Communication Intelligence
Thirty-five years ago, the question was “what did that call cost?” Today it’s “what is about to happen to our customers, and who needs to know right now?” Here is how RSI got from one to the other, and why it matters to your bottom line.
By call accounting blog · November 3, 2026 · 9 min read
Picture a finance manager in 1995 with a thick printout from the PBX. Every line is a call: extension, number dialed, duration, cost. It was tedious, and it was also revolutionary. For the first time, a business could see where its telecom money went.
That ledger mindset, record it, total it, bill it back, is where call accounting began. It is also where Resource Software International (RSI) began, founded in 1990. What has changed since isn’t the instinct to measure. It’s everything that now needs measuring.
01 · THE EVOLUTIONFive eras of one idea
The ledger
Record every call, allocate cost to the department, client or room. Catch toll fraud when the bill arrives.The report
Cradle-to-grave call tracking. Long-duration, missed and abandoned calls become visible, with reports scheduled to land in the right inbox.The live view
Real-time dashboards and alerts that fire within seconds, so problems are handled before a customer feels them.The conversation
Recordings, transcripts and sentiment. Not just that a call happened, but how it went.The ecosystem
Calls, messages, meetings, contact center queues and AI agents, understood together, across platforms, and askable in plain English.Each era didn’t replace the last. It absorbed it. Cost allocation and billing integration are still there, now sitting beside journey analytics, omnichannel views and agentic AI reporting. That layering is why RSI describes its direction as AI-driven communication intelligence rather than simply analytics.
02 · THE COMPLEXITYOne PBX became an ecosystem
The modern enterprise rarely runs one system. A company may have a cloud UC platform for employees, a contact center platform for customers, Microsoft Teams for collaboration, Zoom for meetings, a legacy PBX at a branch office, and now AI agents answering a share of inbound contacts. Each one reports on itself, in its own vocabulary, with its own definition of “abandoned.”
When every platform keeps its own score, nobody can tell whether the team is winning.
That’s the gap RSI’s Shadow portfolio fills: a single analytical lens that sits across the platforms you already own.
03 · ACROSS PLATFORMSOne lens, many platforms
RSI offers one lens for many platforms (premise and cloud) including Avaya, Cisco UCM and Webex Calling, Zoom, Intermedia, Mitel, Asterisk, RingCentral (and its brands Avaya Cloud Office, AT&T Office@Hand, BT Cloud Work, Telus Business Connect, Unify Office and Verizon).
The portfolio itself is deliberately modular, so each organization takes what fits:
CoreShadow Analytics
Traffic analysis, call journey, live monitoring, reporting, omnichannel and agentic AI views, and the AI Builder.
CX-liteShadow AgentCX
Dispositions, break states, callbacks, campaigns and VIP “cherry pick” on everyday RingCentral queues, without full contact center seat licenses.
NotificationsShadow Bot
An embeddable team-messaging bot for real-time notifications, dispositions and break codes.
AI connectionShadow Analytics MCP
Lets AI tools such as Claude, ChatGPT and Microsoft Copilot query call volumes, agent performance and journey data on demand.
The benefit of a cross-platform lens isn’t just convenience. It’s comparability. When the same definitions apply to the branch on Avaya and the headquarters on RingCentral, leadership stops arguing about whose numbers are right and starts discussing what to do.
04 · OMNIDIRECTIONALEvery direction, every channel
Traditional call accounting looked in one direction: outbound calls, because that’s where the cost was. Communication intelligence looks everywhere:
Inbound
Demand, queues, abandonment, service levels, repeat callers.
Outbound
Campaign dialing, account codes, cost and connect outcomes.
Internal
Team messaging, meetings and adoption of the tools you pay for.
Human ⇄ AI
Calls handled by AI agents, handoffs to people, resolution and satisfaction side by side.
And the data itself flows in every direction too. It comes in from your platforms, moves across them into one journey view, and goes out to the places decisions happen: scheduled reports by email, webhook or SFTP, live boards on a wall TV, alerts in a chat channel, and now your own AI assistant through MCP.
The call journey, end to end
Take a single customer call. It hits an IVR, lands in a queue, is answered by an AI resource, is transferred to a billing specialist, and is put on hold twice. Siloed reporting sees three disconnected fragments. Shadow Analytics follows the whole path, from IVR to agent to AI resource to resolution, and lines it up with the recording and transcript, so the question “what happened on that call?” has one answer.
05 · UNIFIED EXPERIENCESOne place to look, one place to act
“Unified” gets overused, so here’s what it concretely means inside Shadow Analytics. The Omnichannel page shows every channel’s contacts for a period on one screen, with how many were resolved and where contacts slipped through. The Agentic AI page compares interactions AI took part in with those only people handled, by volume, resolution, satisfaction and speed. Interaction Insights shows how customers felt across transcribed calls by queue, agent and keyword. Searchable message history means a disputed text isn’t a scavenger hunt.
On the agent side, AgentCX brings those same lenses to the people doing the work: live availability, queue KPIs, leaderboards, disposition codes and timed wrap states. For organizations on RingCentral’s standard queues, this is a practical way to get contact-center-style discipline without buying contact-center-style licensing for everyone.
