Jump

Models run every Jump product. Meet transcribes and summarizes meetings into structured notes in the adviser's own style and the firm's template, extracts tasks and financial data points, drafts recap emails and answers questions across past conversations through the AI Associate. Onboard's Document Intelligence parses statements, tax returns and estate documents, and Grow reads sentiment and recurring topics across a book of clients to raise signals.

Strip the models out and what remains is a calendar connector, scheduling with reminders, and write paths into other systems with nothing to write. Scheduling and conflict handling are conveniences around the meeting work rather than what firms buy.

Jump drafts and proposes emails, CRM updates and follow ups, then stops. Its security page says nothing reaches a client or a system of record until an adviser has reviewed and approved it, and every note, task and record update can be edited before it syncs. A firm that wants sign off before data leaves the platform can require an attestation in which the adviser confirms review of each output. Enterprise adds custom attestations, a compliance dashboard and audit trails of what happened.

Firms also choose the capture mode, from no recording to full audio and video, and set retention, consent and access policies centrally. Role based access controls and hierarchical permissions limit who sees what.

Support material describes the pipeline as speech to text transcription followed by generative summarization into notes, tasks and drafts. No transcription accuracy figure, measure of summary fidelity, benchmark or validation method is published, and nothing describes behavior on unfamiliar terminology or how often models are revalidated. Automated red team testing covers prompt injection and similar weaknesses rather than note accuracy.

A summary that leaves out a client instruction looks complete to the adviser reading it, and once approved it is synchronized into the firm's records, which a compliance reviewer will later rely on. Where a firm chooses full recording, the audio remains available to check the note against.

In the T3/Inside Information 2026 Software Survey, Jump had 22.68 percent of the AI notetaker market, against 10.22 percent for Zocks, the next largest. Jump also cites the Kitces 2026 Research Report, which put it first for customer satisfaction, and it averages 4.7 out of 5 across 105 reviews on G2. Named case studies report outcomes measured by Jump.

Athlon Advisors grew assets under advisement 43 percent after adopting Jump, and Byas Wealth Management saves 25 minutes a meeting, with its founder Jeff Byas quoted on the quality of the summaries. Across its customer base Jump reports 12 hours saved per adviser a week and 42 percent fewer open service tasks. Its chief executive, Parker Ence, says an enterprise RIA ranked Jump first among more than 40 AI vendors it evaluated.

Customer data stays out of training under the master services agreement. It says Jump does not use personal information or customer data to train its AI models, and that its third party service providers will not process customer data for training either. The security page adds that its agreements with model providers include zero retention and no training clauses, and that recordings, transcripts and summaries are never used for training or fine tuning.

Support material explains that the language models are pre trained, so customer content is not needed, and cites supervisory caution about client personal information surfacing from a model later. Firms control retention, can delete manually or automatically, and can choose no recording so that no audio file is ever created. Nonpublic and personally identifiable information can be redacted from transcripts.

Customer data is not sold, monetized or used for advertising, Jump says, and meeting recordings, transcripts and summaries are never used for training. Nonpublic personal information and personally identifiable information can be redacted from transcripts automatically. The firm controls retention, with manual and automatic deletion and custom retention periods set to its own books and records policy, and meeting data stays until the firm removes it.

Data, backups included, is stored in the region the customer designates. The master services agreement has the customer warrant that it holds the permissions needed for Jump to process client personal information. GDPR and PIPEDA compliance are stated. Jump's compliance and security pages make no reference to GLBA or Regulation S-P and give no breach notice deadline.

A SOC 2 Type II examination is in place, assessed by an independent auditor over a continuous 12 month period, and Jump publishes a trust center at security.jump.ai that it says lists current certifications and subprocessors. Controls are monitored continuously through a compliance automation platform.

Data is encrypted at rest and in transit, access requires multifactor authentication, and production code needs independent peer review and static application security testing before release. An outside firm runs annual penetration tests. The cloud setup follows a zero trust approach with continuous vulnerability scanning, threat detection, multiple availability zones and a tested continuity and recovery program, and a status page is public.

Employees pass background checks, sign confidentiality agreements, train on security awareness and work on managed devices. Automated red team testing probes for AI weaknesses such as prompt injection.

As a software supplier, Jump holds no license and names no regulator overseeing it, and its compliance features are built around customers' obligations. Capture modes, retention periods, consent protocols and recording disclosures are configured by the firm to match its supervision policy. Records are produced for RIA, independent broker dealer and hybrid firm compliance and supervision.

