A medication reconciliation platform assembles a patient’s medication history from internal, external, and patient-reported sources, detects discrepancies between those lists, and routes them for resolution at admission, transfer, and discharge. It supports the reconciling clinician. It does not resolve discrepancies, stop medications, or decide therapy.
Reconciliation is labour. External fill history arrives incomplete, patient recall is imperfect, and someone still has to sit with the patient and work through the list. The bottleneck in most hospitals is pharmacy technician hours, not software. Taction builds platforms that make those hours count for more, and we say plainly which part of the problem software cannot touch.
Our experts are ready to understand your business goals.
Reconciliation software gathers candidate medication lists, normalises them to comparable products, identifies where they disagree, and presents the differences to a clinician who resolves each one. Our AI medication reconciliation work covers the model-assisted extraction and matching side of that problem, and this page covers the platform: sources, workflow, roles, transition points, and audit. Both sit inside the wider healthcare software development practice. The platform is the part that determines whether a technician can complete a reconciliation in one sitting. Everything else in the product is downstream of whether that sitting can be completed.
Internal orders, external fill history, prior encounter lists, and patient-reported medications gathered into one view. Source provenance stays visible per entry, because a clinician resolves differently depending on origin. Retrieval timestamps are shown.
Differences in presence, dose, frequency, route, and duplication are typed and ranked. Discrepancy typing matters more than volume, since an omission and a duplicate need different resolution paths. Volume without typing is noise.
Admission, internal transfer, and discharge each require a distinct reconciliation event. Transition triggers come from admission, discharge and transfer messages so events are created automatically. Each event carries its own completion state.
Each discrepancy routes to the role authorised to resolve it, with technician, nurse, pharmacist, and prescriber steps separated. Role separation keeps prescriber time on decisions only they can make. Escalation paths are explicit.
The reconciled discharge list is produced once, in plain language, with changes from the pre-admission list marked. Change marking is the part patients and downstream clinicians actually use. A clinician reviews it before release.
It does not resolve a discrepancy, continue or stop a medication, judge adherence, or complete a reconciliation. Every resolution is a clinical action taken and recorded by a named clinician.
The work that decides success is unglamorous: product normalisation, patient matching on external queries, and a workflow a technician can run for six hours without fighting it. We build those first. Where the requirement is broader medication workflow rather than reconciliation specifically, our pharmacy application development work covers dispensing and operations, and medication management covers the general topic. Reconciliation is narrower and harder than either, because it is fundamentally about disagreement between sources. So we build normalisation and matching first, and treat the interview interface as a product in its own right.
Medication history queried from external sources with consent handling and failure paths, using our Surescripts integration work. Query failure is a designed state, not an error screen. Partial results are labelled as partial.
Products normalised to comparable concepts so a dose form difference does not read as a discrepancy, through our clinical data integration practice. Normalisation quality determines false positive volume. False positives cost user trust fastest.
Rules for presence, dose, frequency, route, therapeutic duplication, and high-alert classes configured by pharmacy. Rule tuning stays in pharmacy hands rather than requiring a release. High-alert classes are prioritised separately from routine differences.
Interfaces designed for the interview itself: keyboard-driven, resumable, and tolerant of partial information. Interview flow is where most reconciliation products lose their users. Sessions resume exactly where an interruption left them.
Reconciled discharge list with marked changes, produced for the chart and for the patient. Patient-facing output is reviewed by a clinician before it is released. Changes from the pre-admission list are marked explicitly.
Reporting on reconciliation completion by transition point, unit, and role, with a full audit trail. Completion evidence supports accreditation review and internal quality reporting. Outstanding events are reportable by unit and ageing.
We publish no figures on discrepancy rates, adverse drug events, readmissions, or reconciliation time, because those depend entirely on your population, your technician staffing, and your current practice. What we deliver is instrumentation so your team measures impact against its own data. There is a limitation specific to this category that we state early: no platform knows what a patient actually took. Fill history shows dispensing, not ingestion, and patient recall is what it is. The software organises evidence for a conversation. It does not replace the conversation. The items below reflect that boundary.
Every candidate list appears together with provenance and retrieval timestamps. Consolidated sources remove the tab-switching that makes reconciliation slow and error-prone. Retrieval failures are shown rather than hidden as an empty list.
Differences are classified and prioritised so high-alert classes surface first. Ranked discrepancies direct scarce clinician attention to the entries where error carries real harm. Routine dose form differences do not compete for attention.
Admission, transfer, and discharge reconciliations are tracked separately with their own completion state. Event separation prevents a discharge list inheriting an unreviewed admission assumption. Discharge reconciliation starts from a reviewed admission list.
