Tower CRM

Model features assist the work rather than constitute it, and the meters confirm that reading.

What exists: call transcription with generated summary notes written back to the contact record, email drafting, and a close probability prediction computed from historical patterns. The vendor states that intelligence is built into how a customer already works rather than presented as a separate product, and every plan including the free one carries some form of it.

What a customer pays for is different. Billing is per workspace with a seat allowance, and the tiers step on contact limits, user counts, integrations and reporting depth. Not one unit of account measures inference, and the model features are listed alongside conventional capabilities rather than above them.

The transcription and summary combination is the one doing real work, because an agent finishing a call has a written record in the contact timeline without typing it, which is the point at which record systems usually fail at this size.

Strip the model layer out and the industry configuration, pipelines, automations and marketing all still function.

Ask what the close probability is calculated from and how it is presented.

One automation escalates to a human when it has not produced an outcome, which is the right design.

That is the renewal sequence in insurance mode. It tracks the expiry date for every client, sends reminders at set intervals ahead of it, and then creates a task for the agent if the policy still has not been addressed. Most automation stops at sending. This one notices that sending did not work and hands the problem to a person, which is the difference between a reminder system and a control.

Around it: automations described as running follow up, email campaigns with enrolment triggers and sequences, and reporting with pipeline visibility at the upper paid tier.

Unlimited automations are included on every plan including the free one, which is unusual and means the constraint on what a customer builds is their own design rather than their tier.

What two passes could not locate: any audit trail, any approval workflow, any change history and any permission or role model below the top tier, where single sign on first appears. For a shared agency workspace holding client policy data, role separation would ordinarily arrive earlier than that.

Ask what permissions exist below the enterprise tier.

Four model capabilities and no technology named behind any of them.

Two passes across the home page, pricing page, feature listings and independent coverage located no provider, no model name, no version, no statement of what data is transmitted for inference, no training or retention position and no accuracy figure for transcription, summarisation, email drafting or close probability prediction.

Two of those deserve specific attention. Call transcription processes recorded conversations between an agent and a client, which for an insurance agency means discussions of coverage, claims and personal circumstances, and nothing describes where that audio goes. And close probability prediction feeds forecasting, so a number with no stated basis and no error rate ends up informing what an agency expects to earn.

The vendor publishes a machine readable file and a dedicated guide for using the product alongside a named commercial assistant, so it has evidently thought about how models interact with its material. It has published nothing about the models inside its own product.

Ask which provider processes call audio and what the close probability is trained on.

Founded this year, and the evidence base is what that implies.

Two passes across the home page, pricing, footer and independent listings located no named customer, no logo, no case study, no testimonial, no review presence on any platform, no customer count and no usage figure of any kind.

The one credential offered is the founder's own record: that they have deployed record and revenue operations systems at hundreds of business to business companies, and that every feature comes from that experience rather than a product roadmap meeting. A manifesto page and an about page carry the argument further. That is operator experience rather than customer evidence, and the vendor does not pretend otherwise.

What it does do is state the stage honestly. Early customers are told they get direct founder access, optional hands on onboarding and a real voice in what gets built next, and founding customers lock in current pricing. Those are the terms of an early access arrangement described plainly rather than dressed up.

A buyer should read the grade as reflecting company age rather than concealment. There is nothing here yet to verify.

Ask how many workspaces are live and whether any will take a reference call.

Email campaigns, sequences and messaging automation, with no recipient controls described anywhere.

Two passes across the pricing page, feature listings, industry pages and independent coverage located no unsubscribe mechanism, no suppression list, no do not contact handling, no consent basis, no reference to the carrier registration framework governing business text messaging, and no do not call screening despite two named dialer integrations at the upper tier.

The free tier caps sending at one hundred messages monthly, which limits exposure at entry, and no cap is stated on paid tiers.

The insurance context makes some of this sharper than it would be elsewhere. Renewal automation sends reminders at fixed intervals to existing policyholders, which is service communication to people with a live relationship and a legitimate reason to be contacted. That is the easier case. Cross sell pipelines and campaign sequences to the wider book are marketing, and those carry the obligations that nothing here addresses.

The real estate mode adds text message automation with the same gap.

Ask how unsubscribe is handled in campaigns and what applies to text messaging.

Two documents published, against holdings that warrant considerably more.

The footer carries terms and a privacy policy. Two passes located no processing agreement, no subprocessor list, no retention schedule, no deletion commitment, no residency statement, no rights request route and no named privacy contact.

What sits in the platform is the concern. In insurance mode this holds policy fields, renewal dates, multi policy household structures and, at the upper tier, live policy and account data synchronised from an agency management system. That is a client's insurance position: what they own, what it is worth, when it renews and who else in the household is covered. Call transcripts are added on top.

Insurance agencies handle information their own regulators treat as sensitive, and an agency adopting this becomes responsible for where that data sits. A processing agreement and a subprocessor list are the documents that question is answered with, and neither exists.