06 · FORECASTINGFrom looking back to looking ahead
Reporting tells you what happened. Forecasting asks what that history implies. Shadow Analytics supports it from several angles:
DemandTraffic Analysis
Busiest days and hours, repeat callers and call sources, highlighting volume trends, peak periods and capacity needs.
TrendsHistorical trend lines
Any time period of historical metrics, shaped by seasonal change, workforce levels and business growth.
Early warningEcosystem health
Train the system to watch for bottlenecks, anomalies and potential outages across your environment.
Ask anythingAI Builder
Describe the question in plain English and get a report or dashboard without waiting for a developer.
A fair caveat: a forecast is only as honest as the history under it. Standard plans keep one year of call history, which covers one full seasonal cycle. Plan your long-range forecasting with that in mind, and use the report repository if you need to keep results longer.
Perspectives: the same data, from your seat
A contact center director and a night-shift supervisor shouldn’t get the same answer to “how are we doing?” In Shadow Analytics, a Perspective tells a connected AI assistant what to lead with, which trade-offs to favor and what to avoid: service levels or cost, operational or strategic, how far ahead to look, whether forecasting comes first. The assistant adapts its emphasis to you. It never changes the figures, and that’s the point. Everyone starts from the same facts and gets a view fitted to their job.
07 · USE CASES & ROIWhat it’s worth
ROI claims in this industry are often vague, so here are five use cases worked through with arithmetic you can check. These are illustrative models, not customer results. Swap in your own volumes and rates; the structure is what matters.
Use case 1 · Contact center
Recover the calls you were losing
Abandoned calls are customers who tried to reach you and gave up. With abandonment visible by queue and hour, and AgentCX callbacks that return abandoned and voicemail callers, you can win a share of them back.
2,000 calls/month × 8% abandoned = 160 lost
Recover 25% = 40 callers reached
20% convert × $1,500 average value = 8 deals
≈ $12,000 per month in recovered revenue
Use case 2 · Operations
Staff to demand, not to habit
Traffic analysis shows which days and hours actually drive volume. If a team is overstaffed in quiet windows and stretched at peak, moving schedules costs nothing.
100 agent-hours/month moved from quiet to peak windows
× $30 loaded hourly cost avoided in overtime or temp cover
≈ $3,000 per month, with shorter waits as a bonus
Use case 3 · Cost control & security
Catch toll fraud in minutes, not at invoice time
Real-time alerts fire within seconds when a live measure crosses a line you set, such as international call volume or long-duration calls.
5 fraudulent concurrent calls × $0.50/min = $150/hour
Found on the monthly bill after 48 hours: $7,200
Caught by an alert in 15 minutes: $37.50
≈ $7,160 avoided in a single incident
Use case 4 · Professional services
Bill the time you were already spending
AgentCX can restrict or tag outbound calls with client/matter account codes. Phone time becomes attributable, and unbilled time becomes billable.
10 fee earners recover 15 minutes/week each = 2.5 hours/week
× $300/hour × 50 weeks
≈ $37,500 per year
Use case 5 · Management time
Stop building reports by hand
Hundreds of pre-built templates, scheduled delivery and the AI Builder turn a weekly spreadsheet chore into something that arrives on its own.
6 hours/week manual → 1 hour/week = 5 hours saved
× $45/hour × 52 weeks
≈ $11,700 per year
| Use case | Lever | Illustrative value |
|---|
| Abandoned-call recovery | Callback + queue visibility | $144,000 / yr |
| Demand-based staffing | Traffic analysis | $36,000 / yr |
| Toll-fraud alerting | Real-time alerts | $7,160 / incident |
| Billable call time | Account codes | $37,500 / yr |
| Reporting time | Scheduling + AI Builder | $11,700 / yr |
Annual figures are the monthly models × 12. Assumptions are hypothetical and will differ by organization.
ROI = (Annual benefit − Annual cost) ÷ Annual cost
For the cost side, ask RSI or your reseller for a quote sized to your platform and seats, then run your own numbers through the formula. Deliberately conservative inputs still tend to tell a clear story, because most of the benefit comes from a handful of behaviors: fewer lost customers, better-matched staffing, faster detection and less manual effort.
08 · WHAT’S NEXTIntelligence you can talk to
The most interesting shift is that analytics no longer has to be a destination you visit. With Shadow Analytics MCP, the people in your organization can ask a question in the AI tool they already use, “which queue had the worst abandonment last week and what changed?”, and get an answer drawn from live communications data, in the framing their Perspective prefers. No custom integration project, and the same figures as the dashboards.
That’s the arc from ledger to intelligence. The printout became a dashboard, the dashboard became a journey, and the journey is becoming a conversation. What stays constant, thirty-five years on, is the original instinct: measure what matters, then act on it.
See your ecosystem as one picture
Book a walkthrough of Shadow Analytics, Shadow AgentCX and Shadow Analytics MCP on your platform.
Request a demoExplore productsResource Software International Ltd. (RSI) · Oshawa, Ontario, Canada. Platform names and logos belong to their respective owners. Use-case figures are illustrative models, not guarantees or customer results.
#BLOG #POPULAR
#CALLACCOUNTING
#CALLACCOUNTINGSOFTWARE #SHADOWANALYTICS
#SHADOWMCP
#COMMUNICATIONINTELLIGENCE
