mPulse

The model is the engine of a core module, and the platform carries other value. Predictive Analytics is one of the company's named products: its Decision Point Models page describes more than 140 predictive models, built with machine learning, that estimate each member's risk of missing quality measures, adverse health outcomes and disenrollment, and says outcome data from each campaign feeds back into the models. That module does not exist without them, and it decides which members are contacted, through which channel and when.

The rest of the platform stands on its own: core communications at a scale of billions of touchpoints a year, member and provider portals, a health content library, print and mail, and acquisition and payment tools, most of it arriving through acquisitions since 2023. A conversational layer answers member messages on top.

Nothing published says what share of a program is model driven targeting and what share is scheduled campaign delivery, and the second is a communications operation with a conversational layer.

Substantial autonomy over contact, with a described human path. Campaigns run automatically at scale, the conversational layer answers member messages and returns resources without a person, and published material shows programs handing off mid conversation to a live agent when the member needs one.

The autonomy that matters is upstream. The system decides which members are contacted, through which channel and at what moment, and a member the model does not select receives nothing. The automated actions are communications rather than clinical decisions, and escalation to a person is a described feature rather than an inference.

The approach is named without the discipline behind it. The Predictive Analytics page describes more than 140 predictive models, built with machine learning and AI, estimating each member's risk of missing quality measures, adverse health outcomes and plan disenrollment; population and member views that filter on hundreds of characteristics and model predictions; and a closed loop in which campaign outcome data feeds back into the models. The platform also carries conversational AI and natural language processing for two way member messaging.

What is missing is everything a buyer would use to check it: no feature list, no training data description, no validation method or accuracy figure for any of the 140 models, no intent recognition or escalation rate for the conversational layer, and no version or retraining schedule, so a customer cannot tell when a model that decides who is contacted has changed. For a system whose main effect is selecting members, the feature set is the substantive disclosure.

Nothing establishes who else sits between a member record and an answer. The home, predictive analytics, terms and privacy pages name no model provider, cloud host or subprocessor. The terms refer to the company's third party licensors, who share its warranty disclaimer and liability cap, without naming them; the privacy policy names Google Analytics only as an optional, client controlled integration on the white labeled enrollment platform, capturing non identifying data.

The chain matters here because it is long by construction. The platform delivers across text, voice, web, portals and physical mail, which implies carriers, messaging aggregators and print operations handling member contact details and message content, and its conversational layer may call external language services. Three acquisitions since 2023 add inherited infrastructure a buyer cannot see.

Real deployments with large denominators, none of them named. The company's CAHPS case study describes a managed care organization serving about 20 million members across 16 states, where predictive models scored 541,316 members across 13 Medicare contracts, members at high risk of a negative survey response were contacted by a plan representative, and a mock CAHPS survey measured the effect afterward. Other published program results include a 1 star CAHPS improvement at a leading national health plan and a 46 percent increase in proxy survey completion, and the company states it serves more than 450 organizations.

The vendor is the claimed source and the plans are anonymous: figures come from the company's own deployments, with enough method in places to see what was done but no named customer and no independent evaluation. The Forrester Wave recognition the company cites is an analyst assessment of the offering and strategy, not evidence of what the product does for members. The endpoints measured are survey scores, retention and return on investment; nothing published measures whether members were healthier.

General assurances and one specific boundary, with the AI questions unanswered. The privacy policy commits that consumer data on the white labeled enrollment platform is not sold or used for marketing or advertising and is not linked to the corporate website, and sets out GDPR data subject rights with a data protection officer; the terms say inactive subscriber data may be removed after 60 days. The Predictive Analytics page says campaign outcome data feeds back into the models, so member responses do train them.

What is not published: whether that training pools data across client plans, how long message and call content is kept, and how the member records and communication histories brought in by three acquisitions since 2023, including 150 million member communications with one of them, are separated and governed after integration. Those records were collected under different agreements with different customers. No safety engineering is published either: how a mis-targeted campaign or a wrong conversational reply is detected and handled.

A status is stated and not supported. The home page describes the platform as HITRUST certified and HIPAA compliant, and the footer carries a HIPAA badge. The privacy policy published at mpulse.com/privacy sets out data subject rights under the GDPR, names a data protection officer at compliance@mpulse.com, and describes the white labeled enrollment platform's separation from marketing, but does not mention business associate status, protected health information or business associate agreements. No agreement, subcontractor position or breach notification commitment appears on the home, predictive analytics, terms or privacy pages.

A HIPAA compliance claim is a self assertion; the HITRUST certification beside it is an outside assessment of security controls, which speaks to security rather than to HIPAA terms.

The platform handles member eligibility, claims and communication data for health plans, so business associate status is the expected posture and the terms of the agreement are the buyer's question.

