by Ta-Tech Solutions All documents

02 - Competitive Landscape & Winning Strategy

CivicLoop by Ta-Tech Solutions Purpose: Know exactly who we are up against, where every one of them fails, and the specific lane CivicLoop takes to win the Prince George's County 311 / Community Relations engagement.


1. The gift: the incumbent is failing in public

Prince George's County currently runs PGC311 - a web portal (pgc311.com) plus iOS and Android apps, operated by the Office of Community Relations, built by a vendor called Connected Bits (the Android package is literally com.connectedbits.connectedcounty311).

It is failing, measurably, right now:

This changes our entire approach. We do not have to argue in the abstract that a better 311 is possible. We show PGC311 failing to submit a request, then submit the same request through CivicLoop in under a minute, by voice, in Spanish. A live, demonstrable incumbent failure is the rarest opening in government procurement. We build the whole pitch around it.

And the County runs a second, separate system for the other half

Prince George's County also runs MyPGC - "sign up for emails, alerts, notifications" - a separate opt-in resident-communication system, distinct sign-up, distinct vendor (the "emails, alerts, notifications" pattern is the signature of Granicus govDelivery, the outbound-comms market leader).

So the County's resident-facing service layer is split in two:

A resident must know about, and sign up for, both - and even then, the inbound system does not talk to the outbound one, so reporting a pothole does not subscribe you to news about that pothole. This is the fragmentation failure mode (Section 3) made concrete. CivicLoop's living record collapses inbound and outbound into one thing: report it, and you are automatically and specifically informed about it. CivicLoop can also feed MyPGC rather than replace it - the anti-lock-in posture in practice.

2. The competitive field

The 311 / citizen-services software market is real and crowded, but the leaders lead on distribution and procurement relationships, not product quality. That is the structural weakness we exploit.

The incumbents

SeeClickFix (owned by CivicPlus) - the category-defining resident-reporting app. Web + mobile, photo/GPS reports, public map, status updates. Widest small-and-mid-market footprint, strong resident brand recognition. But it is a reporting funnel, not an intelligent system: weak analytics and reporting, no integrated work-order or labor tracking, poor integration with other municipal systems (duplicate data entry), and effectively no AI. CivicPlus sells it inside a broad bundle (websites, agendas, notifications) where each module is middling.

Accela - owns the enterprise / large-city tier through a permitting beachhead, and absorbed PublicStuff (which now powers Philadelphia's 311 and, by app-store accounts, has been broken for months). Accela is high-cost and widely disliked: six-and-seven-figure deployments (Seattle spent $12.3M), steep learning curve, specialist-only configuration, long implementations, and high-profile botched launches - the Seattle permit system was called "a big disaster" in the local press. Small and mid-size jurisdictions cannot absorb that risk.

Granicus (govDelivery) - dominates outbound citizen communication (email/SMS/social mass comms, segmentation) and holds large-agency relationships. But its service-request product (govService) is far less mature than its communications products. Granicus owns proactive comms; it does not own service-request intelligence.

QScend / QAlert - a solid, call-center-oriented citizen-request manager, more usable and better-priced than Accela, now under Catalis. Dated UI, minimal modern self-service or AI.

GOgov, Comcate, Rock Solid / OneView - competent mid-market players for small-to-mid municipalities. Simple, affordable, thin on advanced workflow, analytics, and AI. Rock Solid is tied to the cost and complexity of Microsoft Dynamics.

Salesforce Public Sector Solutions - powerful and genuinely extensible with a strong security posture (Government Cloud), but it needs expensive system integrators, long timelines, and per-license cost. Overkill for a county that wants a working 311, not a platform project.

The AI-native challengers

Polimorphic - the best-funded newcomer ($28M raised, $18.6M Series A). A full AI platform: conversational chat, voice AI for phone calls, an AI "answers engine," forms, CRM, workflows, permitting. It publishes hard ROI numbers (claims of 90% voicemail reduction, 75% fewer walk-ins). This is the company to beat on AI. Its relative weakness: it is strong on the AI front door (intake, deflection, answers) and thinner on the back half - closed-loop, proof-of- resolution status updates and field-crew workflow.

Citibot - AI chat and voice for resident engagement, SMS/web, 24/7, 70+ languages. Genuinely AI-forward on conversational multilingual intake. Lighter on back-office workflow and CRM depth.

How AI-natives differ from incumbents

The incumbents bolt a chatbot onto a forms-based CRM and call it AI. The AI-natives treat AI as the front door - conversational intake over text, voice, and phone that deflects calls and answers questions instead of just filing tickets. The AI-natives are winning new contracts because they can show ROI numbers the incumbents cannot.

