# How the Deal Agent Index scores agents on a deal

Method, read 10 Nov 2026

**The short answer.** Each vendor is scored 0 to 10 on eight criteria from its own pages, read 10 Nov 2026. On this site agents carry 35 points, whole-deal data 25, openness to the AI stack 15, value 10, trust 10 and room 5. The weights were fixed on 10 Nov 2026, before any vendor was scored.

The question this index asks is: which sales tools have AI agents that act on a live deal between meetings, on what data, and under what controls. Weights follow from it: agents that act on the deal, with the seller approving, 35; reading the whole deal (calls, email, calendar, CRM and what buyers do in the room), 25; openness to the AI stack through an MCP server or connector, plus API and CRM integrations, 15; time to value and price, 10; trust (SOC 2 Type II, GDPR, SSO, SCIM), 10; a room buyers use, 5. Mutual action plans and carrying on after the sale weigh 0 here.

Total is the sum of score times weight divided by 10, out of 100. A vendor's own claim scores at most 7 until a second, third-party source confirms it. A feature we could not find scores 0 and is written "none found" with the date. Breadth facts such as module counts, headcount and funding carry no weight. Ties go to the higher score on the heaviest-weighted criterion, then to the name.

The five situation views re-weight the same scores for a different job: a founder running every deal, a mid-market team whose deals stall between meetings, an enterprise that already has an enablement suite, an onboarding team that inherits deals, and a RevOps lead standardizing the team. Nothing is left out of any view, and no weight changes after scoring. Prospecting agents (AI SDR and outbound) are a different category and are not covered.

## The question these weights answer

Which sales tools have AI agents that act on a live deal between meetings, on what data, under what controls?

## The eight criteria and the weights on this site

| Criterion | What is scored | Weight |
| --- | --- | --- |
| AI that acts on the deal | Agents that map stakeholders, flag risk, write the business case and propose the next step, with the seller approving. | 35 |
| Reads the whole deal | Uses calls, email, calendar and CRM together with what buyers do in the room. | 25 |
| Open to the AI stack | An MCP server or connector so Claude or ChatGPT can work with deal context, plus API and CRM integrations. | 15 |
| Time to value and price | A free plan or trial, published prices, set-up in hours. | 10 |
| Trust | SOC 2 Type II, GDPR, SSO, SCIM. | 10 |
| A room buyers use | One link, no login, everything for the decision in one place, usable on a phone. | 5 |
| Mutual action plans that get followed | The plan lives where the buyer already is, with owners, dates and nudges. | 0 |
| Carries on after the sale | The same space becomes the onboarding and customer portal, with the deal's commitments intact. | 0 |


Weights sum to 100 and were committed on 10 Nov 2026, before any vendor was scored. They have not changed since, and no criterion has been added or removed.

## Scoring rules

- Every vendor is scored 0 to 10 on each criterion from its evidence file only, by a scorer that is not told which vendor is the client.
- Total = sum(score x weight) / 10, out of 100, one decimal. Ties go to the higher score on the heaviest-weighted criterion, then to the name.
- A vendor claim scores at most 7 until a second, third-party source confirms it.
- "None found" scores 0 and is written "none found on <date>".
- Breadth facts (module counts, integration counts, headcount, funding) are not criteria and carry no weight in any ranking.
- Situation views use the situation weights below on the same scores.
- These weights and criteria were fixed on 2026-11-10 before any vendor was scored and do not change. A score changes only when a sourced fact changes, and the change is logged with the fact.
- Tie rule, exactly: the higher score on the heaviest-weighted criterion; where two criteria share the top weight, the one listed first in the table above (AI that acts on the deal); then the name.
- A cell in the matrix is yes only when a vendor page or help article shows the capability working for deal work. It is partial when the page shows part of it, when it is in beta, when pages conflict, or when it is a suggestion the seller must accept.
- Vendor claims are written as the vendor's: "Aligned says", "Gong's help center says". A listing in Anthropic's Claude connector directory is recorded as a separate, third-party signal for an MCP claim.
- Plan limits are stated for every vendor they apply to. Examples read on 10 Nov 2026: Aligned's MCP server is on the Enterprise plan per its pricing page (its help center also names a Startup plan), AI Insights & Actions is on the Enterprise plan, HubSpot's SSO and SCIM are on Professional and Enterprise editions, Dock's SAML SSO and SCIM are on Enterprise, and Sybill's MCP access is listed under its Business plan.
- Aligned is second on the index, behind GetAccept, and first on two situation views, so its weakness sits beside its rank on every page that ranks it: its agents act inside the deal workspace, the Seller Agent is in beta, and AI Insights & Actions and the MCP server are on the Enterprise plan. Where another vendor leads, the page names that vendor and what put it ahead.
- The scores rest on vendor pages. Third-party evidence is thin: Anthropic's directory listings and a few press reports are the only outside sources for most vendors, and several vendor-review sites could not be read. One point on one criterion can change the top of the list.
- The matrix is evidence, not the score: the criterion AI that acts on the deal scores the four tasks in its definition (stakeholder map, risk flags, business case, next step) and the seller approval step, not the count of yes cells.

