A TekStack perspective on AI, customer context and the complete service lifecycle

HubSpot is pointing toward the right future: AI agents need governed access to customer data, context and business capabilities wherever people work. The harder question is whether HubSpot contains enough of the customer lifecycle to serve as the harness for service-driven businesses.

The customer harness is a compelling idea

Recently, HubSpot repositioned itself around an interesting idea: the Customer Relationship Harness.

The premise is compelling. AI already knows a great deal about the world. What it lacks is context about your customers, your commercial relationships and the work your company is doing for them. HubSpot wants to become the governed layer connecting that customer context to AI, allowing agents to work with it from HubSpot, ChatGPT, Claude and elsewhere.

HubSpot is right about the direction. Salesforce is talking about much the same thing with Headless 360. The application itself matters less when AI becomes another way of interacting with the business.

Where the real question starts is underneath all of that: what data and operational context can the AI actually get to? An AI harness is only as useful as the customer relationship behind it.

At TekStack, we agree. But the Customer Harness idea raises a more fundamental question: what exactly is the harness connected to? An AI harness is only as intelligent as the customer and operational context behind it.

HubSpot has evolved considerably

It would be inaccurate to characterize today’s HubSpot as simply a marketing automation system with a CRM bolted on.

The product now stretches across Marketing Hub, Sales Hub, Service Hub, Content Hub, Data Hub, Revenue Hub, Smart CRM and Agent Hub. HubSpot also has a large integration ecosystem.

Its APIs shouldn’t be dismissed either. HubSpot offers REST APIs, OAuth, webhooks, custom objects and APIs spanning contacts, companies, deals, tickets, invoices, subscriptions and many other objects.

The argument is not that HubSpot has no APIs. But is connecting to a proprietary SaaS product the same as owning an extensible customer operating platform?

Even more importantly: does the underlying platform contain enough of the customer lifecycle for AI to understand the business? For many service-driven companies, we believe the answer is no.

Customer 360 needs to mean more than marketing, sales and support

For a straightforward SaaS company, a good portion of the HubSpot customer base, the customer journey might look roughly like this:

Market → Sell → Support → Renew

 

HubSpot covers that model increasingly well. But consider a 100-person IT company, MSP, VAR or professional services firm. Its lifecycle looks more like:

Market → Sell → Scope → Contract → Onboard → Resource → Deliver → Track Time → Support → Bill → Renew → Expand

 

Now imagine asking an AI agent, “Which customers are at risk?” A marketing and CRM platform may know email engagement, meetings, opportunities, support tickets, health scores and website activity. For a services company, however, the answer may depend on:

  • Project budget consumed
  • Milestones missed or project slippage
  • Resource availability and capacity
  • Billable hours and time write-offs
  • Support SLA performance
  • Project margin and unbilled work
  • Agreements, entitlements, invoices and renewal dates

If you run a services business, a customer relationship doesn’t stop when the deal closes. In a lot of ways, that’s where it starts.

You must scope the work, staff it, deliver it, support it, track the time, manage the margin, invoice it and hopefully expand the relationship. That operational history is customer data too.

Where HubSpot is native, and where another stack begins

HubSpot is broad, but it does not attempt to replace a full professional-services automation platform. Its own marketplace includes products such as PSOhub and BigTime to add project management, resource planning, time and expense, invoicing, utilization and profitability capabilities.

The central difference is architectural. With HubSpot, customer engagement lives in HubSpot while deeper delivery operations often live in a PSA and financial operations live in an accounting or ERP platform. A completely different tool, different database, different licensing arrangement, different application support.

That changes the customer harness conversation.

If the information AI needs is spread across CRM, PSA and financial systems, then the real challenge is connecting the commercial and operational context of the customer across the stack.

This is something we’ve been thinking about at TekStack for years.

TekStack was built around a different operating model: sales, projects, service, billing and recurring revenue share Microsoft Dataverse as a common customer foundation.

  • Lifecycle depth: CRM, onboarding, project delivery, support, billing and renewals share one operating model.
  • Professional-services operations: projects, resources, time, utilization, budget and profitability are native rather than attached through a separate PSA.
  • Microsoft-native architecture: Teams, Outlook, SharePoint, Power BI, Power Automate, Copilot and Dataverse are part of the surrounding platform, not merely integrations.
  • Customer control: the solution operates in the customer’s Microsoft environment and can be extended with Dataverse tables, Power Apps, workflows and APIs.
  • Agent context: AI can reason over commercial, delivery, support and financial signals from the same governed customer model.
  • Consolidated economics: the value proposition is not a separate Hub for every department, although Microsoft licensing and capacity still need to be considered.
  • Built in application support, managed services, and implementation: All included in annual pricing, and no need to contract third party system integrators or partners.