3. The documented failure modes - every way 311 systems break

Our research converged on a consistent set of failures. CivicLoop's architecture is organized to close each one.

Failure mode Evidence How CivicLoop closes it
The "black hole" - requests vanish, no status, no proof Chicago Inspector General found requests marked "completed" with work undone; NYC311 reviews report tickets closed within minutes, no work done; residents explicitly demand photo proof of resolution The living record (block 3) + proactive comms engine (block 6): every status change notifies the resident; resolution requires a proof photo
English-only / weak language access Multilingual jurisdictions running English-only apps; resident complaints about language barriers 35-language voice + text intake on every surface (blocks 1 & 6) - and in PGC, this maps to county law (section 5)
Broken / abandoned apps PGC311 at 1.5 stars, can't submit; PublicStuff broken for months; apps that redirect to buggy department forms A product that actually works, demoed live, on the proven Ta-Tech engine
Mandatory accounts, password loops Recurring resident complaint across systems Report without an account - phone number or anonymous; account optional, only for tracking convenience
Weak analytics Repeatedly cited for SeeClickFix and others - cities cannot see request patterns The director dashboard (block 7): category mix, SLA performance, geographic heat-map, trend lines
No work-order / back-end integration Academic case studies (NYC311, Philly311, SF311, Orange County FL) converge on this as the #1 city-side failure - the 311 CRM is not connected to departmental systems, so no real-time status Open311 API + integration layer (block 8); the agent console is the work-order surface for departments that lack one
Brutal implementations / cost overruns Accela-class failures; general government skepticism rooted in prior failed CRM projects Built on a proven engine, deployed in a focused scope, offered as a no-cost pilot - the County risks nothing
Departmental turf / routing resistance Case studies cite "liaising between the 311 center and municipal departments" as a perennial challenge AI auto-routing with visible rationale - the routing is transparent and overridable, so departments trust it instead of fighting it
Surprise SLA breaches Reactive-only systems; nobody sees the breach coming The prediction layer (block 5): SLA-breach forecasting early enough to act

4. The open lane - what nobody does well

Three capabilities exist in the market in pieces. No vendor combines all of them:

  1. AI multilingual voice + SMS intake - Citibot and Polimorphic do this well.
  2. Proactive, closed-loop, proof-of-resolution status - Granicus owns proactive comms but not service-request intelligence; the AI-natives own intake but their resolution-stage updates are thin.
  3. Offline / low-bandwidth operation - barely addressed by anyone; most systems assume a smartphone with good connectivity.

The intersection of those three, plus genuine analytics, is empty. CivicLoop takes the intersection. That is not a feature list - it is a defensible position, because the incumbents would have to rebuild their core to follow us there, and the AI-natives would have to build the back half they skipped.

4a. Four "no rival has this" differentiators

Beyond the three-way intersection above, CivicLoop now carries four operational features no surveyed rival (SeeClickFix, Accela, Polimorphic, Citibot, QScend, GOgov, Comcate, Rock Solid, the rushilpatel21/CivicPulse open-source clone, or the Romulus 311 stack) ships. Each is concrete, demoable, and shipped today.

  1. Autopilot dial. A single dial on /admin, set by the county admin: off (no automation), route (AI auto-routes to department), full (AI auto-routes AND auto-assigns to the lightest-loaded agent AND sends the assigned-resident notification). Every step is logged with actor='ai' and AUTOPILOT: prefix and is reversible. No rival lets a county choose its own automation level on a single, governable dial. Polimorphic delivers AI as a black box you cannot tune; the incumbents deliver no automation at all.
  2. Equity dashboard. Median resolution + SLA on-time percentage, broken out by council district, with a one-line headline ("District X waits 2.3x longer than District Y"). Turns the County's own data into a public-trust artifact. No surveyed rival ships an equity panel as a first-class dashboard tile - they ship "analytics" as raw tables a Director has to read.
  3. Public transparency portal. /public, anonymous, no account, no PII, shows last-7-day filed / resolved / open / median resolution / SLA on-time / NPS, top categories, by-district, an anonymized SVG scatter map, plus a "save / print PDF" button and a CSV at /api/public/weekly.csv. SeeClickFix has a public map but it is a marketing surface, not a governance surface; nobody else exposes weekly SLA and NPS to the public in machine-readable form.
  4. Predictive issue forecast. Per (category, council_district), predicted_issues holds a 7-day projection computed from a 12-week mean times recent seasonality, with a confidence score and a director "Run forecast now" button. Polimorphic talks about intake AI; nobody surveyed ships an outbound operational forecast a Director can act on.