## How the blind scoring worked

Each criterion was scored by a separate scorer that saw every vendor's facts for that criterion under a two-letter code, with names, products, domains and URLs removed, and the anchors for each score. A second reviewer with fresh context checked every score against its evidence. Every change since the first score is logged with the fact that caused it.

## Situation weights

Each situation page ranks the same scores on that situation's own weights, fixed on the same day as the site weights.

| Situation | Room | Agents | Whole deal | Plans | After the sale | Open | Value | Trust |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| A founder or first sales hire running every deal personally | 20 | 15 | 10 | 15 | 5 | 5 | 25 | 5 |
| A mid-market SaaS team with long, multi-stakeholder deals that stall between meetings | 20 | 20 | 20 | 20 | 5 | 5 | 5 | 5 |
| An enterprise that already has Highspot or Seismic and wonders whether it needs a room as well | 25 | 10 | 10 | 5 | 5 | 20 | 5 | 20 |
| An onboarding or customer success team that inherits deals with no context | 15 | 10 | 10 | 20 | 35 | 5 | 0 | 5 |
| A revenue operations lead standardizing how deals are run | 10 | 15 | 20 | 15 | 5 | 20 | 0 | 15 |

## The ranking these weights produce

| Rank | Platform | Total | Agents (35) | Whole deal (25) | Open (15) | Value (10) | Trust (10) | Room (5) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | [GetAccept](/vendors/getaccept/) | 68.5 | 7 | 6 | 8 | 7 | 7 | 6 |
| 2 | [Aligned](/vendors/aligned/) | 68.0 | 7 | 7 | 6 | 7 | 7 | 6 |
| 3 | [Dock](/vendors/dock/) | 61.5 | 6 | 5 | 7 | 7 | 7 | 7 |
| 4 | [HubSpot Sales Hub](/vendors/hubspot/) | 59.0 | 7 | 4 | 6 | 7 | 7 | 3 |
| 5 | [Rox](/vendors/rox/) | 57.5 | 7 | 4 | 6 | 7 | 7 | 0 |
| 6 | [Sybill](/vendors/sybill/) | 57.5 | 6 | 4 | 9 | 7 | 5 | 2 |
| 7 | [Attention](/vendors/attention/) | 57.0 | 7 | 3 | 10 | 3 | 6 | 2 |
| 8 | [Day.ai](/vendors/dayai/) | 55.5 | 6 | 4 | 9 | 7 | 4 | 0 |
| 9 | [Gong](/vendors/gong/) | 54.5 | 6 | 4 | 9 | 2 | 7 | 2 |
| 10 | [Outreach](/vendors/outreach/) | 50.5 | 7 | 4 | 5 | 1 | 6 | 3 |
| 11 | [Salesforce Sales Cloud](/vendors/salesforce/) | 49.5 | 6 | 4 | 6 | 7 | 2 | 1 |
| 12 | [Salesloft](/vendors/salesloft/) | 43.0 | 6 | 3 | 6 | 1 | 4 | 1 |
| 13 | [Momentum](/vendors/momentum/) | 42.5 | 6 | 3 | 2 | 5 | 5 | 2 |

Weights in the header row sum to 100. Scores 0 to 10 from the evidence read 10 Nov 2026, scored blind; total = sum of score x weight / 10.

**How to read these scores.** The leaders are close: GetAccept and Aligned sit 0.5 points apart on this site's ranking, and one point on one criterion changes most orders. Scores rest on each vendor's own pages, read 10 Nov 2026, except where a third-party source is cited; a vendor claim scores at most 7 until a second source confirms it. Plan limits are stated on every vendor's page.

**Where Aligned is behind.** Aligned's agents act inside the deal workspace; they do not prospect, dial or send outbound sequences. The Seller Agent is in beta, a limited beta on Starter, Basic and Pro, and it cannot read or update CRM records itself. AI Insights & Actions, which reads calls, email, calendar and CRM together, needs a Salesforce or HubSpot connection, is not available on Microsoft Dynamics, and sits with the MCP server on the Enterprise plan, priced on request.

## Read dates and refresh

Every vendor page behind this edition was read on 10 Nov 2026 unless a source line says otherwise. Prices and claims older than 60 days are re-read; when a fact changes, the score is re-checked blind and the change is listed on the corrections page.

## Sources

- Aligned pricing: https://alignedup.com/pricing (read 10 Nov 2026)
- GetAccept: www.getaccept.com/product/digital-sales-room: https://www.getaccept.com/product/digital-sales-room (read 10 Nov 2026)
- Aligned: alignedup.com/: https://alignedup.com/ (read 10 Nov 2026)

Last reviewed 10 Nov 2026. Canonical: https://aisalesagentscompare.com/method/