The master services agreement makes the customer solely responsible for verifying AI output before relying on it for decisions, compliance or communications. Nothing published describes how an examiner reviewing a firm's records would see which entries were machine generated.

Jump describes its AI governance as human in the loop, with AI proposing actions and never executing them on its own and every output reviewable before distribution. Two areas get no published evaluation. Transcription accuracy can vary with accent, dialect and speech pattern, and a systematically poorer transcript produces a poorer advice record for the clients affected.

Grow's Signals and Scorecards read sentiment, impactful topics and recurring patterns across a whole book of clients and point advisers toward particular relationships, which allocates adviser time. No testing across speaker groups, fairness review or account of how insights are ranked is published.

The master services agreement provides the service as is apart from the express warranties in section 7(a). Its AI disclaimers make the customer fully responsible for decisions based on AI output and solely responsible for verifying its accuracy, suitability and legality before relying on it. Liability is capped at the fees paid in the 12 months before a claim.

Jump indemnifies customers against third party intellectual property claims, and customers indemnify Jump for claims arising from their own conduct. Utah law governs, with disputes going first to senior management and then to mediation and arbitration. The optional attestation, in which an adviser confirms review of each output, also records who accepted each entry. No published route lets a client see or contest the notes made about them.

Jump refers to speech to text technology and large language models from third party providers without naming any of them. Its product pages, support material and file for AI assistants give no provider, model family, version or inference host. The master services agreement confirms that AI is delivered partly through third party service providers. Jump says its trust center at security.jump.ai lists current subprocessors.

Its agreements with model providers carry zero retention and no training clauses, so the terms of supply are described even though the suppliers are not named. The MCP server lets firms reach Jump data from their own AI tools, a separate path from the models Jump runs.

More than 40 systems connect to Jump in both directions, each with its own integration page. CRMs include Salesforce, Wealthbox, Redtail, Practifi, Advyzon, Microsoft Dynamics 365, HubSpot, Zoho CRM, AdvisorEngine, SmartOffice, XLR8 and Hubly. Planning, tax and estate connections cover eMoney, RightCapital, Holistiplan, Asset-Map, TaxStatus, Wealth.com and MedicarePRO, and Schwab, Orion and Black Diamond supply custodial and portfolio data.

Outputs can map to custom fields, custom objects, events and workflows rather than notes and tasks alone. Contacts are matched automatically and content is formatted to the destination's schema, with the mapping set once and reused for every later meeting. Meetings come in from Zoom, Teams, Google Meet, Webex, GoTo Meeting, RingCentral, Dialpad, Vonage and Intulse, and every plan includes MCP access, with API access on Enterprise.

The service is multitenant SaaS with strict logical separation, so one customer's data never commingles with another's. Data, including backups, is stored in the region the customer designates, and the service spans multiple availability zones. Support material places data on servers in the United States, specified to the state level. Jump runs in the browser and joins virtual, in person and phone meetings. No dedicated, single tenant or on premises option is offered, and the regions on offer are not listed.

Meet, the core plan, costs $100 per adviser a month and includes the notetaker for unlimited meetings, preparation, follow up tasks and emails, CRM and planning sync and scheduling. Onboard and Grow are add ons at $50 per adviser a month each, covering account opening and document intake, and analytics, coaching and opportunity signals. Annual billing saves up to 20 percent, and Meet can be trialed free.

Every plan includes the integrations, client profiles, the AI Associate, automations, the email assistant and MCP access. Enterprise, with API access, SSO, SCIM and SAML, custom attestations and disclosures and a compliance dashboard, is quoted. Advisers reached through aggregator platforms receive pricing negotiated by that platform.

Financial advisers at RIAs, independent broker dealers and hybrid firms are the core users, from independent practices to enterprise home offices. Jump names LPL Financial, Osaic, Cetera, Equitable Advisors, Allianz Life, John Hancock, Manulife, Principal Financial Group, Focus Financial Partners, Sanctuary Wealth and StoneX as deployments, and LPL selected it for its Vendor Affinity Program.

Separate pages address asset management wholesalers, insurance producers and carrier distribution teams, accounting professionals, and banks and credit unions. Jump says it serves more than 50,000 advisers at more than 11,500 firms managing about $12 trillion, and adds more than 2,000 advisers a month. Advisers reached through aggregator platforms receive enterprise pricing negotiated by that platform. GDPR and PIPEDA compliance are stated alongside its focus on the United States.