These four are the differentiators we lead with in objections about "every 311 vendor says AI." We say: every vendor says it; here are four screens you can point at, today, that none of them ship.

5. The Prince George's County wedge - language access is the law

PGC is the most populous Black-majority county in the United States: roughly 58-60% Black, about 22% Hispanic/Latino, about 25% foreign-born. The County school system already serves 20+ languages - Spanish, Amharic, Haitian Creole, Pashto, French, Yoruba, Mam, K'iche', Vietnamese, and more.

Critically, the County has a Language Access for Public Services Act (2017, amended 2020) that legally requires county agencies to provide enhanced language access.

This reframes CivicLoop's multilingual capability entirely. It is not "nice tech." The incumbent PGC311 app is English-centric and cannot even submit a request. CivicLoop's 35-language voice and text intake is the only path we are aware of by which the County's 311 function actually complies with the County's own language-access law. We pitch it as compliance, not as a feature.

6. Procurement table stakes - what we must have to be taken seriously

These are not differentiators. They are the price of being in the room. Documents 09 and 08 specify how we meet each.

7. Open311 / GeoReport v2 - our interoperability and anti-lock-in play

Open311 / GeoReport v2 is an open, collaborative REST API standard for civic issue reporting. It defines a common interface for viewing service definitions and submitting and tracking service requests, replacing proprietary forms and phone-only intake. It has been supported since around 2010 by San Francisco, Washington DC, Boston, and dozens of other cities, and by vendors including SeeClickFix and Connected Bits.

Committing to Open311 GeoReport v2 lets CivicLoop:

Our positioning line: open standard in, CivicLoop intelligence and closed-loop status on top. The County keeps its data and its freedom; we earn the renewal on results, not on lock-in.

8. The six strategic wedges - how we win, in one place

  1. Attack the incumbent head-on. Demo PGC311 failing to submit; then submit through CivicLoop, by voice, in Spanish, in under a minute. Show, do not argue.
  2. Kill the black hole. Make closed-loop status with proof-of- resolution the headline. "Your request is a living record, not a black hole."
  3. Make language access the wedge. Reframe 35-language voice intake as compliance with the County's own Language Access Act, not as a feature.
  4. Be the anti-lock-in vendor. Commit to Open311. Open standard in, intelligence on top, data exportable, no hostage-taking.
  5. Out-trust, do not out-feature. We will not win a feature-count war on day one. We win on: it works, residents trust it, it meets county law, the County owns its data.
  6. Commit to hard ROI. Agree to measurable pilot targets - call deflection, first-pass routing accuracy, SLA-breach reduction, languages served, NPS - and be judged on them.

Two new wedges (7 and 8) have emerged from the shipped wave, and they are the ones we close on:

  1. Lead with Autopilot. The closing demo move is the dial: flip it from off to route to full while the panel watches the next inbound request route, assign, and notify itself in under ten seconds. No rival lets a County dial its own automation in public; this is the moment that ends the "is the AI real" debate.
  2. Lead with Equity and the Public Portal. The Equity panel answers "is the County treating my district fairly" with the County's own data, and the Public Portal puts the answer on a public URL the press and the Council can read without an account. No surveyed rival turns a 311 system into a governance and public-trust artifact this directly.

9. The platform moat - the incumbents sell an app, we start a platform

Every competitor in Section 2 sells the County a product: a reporting app, a CRM, a comms tool. CivicLoop's positioning is structurally different and harder to compete with: 311 is the front door to a County Intelligence Platform (Document 01, Section 9).

This is a moat, not a slogan:

We say it to the panel plainly: the County is not buying a 311 app. It is starting a County Intelligence Platform, and 311 is the door that is open today. That reframes the whole evaluation - from "which 311 app is best" (where incumbents have a distribution head start) to "who can give us a coordinated County" (where we are the only credible answer in the room).

10. What this means for the build

Every one of the eight system building blocks in Document 01 now has a documented competitor failure it is designed to close. The build is not a list of features we think are interesting. It is a systematic shutdown of every known way a 311 system fails a county and its residents. The remaining documents - domain model, roles, architecture, screens, AI components, integrations, security - each trace back to this strategy.


Next: 03 - Domain Model - every entity, every relationship, and the lifecycle of a service request.

PreviousVision & Solution Overview
CivicLoop - Ta-Tech Solutions - Architecture & Design Documentation