Recognized certifications are named without the artifacts. The home page describes the platform as HITRUST certified, and the site footer carries an r2 HITRUST Certified mark and the AICPA SOC badge for service organizations, both presented as the company's own. HITRUST r2 is the assessment's most rigorous tier, which is a meaningful statement for a company handling health plan member data at this volume.

What is not published: the certification's scope (which products and environments it covers, including the portals and print operations acquired since 2023), its date and expiry, which SOC report the badge refers to and its type and period, and any trust portal from which a buyer could retrieve them. The privacy policy names a data protection officer at compliance@mpulse.com.

Population health targeting with operational recommendations, and no device position claimed. The predictive models estimate each member's risk of missing quality measures, of adverse health outcomes and of disenrolling, and the output is who to contact, through which channel and when: a ranking that recommends outreach rather than treatment, diagnosis or testing, which is population health rather than device functionality.

Three other regimes apply and matter more. Telecommunications rules govern automated outbound texts and calls, and the terms require clients to hold the express written consent federal rules demand and to honor opt outs. Medicare marketing and communication rules constrain what a Medicare Advantage plan may say to members and when, particularly around enrollment. And the Star Ratings system, including the federal experience survey, defines the measures the platform optimizes toward.

Nothing published says how the platform enforces consent and Medicare communication rules in each campaign.

Nothing published on how model behavior is governed or tested, and the company's own case study describes targeting logic precise enough to examine. In its CAHPS case study, predictive models scored 541,316 members of a managed care organization to find the cohorts most and least likely to rate the plan negatively on the federal experience survey; members at high risk of a negative response were called by a plan representative, and those least likely to respond negatively were sent mailers. Both choices are rational for a plan paid on that survey, and neither is clinical need: outreach follows a member's likely influence on a score.

No bias evaluation, subgroup performance or governance process appears on the home, predictive analytics, terms or privacy pages, in a population that is disproportionately old, poor and multilingual. A model that learns from response rates will tend to deprioritise the members hardest to reach, who are often the least engaged in their care. The company sells plans a Measures product that breaks performance down by race, language, socioeconomic status and a health equity index, which is the tool that could answer the question for its own models.

Nothing published on what happens when the system is wrong, and the terms disclaim it. The terms of use, which govern the services under a client agreement, provide them as is, disclaim all warranties including that the services will be error free or that defects will be corrected, state that the company takes no responsibility for whether a client's messages arrive, and cap total liability at the amount paid under the agreement. No accuracy commitment, remediation obligation or route for a member to challenge a targeting decision or a conversational reply appears on the home, predictive analytics, terms or privacy pages.

The exposure sits in what the models decide. A member the targeting models do not select receives nothing, and a conversational reply that misreads a member describing a serious problem leaves no record of the miss. Neither failure is one the member can see, and the contract places neither on the vendor.

It reaches into provider scheduling as well as payer data. Published material names real time integration with the dominant record vendor's scheduling module, which is what allows an engagement campaign to book rather than merely remind, and portal capability arrived with one of the acquired businesses.

No interface standard, certification or breadth of record system coverage is described, and the primary data direction is payer side: eligibility, claims and member files rather than clinical records.

Nothing published about where the system runs or where the data rests. The home, predictive analytics, terms and privacy pages describe a hosted platform delivering across text, voice, web, portals and physical mail, and the privacy policy says data on the white labeled enrollment platform is kept within the platform's secured operational environment, but no hosting provider, region, tenancy model or residency commitment is stated anywhere.

The question has a specific edge here. Three acquired businesses since 2023 brought their own portals, analytics and communication data, so whether a plan's program runs on one platform or several joined by a common interface decides where its member data actually sits and how many environments hold it. The terms also reserve a disabling device in the software and the right to remove subscriber data after 60 days of inactivity, which bears on continuity at exit.

No price is published, and the contract structure is. The terms of use govern the services under a client agreement: non refundable one time fees and recurring fees set on the agreement's cover page, an initial term set there too, automatic renewal for one year terms unless either party gives notice, the company's right to raise fees at each renewal, suspension of service if an invoice goes unpaid for ten days, and removal of a client's subscriber data after 60 days of inactivity. So how the product is contracted is public before any sales conversation, though what it costs is not.

What drives cost is unstated and matters: billions of touchpoints a year across text, voice, portals and print imply volume economics under any platform fee, and the predictive, conversational, portal and print components arrived through different acquisitions and may be licensed separately. No fee at risk against a measure is published.

Broad across payer types and buyer categories. Published material names Medicare Advantage, Medicaid managed care and commercial plans alongside health systems, accountable care organizations, value based providers, management services organizations and pharmacies, with more than 450 organizations served.

Depth is greatest in Medicare Advantage, where the measures the platform targets carry the most revenue and where the company reports reaching 80 percent of top rated plans. Clinically it is measure driven rather than specialty driven, so coverage follows quality programs: adherence, preventive screening, chronic condition management. Everything is communication and targeting rather than care, and coverage is United States only.