Sword Health
Each of these is a real decision Sword Health is working through, written so you can make the call yourself — incomplete information, someone who pushes back, no answer key.
Sword Health is a digital health company that delivers AI-enabled, clinician-supervised virtual care for musculoskeletal pain, women's health, mental health, and cardiometabolic conditions, sold primarily to employers and health plans as a covered benefit for their members. In Germany, it sells through B2B2C and B2C sales channels. Its product line includes Thrive, an AI musculoskeletal solution that delivers care for pain at home; Kaia, a digital back pain therapy product delivered as a DiGA; Pulse, an AI cardiometabolic care solution spanning hypertension, pre-diabetes, and GLP-1 support; and Triage, a certified Clinical Triage Line that guides every patient to the right care 24/7 across voice and digital channels. The Phoenix platform is the continuous, connected intelligence layer that powers these AI Care solutions. Sword Health deploys AI Care Managers trained on more than 14 million real care interactions to run the operations of providers, payers, and national health systems, and has partnered with the Hellenic Ministry of Health to build an AI-powered front door for Greece's 10 million citizens.
Your first one is free. No card.
31 live right now.
The rest
Ship a resume path for paused members without exposing stale state
Members who return after account creation are being sent back to the beginning instead of toward care. You must weigh a faster return path against reliable account and entitlement state.
Surface care gaps without turning signals into diagnoses
Population health managers cannot tell whether active chat members are quietly losing continuity of care. You must weigh earlier clinical review against the risk of treating incomplete signals as clinical conclusions.
Cluster feedback in Phoenix without letting AI set the roadmap
Recurring member concerns are being rediscovered across separate feedback reads instead of appearing as one sized signal. You must weigh faster evidence synthesis against the review needed to keep product and clinical judgment in control.
Invite at-risk members into care without implying a diagnosis
Eligible members are opening Pulse without taking a first step into care. You must weigh earlier relevance against clinical boundaries in the first conversation.
Shape value reviews before employers cut their 2027 vendor lists
Employers are questioning whether low enrollment reflects limited access or limited demand as 2027 benefit decisions approach. You must weigh a clear executive decision path against claims, outcome, and ROI evidence that matures at different speeds.
Protect access during entitlement cutover
Pre-cutover replay shows 0.8% disagreement, including active-renewal accounts. You have to weigh temporary access for uncertain records against incorrect entitlement grants that are hard to unwind.
Repair Android links into onboarding
Thirteen percent of campaign entrants reach a generic screen instead of onboarding. You have to weigh a fast routing fix against protecting every existing deep-link path.
Verify billing before the expansion cutover
Automated payment classifications disagree with human review in 1.8% of pre-cutover cases, concentrated in high-value invoices. You need to weigh launch timing against a billing error that becomes expensive to unwind.
Reconcile payments before the cash gap grows
Unmatched payments reached 3.1% of posted collections for three straight days. You need to weigh rapid correction against protecting valid payment posting.
Commit one access model across conflicting subsidiary rules
Automated matching found 4.7% overlapping member records before a plan-year configuration becomes hard to reverse. You must weigh uniform member access against verified local exceptions and trustworthy ROI reporting.
Reconcile the gap between eligibility and billing
The monthly invoice counts 8% more covered lives than the eligibility roster. You must separate a billing error from a timing issue without weakening customer trust.
Prove migrated member history still classifies correctly
Pre-cutover reconciliation shows 1.7% of historical events differ between automated mapping and human review. You need to decide what must be verified before the reporting model becomes permanent.
Repair missing hypertension readings before morning review
Twenty-two percent of yesterday’s hypertension readings are absent from the morning dataset. You need to restore the data without creating duplicate or misleading history.
Close the disputed outcomes-based contract
Finance and clinical reviewers rejected the first payment schedule because they dispute whether outcomes can be fairly attributed. You need to weigh downside protection against measures that preserve appropriate member access and withstand validation.
Recover broker referrals after first discovery
Fourteen of 19 broker-referred opportunities stopped after the first discovery call. You need to weigh a faster broker-ready story against the evidence needed to make ROI claims credible.
Prove audit evidence survives cutover
Six percent of AI-assisted classifications disagree with human review before raw audit detail is shortened permanently. You must weigh retention cost and operating speed against evidence needed for a high-severity investigation.
Repair delayed device readings
Eight percent of submitted readings are missing the five-minute delivery window after a certificate rotation. You must weigh the fastest restoration against preserving secure, observable ingestion.
Prove contact controls before the health-plan launch
Pre-launch sampling finds missing sequence evidence in 6.2% of contact events. You have to meet the launch date while proving that contact preferences reach every follow-up path.
Stop duplicate follow-ups after a rescheduled appointment
Duplicate follow-ups now affect 8.6% of rescheduled appointments. You have to protect timely outreach without sending members conflicting messages.
Lock an enrollment story for GLP-1 support
GLP-1-forward copy lifts response 42% but previously created a 9% expectation-mismatch rate. You need to weigh enrollment lift against durable trust after members enter care.
Improve mobile enrollment from benefit emails
Mobile enrollment from benefit emails has fallen to 16% after the first screen became generic. You need to weigh faster message clarity against precise benefit language.
Approve the hypertension risk-path release
Borderline hypertension cases produce inconsistent route agreement before a launch that changes member guidance and escalation. You have to weigh contracted timing against verified safety for members whose path cannot be cheaply undone.
Lift completion at the final consent screen
Completion after eligibility has fallen from 78% to 66% at the final acknowledgement screen. You have to weigh a clearer path into care against preserving an informed consent moment.
Prove the new eligibility files add up
Pre-cutover tests show 3.4% of eligibility records would be excluded or duplicated. You have to weigh the launch date against verified member access before a hard-to-reverse cutover.
Close the gap after enrollment confirmation
Twenty-two percent of confirmed enrollments wait more than 48 hours for outreach. You have to weigh a fast handoff against keeping eligibility checks dependable.
Verify multilingual routing before the care-path launch
Routing accuracy ranges from 89% to 98% by language in pre-launch tests, with failures around escalation requests. You must weigh launch readiness against the cost of delaying a contracted care path and the risk of a wrong handoff.
Correct stale guidance in follow-up messages
Clinician review found outdated GLP-1 guidance in 9% of sampled follow-up messages. You must weigh a fast retrieval correction against response speed and reliable regression coverage.
Decide what counts as improvement at 90 days
The health plan launch date is fixed, but member reports, clinician notes, and session data disagree 22% of the time on 90-day work function. You have to set the measurement rules now, knowing contracts will lock them in.
Decide on the coalition deal before the vote
The coalition wants a fixed price per eligible member, category exclusivity, and a shared feed for three years, and the board votes in 14 days. You have to land terms you can live with or recommend walking away.
Stop reminders going to terminated members
An audit found 23 messages that reached terminated employees, and the customer wants a cure within 30 days. The fix affects every member and cannot be undone once you attest, so you settle what evidence counts as closure.