Pending reconciliations, ageing, and completion by role become reportable. Workload visibility turns technician staffing into a measurable decision rather than a guess. Coverage gaps by shift and unit become an operational conversation.
Who resolved what, from which source, with what reasoning, and when. Resolution audit matters in accreditation review and in any subsequent case review. You can reconstruct the exact view a clinician had.
Software cannot confirm adherence or replace the patient interview. Fill history shows what was dispensed, and the gap between that and what was taken stays a clinical conversation. We do not claim otherwise.
We sequence around your external data access and your matching quality, because a platform that cannot reliably retrieve and match history is a data entry tool with extra steps. Discovery is paid and time-boxed and produces an itemised fixed-scope estimate with a build or configure recommendation. If your EHR’s native reconciliation module works and the real problem is technician staffing, we will say that, and no software purchase will fix it. Delivery runs in short increments with technicians and pharmacists using working software from the first increment. Technicians test every increment.
We establish which external sources you can query, under what agreements, and with what coverage for your population. Source coverage varies by payer mix and geography. Coverage is measured, not assumed from vendor claims.
Demographic matching thresholds, multiple-match handling, and no-match paths designed using our master patient index practice. Uncertain matches go to human review, never auto-merge. Thresholds are documented and remain reviewable after go-live.
Pharmacy defines what counts as a discrepancy and what is noise. Noise suppression is designed deliberately, because a platform reporting everything is ignored within a fortnight. Pharmacy owns that judgement, not engineering.
Technicians use each increment and tell us where the interview breaks. User testing with actual technicians, not managers, is a condition of our delivery approach. Manager feedback is useful but not sufficient on its own.
One unit, admission reconciliation only, running fully with reporting. Narrow pilot proves matching quality and interview flow before transfer and discharge are added. Transfer and discharge follow only once matching holds up.
Transfer and discharge events added, then handover with rule editing and matching threshold documentation. Handover leaves pharmacy able to tune rules unaided. Your pharmacy team adjusts discrepancy rules without a release.
We build reconciliation platforms as clinical support tools with strict boundaries on what the software concludes. External data retrieval respects consent and source agreements. Sensitive category data is handled architecturally rather than procedurally: we built CHIPSS, a behavioural health system, and consent segmentation of sensitive records is established practice for us rather than a design exercise. Compliance covers HIPAA safeguards, federal substance use disorder confidentiality rules where applicable, audit trails supporting case review, and clear allocation of every clinical determination to a person. Each determination below is assigned to a named role rather than implied.
The platform is clinical decision support. It does not diagnose, prescribe, stop a medication, judge adherence, or complete a reconciliation, and clinicians make every resolution and sign it. Nothing is resolved without a signature.
Whether to continue, change, or stop a medication is a prescriber determination. Whether a reconciliation is complete is a clinician’s attestation. The software records both and generates neither. Neither is inferred from the data.
Substance use disorder and behavioural health records carry re-disclosure limits. We built CHIPSS for behavioural health, and we enforce segmentation in the data model rather than in user training. Training alone is not a control.
Fill history shows dispensing, not ingestion. The interface labels source and date explicitly so source limitations are visible to the clinician resolving rather than buried in documentation. Coverage gaps are shown, not smoothed over.
Where patient instructions are model-drafted from the reconciled list, output is grounded with source traceability, reviewed by a clinician before release, and safety content such as red-flag guidance is rule-enforced.
Match decisions, thresholds, and overrides are logged alongside resolutions. Audit integrity lets you reconstruct exactly which sources a clinician saw at the time they resolved. Threshold changes are versioned like any other configuration.
We have been building healthcare software since 2013, which is over 12 years, and we have delivered more than 200 healthcare projects. We built our own EHR platform, Voyant Health, and CHIPSS, a behavioural health system where consent segmentation of sensitive records was a core requirement. We are ISO 27001 certified, our leadership brings more than 20 years of personal experience in the field, and we work from four US offices in Chicago, Cheyenne, Austin, and Sacramento. We will also tell you when your problem is staffing. That answer arrives in discovery.
We built CHIPSS, a behavioural health system, so segmenting confidential records and controlling re-disclosure is practised work rather than a policy paragraph. Segmentation extends to discharge lists and patient-facing materials.
Building Voyant Health means we understand medication lists, order structures, and how a reconciled list must appear in the chart and on discharge. Chart presentation and downstream propagation are designed together.
More than 200 healthcare projects since 2013, with matching and normalisation estimates drawn from that history rather than from vendor documentation. Coverage and false positive expectations are set from that experience.
Taction is ISO 27001 certified, with documented access control, encryption, and change management that stands up to a customer security review. External source credentials and their rotation practice are documented.
