PKSHA Technology holds roughly a quarter of the Japanese voice bot market and is deployed at Mizuho Bank among other large Japanese enterprises. It is the domestic leader in a market that rewards local context, formal language handling, and long vendor relationships. Kaigen Labs operates in the adjacent lane: international and India-led sales motions that touch Japan rather than live there.
PKSHA ships VoiceAgent and AIAgent products tuned for Japanese-language enterprise customers, with deep keigo (formal honorific language) handling and a strong domestic compliance posture. Kaigen Labs builds and runs multi-channel sales systems for English, Hindi, and increasingly Japanese deployments delivered through partner channels.
The two are not competing for the same buyer as often as a comparison page implies. This guide maps where each one earns its keep, what each one leaves on your desk, and the questions that settle the choice faster than any feature list.
24.7%
PKSHA Technology's share of the Japanese domestic voice bot market.
46.3%
Projected CAGR for the Japanese voice AI market through 2030.
8days
Mandatory cooling off period after Japanese telemarketing contracts.
TL;DR
PKSHA Technology is the domestic answer. A Japanese enterprise buying Japanese AI for Japanese customer communication gets language depth, enterprise references, and a vendor relationship that lives entirely inside Japan.
Kaigen Labs is the cross-market answer. One team designs, builds, and operates your outbound motion across voice, WhatsApp, SMS, and email, with CRM write back and weekly tuning included. How a managed deployment works covers the model in full.
The fork in the road is geography plus ownership. A domestic deployment run to Japanese enterprise standards points at PKSHA. An India to Japan or global outbound motion that needs one system across markets, operated for you, points at Kaigen Labs.
At a glance
This matrix tracks ownership rather than checkboxes. A full disc means the capability ships and someone else runs it for you. A half disc means the pieces exist but your team assembles and maintains them. An outline means you build it yourself. Read every row with the same question: who is running this in month six?
| Capability | Kaigen Labs | PKSHA Technology |
|---|---|---|
| Japanese keigo depthPKSHA leads here; Kaigen delivers keigo through partner deployments | ||
| Domestic enterprise support in JapanPKSHA is the domestic vendor with local teams; Kaigen supports Japan through partners | ||
| Outbound orchestration across voice, text, and emailPKSHA voice and chat ship as separate products; the outbound sequence is yours to design | ||
| CRM write back run for youConnectors exist on PKSHA; your team wires fields, triggers, and retries | ||
| Ongoing tuning and monitoring after launchPKSHA implementation teams deliver the build; the ongoing cadence is set by contract | ||
| Pilot live in two to three weeksEnterprise engagements in Japan typically run months of procurement and build | ||
| Warmup texts before outbound calls | ||
| Failover across voice providersKaigen routes across multiple runtimes; a single vendor stack has no outside backup |
HEAR IT FOR YOURSELF
Reading about voice quality only gets you so far.
The live demo on our homepage runs a real Kaigen voice agent in your browser. Pick an industry, start a call, ask whatever you want. Hang up whenever you have heard enough.
Try the live demo →Voice quality and latency
On Japanese voice, PKSHA holds an edge that is hard to argue with. Keigo is not a translation problem; it is a social register problem, and a vendor that lives inside the market handles it with a depth an international stack matches only through local partners. For a Japanese-only deployment, that depth matters.
Everywhere else, voice has stopped being the deciding factor. The serious platforms all sound natural and respond fast now, which moves the evaluation to the layer around the call: what fires when nobody answers, how the next touch knows about the last one, and who is reading the transcripts when your offer changes. That layer, not the audio, is where the rest of this comparison lives.
Channel orchestration is where the difference lives
PKSHA's products are voice and chat for the Japanese market, where the channel pair that matters is phone plus LINE. They are strong products doing separate jobs. Turning them into one outbound sequence, a text that warms the call, a call that knows about the text, a follow up that references both, is integration work your team scopes, builds, and keeps alive.
Kaigen Labs sells that sequence as the product, because sequencing is what moves the numbers. Research on outbound sequencing puts the pickup lift from a short text sent minutes before the call well above baseline, and industry benchmarks put text open rates far above email. In response studies, a voicemail followed at once by a text outperforms the voicemail alone. In India, where WhatsApp is the default business channel, the same play runs on WhatsApp instead of SMS.
On a Kaigen deployment that shows up as a seven day cadence running on one conversation memory. The agent that calls has read the text thread. The email that follows references the call. The CRM updates at every step, with no Zapier chain holding it together.
DAY 1
WhatsApp / SMS
A heads up text so tomorrow's call arrives expected, not cold.
DAY 2
Text + AI call
Warmup text minutes before the call. Voicemail plus text if no answer.
DAY 4
Proof in writing: a short case study matched to their market.
DAY 5
AI call retry
A second attempt at a different hour, carrying every prior touch.
DAY 7
Soft close
A no pressure message that keeps the door open.
CRM write back and integrations
PKSHA integrates with the CRM and contact center stack a Japanese enterprise already runs, and in that world integration is usually delivered by the implementation project or a systems integrator. The connectors are real. What follows the connectors, field mapping, trigger logic, retries, and the quiet schema change when an admin renames a property, becomes your operations backlog once the project team rolls off.