Advanced security appears only at the enterprise tier, which implies it is not the baseline.

Ask for a processing agreement and where policy data is stored.

No data about people is supplied with the product, and the intake routes are all first party.

Two passes located no contact database, no record count, no enrichment engine, no lead finder, no intent signals and no third party data supplier anywhere in the feature set or pricing. The platform holds relationships the customer already has.

Three intake paths are described and each originates with the customer. Import from spreadsheets or another record system, with a step by step guide inside the application. Synchronisation from the customer's own agency management system, bringing across clients, policies and renewal dates. And direct entry by the agent.

The agency management system route is the interesting one, because for an insurance agency that system is the authoritative record and the data in it came from the agency's own client relationships. Nothing is acquired from outside.

That removes the provenance question entirely: nobody appears in this system because they were bought, scraped or inferred.

Ask whether synchronisation runs in both directions or only inbound.

Every named integration is a partner path and nothing operates without permission.

The named set is specific: an insurance agency management system for live policy and account synchronisation, two dialer platforms for call transcription, mail and calendar services, a team messaging platform, and a general automation service reaching over two thousand applications. Each publishes an interface for exactly this use.

Naming an agency management system by name matters more than the count. Those systems are the authoritative record inside an insurance agency and they do not connect casually; a working synchronisation implies an actual relationship rather than a generic connector pointed at an interface.

Two passes located no scraping, no social network automation, no browser extension operating a third party property, no account renting and no capability whose viability depends on another platform failing to notice it.

That position follows from what the product is. A record system holding an agency's own book of business has no reason to reach into anybody else's network.

Custom integrations are reserved for the enterprise tier.

Ask which other agency management systems are supported.

Call audio and policy data flow through a platform with no published security position.

On the model side, two passes located no provider, no evaluation, no accuracy figure and no constraint description for transcription, summarisation, drafting or close probability prediction. Transcription in particular processes recorded conversations between an agent and a client about coverage and circumstances, and where that audio is sent is unstated.

On the security side the record is empty. Two passes located no security page, no certification, no audit, no encryption statement, no access control description, no incident response process, no breach notification commitment, no named security contact, no vulnerability disclosure route and no status page. Advanced security and single sign on appear as enterprise tier features, which indicates they are absent below.

The combination is what makes this consequential rather than routine. At the upper paid tier the platform synchronises live policy and account data from an agency management system, so a client's full insurance position sits alongside transcripts of their calls, in a product with two published legal documents and nothing describing how any of it is protected.

Ask what encryption and access controls apply below the enterprise tier.

Nothing manufactures a false impression, and one automated capability records people without a described notice.

Two passes located no persona, no synthetic sender, no cloned voice, no rented identity, no generated message presented as personally written and no automated social action. Communication runs from the agent's own address and their own number, mostly to clients who already hold policies with them.

The intake being entirely first party helps here too: people in this system are the agency's own clients and prospects rather than strangers sourced from a database.

The gap is call recording. Transcription with generated summary notes operates on conversations between an agent and a client, and nothing describes what the person on the other end is told before it begins. That matters commercially as well as legally in a country where consent requirements differ by state and where the conversations concern personal circumstances.

Two further gaps sit on the outbound side: email campaigns run with no unsubscribe mechanism described, and the real estate mode adds text message automation with the same absence.

Ask what a client hears before a call is recorded and transcribed.

Unlimited integrations on every plan including the free one, and the named ones are chosen for the vertical.

Including integrations without limit at the free tier is the structural decision worth recording, since gating connectivity is the standard way to force upgrades and this vendor declines to.

The named set is specific rather than generic: an insurance agency management system by name for live policy and account synchronisation, two named dialer platforms feeding call transcription, mail and calendar services, a team messaging platform, and a general automation service reaching over two thousand applications.

The agency management system connection is the one that determines whether this product works for its stated buyer. An insurance agency's authoritative record lives in that system, and a record system that cannot reach it becomes a second place to type things.

Two further surfaces are published for machine consumption: a structured file for language models and a dedicated guide for using the product alongside a named commercial assistant.

What two passes could not locate: any public interface documentation, any authentication scheme, any webhooks or any protocol server. Custom integrations are reserved for the enterprise tier.

Ask whether a documented interface exists below the enterprise tier.

Nothing about the infrastructure appears on any surface examined.

Two passes across the home page, pricing page, feature listings, industry pages and footer located no hosting provider, no region, no data centre, no residency commitment, no tenancy or isolation model, no encryption statement, no backup position, no continuity plan and no status page. Service level guarantees are listed as an enterprise tier feature, which means no availability commitment applies to anyone below it.

The holdings make residency a real question rather than a formality. In insurance mode the platform holds policy details, renewal dates and household structures, synchronised live from an agency management system at the upper tier, alongside call recordings and their transcripts.

Insurance regulation in the United States operates at state level, and agencies carry their own obligations about where client information is held and who may reach it. An agency principal would need the processing location established before connecting their management system.

