AI models

The SIMCOAI models by name - which one answers, what each is for, and how the choice shows up in your call quality and on your bill.

Named

The SIMCOAI models

Every SIMCOAI model is a configuration SIMCOAI builds and runs on top of a frontier base model: its own instructions, its own context routing, its own response deadlines, token budgets and guardrails. What you choose between is the configured assistant, not the engine - which is why the engine underneath can be replaced with a better one without your choice or your call behaviour changing.

Model	What it is for	Speed
SIMCOAI Lite	Ordinary, high-volume questions - opening hours, where an order is, taking a message - at the lowest cost per answer.	Fast
SIMCOAI Swift	Live phone calls. This is what answers your phone by default, because a caller left in silence hangs up.	Fastest - about a second to first word
SIMCOAI Core	The everyday balance of speed and depth, and the default for anything that is not on a live call.	Balanced
SIMCOAI Insight	Policy questions, complex refunds and setup advice - where the answer matters more than a second of latency.	Slower, and deliberately so
SIMCOAI Live	Speech straight to speech, with no transcription step in between. Keeps more of how something was said.	Lowest latency, experimental

These are SIMCOAI's names for SIMCOAI's own configurations. SIMCOAI does not claim to have trained a base model from scratch, and every provider it relies on is named in full in the privacy policy and on the GDPR and DPA page - naming who processes your data is a legal requirement rather than a marketing choice.

Managed

Models are managed for you

You choose a model rather than wiring up a provider account, and SIMCOAI operates it. There are no customer-owned API keys and no custom endpoints.

If a model has a problem, SIMCOAI is designed to fall back to another rather than fail the call outright. Anything affecting service appears on the status page.

Profiles

How the choice is made

In the dashboard you pick a profile and a routing mode. The profile decides which SIMCOAI model answers; the routing mode decides whether that is fixed or varies by what is being asked.

Profile	Model	Notes
Economy	SIMCOAI Lite	Lowest cost per answer
Fast	SIMCOAI Swift	Default for live calls
Balanced	SIMCOAI Core	Default off a call
Premium	SIMCOAI Insight	Needs Pro, or the Premium AI Models add-on

Fixed: Every conversation uses the profile you chose.
Auto by enquiry: Routine calls stay fast, and complex policy, setup, legal, automation and evidence questions are routed to a stronger model where your plan or add-on allows it.

Cost

How model use is billed

Model usage is metered as AI usage against your plan. SIMCOAI Insight consumes more of it for the same conversation, and on most plans needs the premium add-on.

Careful. Switching everything to premium is rarely the right first move. Check whether the problem is actually missing knowledge - see knowledge and documents - because a better model cannot invent facts you never supplied.

See usage and limits for how AI usage is counted, and add-ons for premium model access.

Choosing

A sensible way to decide

Start where the defaults put youSIMCOAI Swift on calls, SIMCOAI Core elsewhere. That handles the majority of front-desk conversation.
Read a day of transcriptsLook for specific failures, not general impressions.
Separate the two failure typesMissing knowledge is a knowledge fix. Missed nuance is a model question.
Change one thingThen compare the next day of calls against the same measures.

Expected result

A decision based on your own transcripts rather than on which model sounds better.