fidelity.com

scanned 9/10/2026, 9:49:23 PM · took 173s · SecuritiesGemini · gemini-2.5-flash · HTTP · Target fidelity.comA measurement at one moment, not a verdict · Methodology

67/ 100

CFair

Fidelity.com provides authoritative answers but forces AI agents to rely on external search engines due to a nearly invisible internal link structure and missing discovery files.

The one fix worth most right nowGet the fix files →

Put the main links in the raw HTML+1367 → 80 · grade C → B

All of them together: 93 (A)

Projected from the audit record. After fixing, re-run the task or re-measure to confirm.

AI agents achieved a 60% task success rate on Fidelity.com, consistently citing the site's own URLs when they could find the right pages. However, the site's internal navigation is highly opaque to automated systems, evidenced by a severely low link following score of 3% and only 6 raw HTML anchors detected on the homepage. Because standard discovery tools like a sitemap.xml (returns 404) and llms.txt (returns 403 Forbidden) are unavailable, agents cannot map the site natively. Consequently, agents had to use heavy bridging—relying on external search engines—to complete tasks related to eligibility, contact details, and recent updates.

What works

Agents consistently cited Fidelity's own URLs (100% own-source rate), ensuring users receive authoritative information.
Tasks related to fees, support, and product comparisons succeeded natively without requiring external search workarounds.
The site demonstrates perfect reliability (100%) once agents successfully land on the correct informational pages.

What to fix

For each item: its current state, why it matters, how to fix it and how to check. The same content comes as files you can hand straight to whoever does the work.

Putting the items together…

Fix files

Files generated from this audit. Paste them in, or hand them to a coding agent. Have someone review them before they go live.

Building the fix files…

Detailed results

The results behind the score, and the measurement record.

Discovery

9/20

Access

24/30

Navigation

17/30

Completion

17/20

1.Can an agent get into the site at all?90/100

HTTPREDIRECT

Evidence

homepage HTTP 200
redirect hops: 1

2.Does the site let agents through rather than block them?70/100

ROBOTSBOT

Evidence

robots.txt contains Disallow rules
First byte after 496ms

3.Does it publish machine-readable guidance?15/100

LLMS.TXTSITEMAPROBOTS

Evidence

/llms.txt — absent (HTTP 403)
/llms-full.txt — absent (HTTP 403)
/sitemap.xml — absent (HTTP 404)
/sitemap-index.xml — absent (HTTP 400)
/sitemap_index.xml — absent (HTTP 400)
/robots.txt — present

4.Is the raw HTML readable without running JavaScript?55/100

HTMLSSR

Evidence

6 links in the raw HTML

5.Does the site state its fees or rates as numbers?100/100

FEESFULL

Evidence

outcome success · verdict full
the final answer cited this site
site-guided navigation 100%
2xx rate 100%
done_unaided

6.Can an agent check who is eligible?85/100

ELIGIBILITYFULL

Evidence

outcome success · verdict full
the final answer cited this site
site-guided navigation 25%
2xx rate 100%
done_big_detour

7.Does the site guide an agent to support?100/100

SUPPORTFULL

Evidence

outcome success · verdict full
the final answer cited this site
site-guided navigation 100%
2xx rate 100%
done_unaided

8.Can an agent find branch or store locations?75/100

LOCATIONSPARTIAL

Evidence

outcome success · verdict partial
the final answer cited this site
site-guided navigation 100%
2xx rate 100%
done_unaided

9.Does the site guide an agent through signing up?93/100

SIGNUPFULL

Evidence

outcome success · verdict full
the final answer cited this site
site-guided navigation 67%
2xx rate 100%
done_big_detour

10.Does it state the refund or cancellation rule with a deadline?65/100

POLICYPARTIAL

Evidence

outcome success · verdict partial
the final answer cited this site
site-guided navigation 50%
2xx rate 100%
done_big_detour

11.Can an agent reach the actual document file?75/100

DOCSPARTIAL

Evidence

outcome success · verdict partial
the final answer cited this site
site-guided navigation 100%
2xx rate 100%
done_unaided

12.Can an agent reach a human contact?80/100

CONTACTFULL

Evidence

outcome success · verdict full
the final answer cited this site
site-guided navigation 0%
2xx rate 100%
done_big_detour

13.Can an agent reach the latest notice, with its date?55/100

LATESTPARTIAL

Evidence

outcome success · verdict partial
the final answer cited this site
site-guided navigation 0%
2xx rate 100%
done_big_detour

14.Can an agent compare two plans or products?100/100

COMPAREFULL

Evidence

outcome success · verdict full
the final answer cited this site
site-guided navigation 100%
2xx rate 100%
done_unaided

WebMCP — what the site declares about itself

Not checked

This audit ran with the WebMCP check switched off. That does not mean there are no tools — it means we did not look. Turn on "Check WebMCP" when re-auditing.

Act-stage check (nothing submitted)

Checks whether an agent could fill in the forms and press the buttons, up to the point of pressing. No form was submitted, nothing added to a cart, nothing paid. Report only; not scored.

The act stage was not checked in this audit.

Per-task results

Task	Verdict	Own-site share	Request success	Followed guidance	Cited own site
fees & rates	Full	100%	100%	100%	Yes
am I eligible	Full	100%	100%	25%	Yes
get support	Full	100%	100%	100%	Yes
find a branch	Partial	100%	100%	100%	Yes
create an account	Full	100%	100%	67%	Yes
refund policy	Partial	50%	100%	50%	Yes
find a document	Partial	100%	100%	100%	Yes
contact a human	Full	100%	100%	0%	Yes
latest notice	Partial	100%	100%	0%	Yes
compare two options	Full	100%	100%	100%	Yes

About this report

Measured: 9/10/2026, 9:49:23 PM GMT+9
Engine: Gemini · gemini-2.5-flash · HTTP

This is what AI agents could do on the site at the time above, measured by actually using it. It is not a verdict on the site or the company, and the result can change when the site does.

Methodology →About AOO-Bot and opting out →