce team productivity |
| Compare team performance | Usage, spend, efficiency, model selection, reliability | Best practices and cross-team optimization |
| Support budgeting and governance | Team spend and usage | Chargeback, policy compliance, teams needing training |
| Measure AI initiative impact | Metrics before/after new tools, workflows, or policies | Whether initiatives improve adoption, efficiency, and cost control |
Pair Usage and Cost with Dashboards for at-a-glance health, Optimizations for savings impact, and Traces for request-level evidence.
Optimizations#
Use Optimizations when you need visibility into how much optimization features save—not just what you spent. Enable context compression in AI Access turns on optimization mode for a user or team; the page then compares Would have paid vs Actual cost and breaks savings down by technique (Cache read, Prompt compression, Memory retrieval), model, and provider. For page layout and filters, see Core Services — Optimizations.
| Use case | What to review | Typical outcome |
|---|---|---|
| Measure AI cost savings | Saved, Would have paid, Actual, Projected annual savings | Answer how much you saved and what the bill would have been without optimizations |
| Analyze optimization effectiveness | Savings breakdown chart by technique | See whether cache read, prompt compression, or memory retrieval drives the most savings |
| Monitor optimization trends | Savings breakdown over 30d / 90d / 180d / 1y | Spot usage growth, efficiency changes, or regressions after config updates |
| Compare models | Model table — Would have paid, Actual, Saved, Savings % | Find expensive models and highest ROI from optimization |
| Evaluate provider efficiency | Filter by Provider; compare Top optimization and Breakdown | Compare Anthropic, OpenAI, Groq, and other providers for placement decisions |
| Identify optimization opportunities | Models with low cache use, heavy prompt compression, or no memory retrieval | Prioritize caching, compression, or memory configuration changes |
| Report AI cost optimization | Summary cards — Projected annual savings, Saved, token counters | KPIs for FinOps reviews and leadership reporting |
| Validate optimization rollout | Savings breakdown and model table before/after enabling Enable context compression | Confirm measurable savings after turning on optimization mode in AI Access |
| Support capacity and budget planning | Projected annual savings and historical Saved trends | Forecast AI spend and savings as usage scales |
| Persona | How they use the page |
|---|---|
| Platform administrator | Monitor optimization performance and tune strategies per model or provider |
| FinOps / cloud cost manager | Track realized savings and avoided cost alongside Usage and Cost |
| Engineering manager | Identify high-cost models and workloads that need more caching or compression |
| AI platform owner | Measure ROI of optimization features and justify infrastructure investment |
| Executive / leadership | Review Projected annual savings and efficiency trends at a glance |
Pair Optimizations with Usage and Cost for total spend and Traces when you need request-level evidence for a specific saving.
Policies & Guardrails#
| Use case | Typical guardrails | Policy pattern |
|---|---|---|
| Redact PII in user prompts | PII detection and redaction on Input, action Redact | Match Chat Completion (or your API request type); Stage Input; Sampling rate 100% for compliance-sensitive workloads |
| Block credential leaks | Secrets on Input and/or Output | Organization-wide or team-scoped conditions; block or redact before content reaches the provider or the client |
| Restrict unsafe or off-topic content | Ban topics, Toxicity, or Sentiment | Scope by Request type, Provider, or Team when only certain workloads need stricter rules |
| Harden against prompt abuse | Prompt injection, Jailbreak, Invisible text on Input | Apply to externally facing Chat or API traffic; start with moderate thresholds and tune from Violation rate |
| Control response safety | Toxicity, Code injection, or Secrets on Output | Stage Output so model responses are checked before they return to the user or calling application |
| Limit oversized requests | Token limit | Match high-volume or untrusted entry points; enforce limits before expensive provider calls |
| Layered governance | Multiple guardrails on one policy (for example PII detection and redaction + Secrets on Input) | One policy with shared Conditions; each linked guardrail keeps its own Type, Threshold, and action |
Reuse guardrails across policies when the same control applies to different scopes (for example one Secrets guardrail linked from both a Chat policy and an API policy with different conditions).