Soliq.
The intelligence engine inside Tilarq.
Soliq is not an LLM. It is the orchestration layer we built so Tilarq products are never tied to a single model it calls multiple LLM APIs (Claude, GPT, Gemini, Llama and others), grounds them on your data, evaluates the output and picks the right one for each job.
Ti as in Tilarq stands for Techliphant Intelligence.
Three layers. One coherent system.
Tilarq products on top. Foundation models on the bottom. Soliq in the middle as the engine that routes calls, grounds answers, runs workflows and proves what happened.
Product Layer
Tilarq CRM today, and every future Tilarq product. This is what your team uses.
Soliq
Model Routing · Retrieval · Agent Workflows · Eval Studio · Guardrails · Audit
Foundation Models
OpenAI · Anthropic · Google · Llama · Mistral · fine-tuned on private infra
Eight capabilities. Built once.
Each one applies to every Tilarq product and every services engagement we deliver. Built so we never have to solve the same AI plumbing twice.
Model Routing
A single interface across Claude, GPT, Gemini, Llama, Mistral and others. Every call goes to the model best at that specific task today.
Retrieval & Grounding
Hybrid vector + lexical search with re-ranking and citation tracking. Answers come from your data, not the model's guesswork.
Agent Workflows
Multi-step actions inside our products updating records, drafting messages, generating documents with tool registries and human checkpoints.
Eval Studio
Every capability ships with a graded eval set. Quality drift is caught and rolled back, not discovered by a customer.
Guardrails
PII redaction, prompt-injection defences, structured outputs and brand-voice constraints applied uniformly.
Audit Logs
Every prompt, response and tool-call logged. Compliance, support and engineering can replay exactly what happened.
Tailored Per Use Case
A scoring prompt is not a drafting prompt. Soliq tunes the model, temperature, retrieval and guardrails for each job in each product.
Model-Agnostic by Design
When a better model lands, we add it to the routing policy. The product on top doesn't change. You inherit the upgrade.
Inside Tilarq CRM today. Everywhere we ship AI.
Every AI feature in Tilarq CRM is a Soliq call underneath. Every AI services engagement we deliver ships on Soliq too.
Lead scoring inside Tilarq CRM
Soliq routes scoring calls to the model that handles your data shape best, calibrated against your closed-won outcomes.
Deal coaching & next-best actions
Grounded suggestions on every open deal drawing from your reps' winning patterns, not generic playbooks.
Conversation summarisation
Calls, emails and WhatsApp threads summarised into structured CRM updates the next person can actually use.
Pipeline forecasting
Probabilistic forecasts that correlate to actual revenue, with every assumption auditable.
Drafting & follow-ups
Outbound and follow-up emails drafted in your voice, with the right model picked per stage and per audience.
Services engagements
Custom AI builds we deliver for clients (document AI, support automation, knowledge search) all ship on Soliq underneath.
Principles baked into the engine.
Right model for the job
No single model wins every task. Soliq picks per call and switches when something better lands.
Grounded, not guessed
Retrieval and citation tracking on every answer that touches your data. No invented facts.
Evals before launch
No capability ships without a graded eval set. Drift detection on real traffic catches regressions early.
Audit everything
Every prompt, response and tool-call logged for compliance and incident response.
No single model. The best one for each call.
Models and retrieval infrastructure. The list grows as new options prove themselves on our evals the product on top doesn't change when it does.
Questions about Soliq.
No. Soliq is the engine that orchestrates LLMs. It calls Claude, GPT, Gemini, Llama and others under the hood and picks the right one for each task so Tilarq products keep getting smarter as models improve, without a rewrite on top.
No single model is best at everything. Pricing, latency, reasoning quality and safety behaviour shift constantly. Soliq lets us choose per use case and switch when something better lands, without touching the product code.
Today, models from OpenAI, Anthropic, Google, Meta and Mistral, plus fine-tuned models on private infrastructure where the use case justifies it. The list grows as new models prove themselves on our evals.
Today, Soliq lives inside Tilarq products and our services engagements. We are not selling it as standalone infrastructure. If you want Soliq working on a custom problem, talk to us about a services engagement.
Techliphant Intelligence. Tilarq the product brand carries the Ti prefix that marks every Techliphant Intelligence product.
Ready when you are
Let's build something exceptional.
Tell us about your business, your stack, and the problem you are trying to solve. We respond with a clear next step usually a 30-minute discovery call, no fluff.