Company age is the mitigating context. A platform launched this year may not yet have infrastructure documentation to publish.

Ask where data is hosted and what uptime applies below the enterprise tier.

Two legal documents, and security positioned as a feature of the most expensive tier.

The footer carries terms and a privacy policy. Two passes located no certification of any kind, no audit report, no examination, no penetration test, no trust portal, no dedicated security page, no encryption or access control statement, no subprocessor list, no incident response process, no named security contact, no vulnerability disclosure route and no status page.

The more telling detail is where security appears in the commercial structure. Advanced security and single sign on are listed among the enterprise tier features alongside a dedicated account manager and service level guarantees. Treating security as something bought at the top rather than provided as a floor is a defensible early stage choice and a buyer should understand it as such.

What sharpens it is the data. At the upper paid tier this platform synchronises live policy and account information from an agency management system, meaning a client's insurance position, alongside recordings and transcripts of their calls.

An agency principal answering to a state regulator has nothing here to file.

Ask what security controls apply on the standard tiers and whether any assessment is planned.

Four tiers published in full, priced per workspace rather than per seat, with one stale figure on the page.

The structure is clear and complete. A free plan at no cost covering one user, a thousand contacts and a hundred messages monthly. A basic plan at eighty nine dollars for up to two users with unlimited contacts, campaigns, automations and custom fields. A professional plan at two hundred and ninety nine dollars including five users, adding agency management system synchronisation, dialer transcription with generated notes, enrolment triggered sequences and advanced reporting. An enterprise tier priced on application with unlimited users, dedicated management, custom integrations, advanced security and service guarantees. Annual billing saves up to twenty percent.

Charging per workspace with a seat allowance rather than per user is the decision that suits the buyer, since a small agency adding a producer does not immediately incur a new licence.

Two commitments go further than most. Unlimited automations and integrations are included on every plan including the free one. And founding customers are told they lock in the pricing currently displayed.

The discrepancy: the page's own metadata describes the basic plan at fifty nine dollars while the page body shows eighty nine. A thirty dollar gap on a single page, almost certainly a stale tag from a recent increase, and it is the version a search engine or assistant would surface.

Ask which basic price is current and what founding pricing locks.

Getting in is documented carefully and getting out is not described at all.

The entry path is explicit: contacts, companies and deals can be imported from spreadsheets or another record system, with a step by step guide available inside the application, and a dedicated page addresses switching from a competitor. That is a vendor that has thought about migration, in one direction.

Outbound routes exist by implication rather than statement. Unlimited integrations on every plan and a connection to a general automation service reaching over two thousand applications mean data can be pushed outward continuously. Where an agency runs an agency management system, the authoritative client and policy record already lives there rather than here, which is the strongest portability position in this record and it comes from the architecture rather than a feature.

What two passes could not locate: any export mechanism, any format, any interface documentation, any deletion commitment, any retention position after cancellation and any account closure process.

The material with no described route out is what the platform adds rather than receives: call transcripts, generated summary notes, campaign history and automation configuration.

Ask whether call transcripts and campaign history can be exported.

Campaigns and sequences run with no sending discipline described anywhere.

The platform sends email campaigns with enrolment triggers and sequences, and the real estate mode adds text message automation. Two passes across the pricing page, feature listings, industry pages and independent coverage located no warm up capability, no sender authentication guidance, no domain or inbox rotation, no address verification, no bounce handling, no placement testing, no reputation monitoring and no sending limits on paid tiers.

Web analytics and attribution are included, so the vendor measures what happens after a message lands without describing how it gets there.

Two factors reduce the exposure relative to a cold outbound platform. The audience is the agency's own book of business rather than purchased lists, so recipients recognise the sender and complaint rates are correspondingly lower. And the free tier caps sending at a hundred messages monthly, which bounds the entry level risk.

The renewal reminder traffic is the highest volume use and the least likely to cause difficulty, since it goes to existing policyholders about their own cover.

Ask what sending infrastructure campaigns use and what limits apply on paid tiers.

The industry modes are data model differences rather than template swaps, which is why the approach works.

Insurance mode arrives with policy fields, expiry date tracking, renewal automations, cross sell pipelines and multi policy household structures. Real estate mode arrives with listing stages, buyer pipelines, commission tracking and showing follow ups. Those are different objects and different relationships, not the same system with new labels, and they address the actual reason vertical teams abandon general record systems: the fields they need do not exist and the ones that do go unused.

A multi policy household is the clearest example, because modelling a family with several policies is not something a generic contact and company structure handles at all.

A third mode serves any business team with a clean flexible configuration, so the product is not exclusively vertical.

A comparison hub measures the product against three named competitors and, candidly, against the spreadsheet the vendor assumes a prospect is currently using.

The positioning is visibly in transition. Page titles and descriptions now lead with insurance while the footer still describes a general product for high growth teams, which suggests a recent narrowing.

Ask which further industry modes are planned.