Prompt-based agents vs Conversational Pathways (2026)
Samuel Oram-Jones, 6 min read
Prompt-based agents vs conversational pathways is a choice about how much of a call is scripted. A prompt-based agent follows one set of instructions from start to end, while conversational pathways split a call into rule-linked steps. This page covers where each fits, what Retell says about its version and how Bland combines both as of November 2026.
What do prompt-based agents and conversational pathways mean on Bland?
On Bland, a prompt-based agent keeps its instructions, tools, and knowledge in one prompt that its language model follows for a whole call, while conversational pathways add steps, branches, and checks to the parts of a call that need them. Bland's agent builder holds prompts, tools, knowledge, variables, and pathways inside one agent, so a team isn't picking between two products, only deciding where structure belongs.
A prompt carries a call well when the goal is open and the order of questions doesn't matter, such as answering a question about opening hours. A pathway carries it better when a step has to happen every time. Bland describes pathways as branching, multi-turn conversation flows with state and tool calls.
Bland also lists a custom code node among its enterprise features, which runs custom logic inside a pathway during a live call.
How does Retell describe single prompts and Conversation Flows?
Retell sells the same choice as two separate agent types, a single prompt agent and a Conversation Flow agent built from nodes, and it steers teams towards the first.
Retell markets flows for "precise call control" and also gives its reason "why to start with a single prompt agent first". A team on Retell picks a build path before it knows how callers behave, whereas on Bland the pieces already share one agent, so no such choice comes first.
Which fits better, a prompt-based agent or Bland pathways?
A prompt-based agent fits open, low-stakes calls where a wrong turn costs little, and Bland pathways fit calls with steps that must happen every time. A Bland agent can combine the two, with a prompt handling the conversation inside each step and a pathway holding the rules between steps.
A call that must verify a caller, read back a figure, or hand over to a person suits a pathway, because those steps can't be left to chance.
Use a prompt when
- The goal is open, such as answering a question about opening hours
- The order of questions doesn't matter
- A wrong turn costs little
- Callers are likely to change course halfway through
Use pathways when
- A caller must be verified before anything else happens
- A figure has to be read back exactly
- A call must hand over to a person at a set point
- A step has to happen on every call
A typical pathway step is a node that collects and confirms a rental record number, with one branch for a confirmed record number and one for a failed collection. A prompt could ask for the same number, but a pathway lets a team require that the step runs on every call and that a failed collection goes to a defined next step.
Take a rental return call as a worked example.
A prompt-led version gives the agent one instruction to help the caller return a vehicle and leaves the order of questions to the language model. A pathway-led version puts the record number in a step of its own, with a confirmed branch and a failed-collection branch, while a prompt still handles the conversation inside that step.
A team can start with the prompt and add a step only where a call has gone wrong or a step can't be skipped.
Before a change reaches callers, Bland moves an agent from staging to production only when chosen simulated test cases pass chosen judges.
Caller behaviour is what separates the two in practice. A caller who gives a date of birth before being asked, or changes an appointment request halfway through, tests a rigid sequence hardest. A prompt absorbs those detours easily and a pathway keeps the steps that must not be skipped, so putting a pathway only around the steps that need one limits how often a detour breaks a call.
Bland also measures how fast a caller speaks and adjusts how soon the agent cuts in, slowing down for a caller who says they have a speech impediment.
A team that isn't sure where to start can ask Norm, Bland's agent-building copilot, to help build a first version. A pathway is built by connecting nodes, each holding its own instructions and a plain-language condition for moving on.
How do Bland pathways carry over to iMessage and text?
Bland runs branded iMessage agents natively on Enterprise, and they share the same pathways as its voice agents, so logic built for a phone call carries over to a text thread. Those agents send and receive photos, videos, files, voice memos, and tapbacks, and Bland also runs SMS and web chat.
Retell doesn't offer iMessage agents as of November 2026, so a team that builds a Conversation Flow there reaches customers by phone and SMS.
What do Bland pathways cost?
Bland pathways cost $0.14 per minute on the self-serve Start plan and $0.12 per minute plus a $299 monthly platform fee on self-serve Build, before phone, and Build is cheaper than Start above about 14,950 minutes a month. Bland includes pathways, integrations, and automations on Start, Build, and Enterprise, with no token charges on top of the per-minute rate.
Every rate is listed under pricing, and either self-serve rate covers the language model, transcription, and speech, and phone is billed separately, on a customer's own carrier or on Bland's at pass-through cost.
Enterprise pricing is custom, set with Bland's sales team.
Enterprise is the plan regulated teams buy. It adds native iMessage agents alongside SMS and web chat, warm transfers, guardrails, a BAA, SSO, data residency, dedicated orchestration with dedicated GPUs run by Bland, and forward-deployed engineers who work beside a customer's team, with concurrency contracted to its volume and custom pricing.
What do pathways look like in production?
Bland is full stack, so one company builds the language model, speech, and transcription that run a pathway, and its language model is a closed-source model that Bland developed itself. Calls run at sub-200ms latency with a 99.9% uptime SLA on every plan.
Bland serves more than 250 enterprise customers, with over 1.3 billion calls handled. Idaho Housing and Finance Association saves $750,000 a year by replacing its IVR with Bland, which handles 4,000 inbound calls a day and cut average call time by 20%.
"We save thousands of dollars a day."
Greg Blake, Chief Information Officer, Idaho Housing and Finance Association
Teams weighing other build choices can read No-code vs custom-code voice AI agents or Zendesk AI vs Fin, and Bland vs Retell sets out the full comparison. A demo on a real call workflow shows a pathway running, and Best Air AI competitors and Best Ada alternatives cover the wider market.
Frequently asked questions
Is a prompt-based agent or conversational pathways better for phone calls?
Neither is better on every call: a prompt-based agent suits open, low-stakes conversations, and conversational pathways suit calls with steps that must happen every time. Bland puts both on one agent, with a prompt inside each step and a pathway between steps.
Is Bland only for developers?
No, Bland isn't only for developers. Pathways gives teams a visual builder, Norm helps build agents, and on Enterprise forward-deployed engineers build alongside a customer's team.
What does Retell say about Conversation Flows?
Retell says flows give "precise call control" and recommends starting with a single prompt agent first.
Does Bland's rate include phone?
No, Bland's rate covers its language model, transcription, and speech, and phone is billed separately.
Does Retell support iMessage?
Retell doesn't offer iMessage agents as of November 2026. Bland runs branded iMessage agents natively on Enterprise, which send and receive photos, videos, files, voice memos, and tapbacks, and share the same pathways as its voice agents.