If your native module works and the constraint is technician hours, we say so. That answer loses us the project and saves you a platform that changes nothing. We put it in writing.
Four US offices in Chicago, Cheyenne, Austin, and Sacramento, with delivery overlapping your hours for the technician testing this work depends on. Escalation reaches a named delivery lead rather than a support queue.
Reconciliation platform pricing turns on how many external sources you query, the state of your product normalisation, and how many transition points are in scope. The tiers below cover engineering. Third-party licensing, cloud infrastructure, data subscriptions, and hardware are separate from engineering cost and itemised clearly. External medication history access typically carries per-transaction or subscription fees from the network and a drug terminology licence alongside it, and both are recurring vendor costs we quote as line items rather than folding into a build figure. Sensitive record segmentation is scoped and priced as its own workstream.
$40,000 to $80,000 for admission reconciliation at one site with external history retrieval, normalisation, discrepancy detection, technician workflow, and completion reporting. One site, admission only, with existing external source access.
$80,000 to $200,000 for admission, transfer, and discharge events with role-based resolution, discharge list production, patient materials, rule configuration, and audit reporting. This tier covers most single-facility programmes we are asked to scope.
Starting at $200,000 for multi-facility deployments with multi-source retrieval, consent segmentation for sensitive records, post-acute transfer handling, and multi-EHR integration. Facility count and segmentation requirements drive the figure more than features.
A paid, time-boxed discovery phase produces a source coverage assessment for your population, a matching strategy, a build or configure recommendation, and an itemised fixed-scope estimate. The assessment is yours whether or not we build.
Source count, normalisation state, transition points in scope, consent segmentation requirements, and EHR instance count. Sensitive record segmentation adds real architectural cost. Poor product normalisation in the source catalogue costs the most.
Budget annually for support, terminology updates, rule tuning, matching threshold review, and EHR upgrade regression testing. Network transaction fees are vendor costs quoted separately. Coverage changes as source participation shifts, so review is recurring.
If your technicians are reconciling from three screens and a phone call, start with the source coverage assessment. A paid discovery phase gives you a realistic view of external history coverage for your population, a matching strategy with threshold recommendations, a normalisation assessment, a build or configure recommendation, and an itemised fixed-scope estimate. If your native module is adequate and the constraint is technician hours, you keep the assessment and spend nothing further with us. Talk to our team about your transition points and your sensitive record requirements.
These are the questions pharmacy directors, informatics leads, and quality teams raise before scoping reconciliation work. Several concern boundaries we hold firmly: what the software concludes, and what remains a prescriber’s or a clinician’s call. One concerns whether you should buy software at all, where the honest answer is sometimes no. Where an answer depends on your payer mix, your external source coverage, or your EHR module, discovery resolves it quickly and the source coverage assessment is worth having regardless. The source coverage assessment is worth having even if you decide to build nothing.
No. It assembles sources, normalises products, detects and ranks discrepancies, and routes each one to the role authorised to resolve it. A clinician resolves every discrepancy and attests completion. Auto-resolution would move a prescribing judgement into a rules engine operating on incomplete external data, which is exactly the wrong place for it.
That page covers the model-assisted layer: extracting medications from unstructured text, matching products, and suggesting likely discrepancies. This page covers the platform around it: source retrieval, consent handling, transition events, role-based workflow, discharge output, and audit. Most organisations need the platform first, and the model layer is an enhancement inside it.
It is useful and incomplete. It shows what was dispensed by participating pharmacies, not what a patient took, and coverage varies with payer mix and geography. We label source and date on every entry so the clinician resolving sees exactly what they are relying on, and we assess your realistic coverage during discovery rather than after.
Architecturally. Federal confidentiality rules limit which records may be received and re-disclosed, so segmentation is enforced in the data model and in every downstream output rather than left to user training. We built CHIPSS, a behavioural health system, where that requirement drove the design, and we scope it explicitly rather than treating it as configuration.
We make no claim that it will, and we would treat any vendor figure here sceptically. Those outcomes depend on your population, your staffing, and your discharge process. What the platform provides is organised evidence, ranked discrepancies, completion visibility, and an audit trail, so your team can measure against its own baseline.
A clinician does, per transition event, and that attestation is recorded with the sources and resolutions behind it. The system tracks completion state and can report on outstanding events, but it never marks a reconciliation complete on its own, including where every discrepancy happens to have been resolved.
Your email address will not be published. Required fields are marked *
Our expert reaches out shortly after receiving your request and analyzing your requirements.
If needed, we sign an NDA to protect your privacy.
We request additional information to better understand and analyze your project.
We schedule a call to discuss your project, goals. and priorities, and provide preliminary feedback.
If you're satisfied, we finalize the agreement and start your project.