What each approach covers

Business capability HubSpot TekStack
CRM and customer 360 Native Native
Marketing automation Excellent / native Sufficiently Great / native
Sales pipeline and sequences Native Native
Help desk, cases and knowledge Native Native
Customer success Native Native
AI agents Native HubSpot agents Microsoft Copilot + TekStack context
Project management Lightweight / add PSA for depth Native PSA
Project financials Add PSA Native
Resource and capacity planning Add PSA Native
Time, expense and utilization Add PSA Native
Project profitability Add PSA Native
Agreements and recurring services Partial / configurable Native
Billing and services invoicing Revenue tools; add systems for depth Native services workflow
Microsoft 365 and Dataverse Integrations Underlying platform

 

A pricing example: a 100-person services company

Consider a company with 5 sellers and marketers, 5 managers, 20 support professionals, and 70 project managers and delivery resources. The following is an illustrative front-office scenario using publicly posted annual list pricing. It is not a quote, and bundles, contact volumes, discounts, credits and negotiated terms may materially change the total.

Component Illustrative assumption Monthly
Marketing Hub Professional Base subscription; 3 Core Seats; 2,000 marketing contacts $800
Sales Hub Professional 5 seats × $90 $450
Service Hub Professional 20 seats × $90 $1,800
Additional Professional Core Seats 5 seats × approximately $50 $250
HubSpot front-office subtotal   $3,300
Approximate annual subscription   $39,600

Public Professional onboarding fees can also apply. Current posted amounts include $3,000 for Marketing Hub Professional and $1,500 each for Sales Hub Professional and Service Hub Professional, although commercial treatment should be confirmed with HubSpot.

The approximately $39,600 annual figure is a front-office subtotal. It still does not equip the 70 people responsible for delivering customer work.

 

Those delivery professionals still need project management, resource planning, time and expense, utilization, project budgets, profitability and invoicing. In a HubSpot architecture, that normally means adding a PSA such as PSOhub, BigTime, Accelo, Productive or Scoro, plus integration and potentially an accounting or ERP platform.

A responsible comparison adds four things to the HubSpot subtotal:

  • PSA subscriptions for delivery and operational users
  • Implementation and data migration for both systems
  • Ongoing CRM-to-PSA-to-finance integration
  • AI credits, contact tiers or API capacity where usage requires them

Because third-party PSA pricing varies by vendor, edition, contract and which of the 70 delivery users need full versus time-entry access, the delivery layer is intentionally not estimated here. The important point is that it is additional. The front-office subtotal is not a like-for-like total cost of ownership comparison against a platform that includes PSA and project financials.

The cost of AI adds another dimension

HubSpot increasingly prices agentic functions through HubSpot Credits. Its public catalog lists additional credits at $10 per 1,000, while different actions consume different quantities. HubSpot also commercializes additional API capacity.

There is nothing inherently wrong with cloud limits or consumption pricing. Microsoft platforms also have licensing, capacity and service-protection limits. The strategic question is whether an AI harness should add another collection of product, seat, contact, API and consumption meters to the customer stack.

Where HubSpot remains stronger

A credible comparison should also be clear about HubSpot’s advantages:

  • More mature marketing automation and demand-generation capabilities
  • Strong brand recognition and broad adoption

For a smaller SaaS company whose lifecycle is primarily generate demand, sell a subscription, provide support and renew it, HubSpot could be the right answer. The calculation changes for MSPs, systems integrators, VARs, consulting firms and other service-driven businesses because those companies do not merely engage customers. They perform work for them.

The real battle is no longer CRM

The next platform battle is more about context than CRM. AI needs to know:

  • What you sold the customer, and what’s been delivered.
  • Was the project profitable?
  • Is support meeting its commitments.
  • What they’ve been billed.
  • Is anything coming up for renewal?
  • What needs to happen next.

If we’re going to give AI meaningful responsibility for managing the customer relationship, then context is the harness it needs.

 

About TekStack

TekStack is a Microsoft-native platform built for service-driven businesses that want their customer and operational data in one place. CRM, project delivery, service, billing and recurring revenue share the same Microsoft Dataverse foundation, giving people and AI a more complete view of the customer relationship.

Learn more about how TekStack connects the customer lifecycle at tekstack.com.

 

Pricing assumptions and sources

Pricing shown in U.S. dollars and based on publicly posted annual pricing reviewed in November 2026. Verify current pricing and eligibility directly with each vendor before making a purchasing decision.

  • HubSpot, Marketing Software Pricing: Marketing Hub Professional starts at $800/month annually, including 3 Core Seats and 2,000 marketing contacts; additional Core Seats start at $45/month; $3,000 onboarding is listed.
  • HubSpot, Sales Software Pricing: Sales Hub Professional starts at $90/seat/month annually; $1,500 onboarding is listed.
  • HubSpot, Service Software Pricing: Service Hub Professional starts at $90/seat/month annually; $1,500 onboarding is listed.
  • HubSpot App Marketplace, PSOhub: separate subscription adding projects, resources, time, expense, invoicing and profitability.
  • HubSpot App Marketplace, BigTime: separate PSA subscription connecting HubSpot deals with project delivery.
  • TekStack internal platform materials: CRM, PSA, service, billing and renewals on Microsoft Power Platform and Dataverse.