Kaigen Labs ships native write back for the CRMs we deploy most, HubSpot, Salesforce, Airtable, Pipedrive, and Close: lead status, call summaries, sentiment, structured qualification fields, and the transcript, landed on the right object in the right pipeline. Anything outside that set gets wired during the Build phase, usually in days. When a connector breaks at 2am, it pages the Kaigen team, not yours.
Deployment model: who owns the build, monitoring, and tuning
Neither of these is a DIY developer platform, which makes this section subtler than most comparisons on this blog.
PKSHA runs enterprise engagements: implementation teams that design, build, and deploy over multi-month projects, matched to the procurement rhythm of large Japanese organizations. If your governance expects a domestic vendor, a named project team, and a multi-year relationship, that model is a feature, not a cost.
The Kaigen team runs a tighter loop built for sales motions rather than enterprise programs. The Kaigen Method has four phases: Assess, Build, Deploy, Optimize. We map the funnel and find where leads go cold, build the playbook and the agents across every channel wired into your CRM, launch on a focused slice of volume with most pilots live in two to three weeks, then tune weekly on evals from real conversations with monthly performance reviews.
01
Assess
Funnel mapping, leak analysis, baseline metrics.
02
Build
Conversation playbook, agents on every channel, CRM wiring.
03
Deploy
Focused launch, live in two to three weeks, human in the loop.
04
Optimize
Weekly evals and tuning, monthly performance reviews.
The phase gates exist because the failure pattern in this category is well documented. A preliminary MIT Media Lab report found that 95 percent of generative AI pilots show no measurable financial return within six months, and the miss is rarely the model. It is the absence of anyone owning the system once the launch excitement fades.
Built, not assembled. Managed, not abandoned.
The Kaigen Labs operating principle.
Compliance and security
Domestic compliance is a genuine PKSHA strength. Japanese enterprise buyers, banking above all, hold vendors to exacting data handling standards, and a domestic vendor deployed in that tier has cleared them. If your entire risk surface sits inside Japan, that posture is a strong argument for staying domestic.
Outbound calling across borders multiplies the rulebooks. Japanese telemarketing rules require the caller to state the business name and the purpose of the solicitation at the start of the call, and contracts closed by telemarketing carry the eight day cooling off period noted above. In the United States, the FCC confirmed in February 2024 that AI-generated voices count as artificial under the TCPA, so outbound AI calls must disclose the artificial voice up front. In India, TRAI requires designated number series, 140 for promotional and 1600 for transactional, with explicit consent and DND respect. Our voice AI compliance guide walks the full map.
The operational difference: on a deployment you operate yourself, you keep three regulators' rules current in your scripts and dialing logic. On Kaigen, disclosure language, number series, and consent checks live inside the prompts and the dialing layer we operate, and the Kaigen team updates them as the rules move.
Languages and regional fit
For Japanese depth the ranking is not controversial: PKSHA first for the domestic enterprise tier, and the matrix above says so. Kaigen's Japanese deployments run through partner channels, which works when the motion is mid-market outreach and the winning ingredient is the sequence rather than the deepest possible keigo.
Kaigen's language work runs wide rather than deep in any single market: English and Hindi as defaults, vernacular handling for Indian cities beyond the metros, where industry surveys consistently find most leads prefer Hindi or a regional language over English, and local phone numbers per country because an unfamiliar country code depresses pickup. Language and numbers are set up during Build, not left as homework. The multilingual voice AI guide goes deeper.
WHEN PKSHA TECHNOLOGY WINS
Pick PKSHA Technology if…
- You are a Japanese enterprise and the deployment stays domestic
- Strict keigo handling and Japanese compliance sit at the top of your requirements
- Your buyers are Japanese banking, insurance, or BFSI customers at Mizuho tier
- A long procurement relationship with a domestic vendor fits your governance
- International expansion is not on the near horizon
WHEN KAIGEN WINS
Pick Kaigen Labs if…
- Your motion spans markets, India to Japan, or India to the UK and US, and needs one system
- You want voice, SMS, WhatsApp, and email orchestrated as one sequence
- You want CRM write back with no glue code on your side
- You have no AI engineering team and no plans to hire one
- You want someone else monitoring calls and tuning prompts as your offer evolves
MAP YOUR MOTION
Want us to sketch this for your sales motion?
Twenty-minute call. You bring the sales motion you are trying to scale; we sketch the agent, the channels, the integrations, and the metrics we would target. No deck, no pitch.
Book a 20-minute audit →A concrete walkthrough: India to Japan expansion
A growing number of Indian SaaS and EdTech companies are looking at Japan as the next market. That motion looks nothing like domestic Japanese enterprise sales. It is a foothold play: mid-market Japanese accounts, reached through a sequence that respects Japanese business norms, feeding a local partner team.
Here is the shape of that deployment with Kaigen Labs running behind a Japanese partner.
Day zero (account research): An ICP-aligned Japanese account list lands in Kaigen, segmented by company size and procurement profile, with the language layer set to Japanese keigo by default.
Day one (formal email introduction): A Japanese-language email with the honorifics expected for first contact, a one line value proposition, and a calendar link for a discovery call.
Day three (AI voice call with disclosure): The agent calls, opens with the business name and solicitation purpose disclosure Japanese telemarketing rules require, asks two qualifying questions in keigo, and either books a meeting for the partner sales team or schedules a follow up.
Day five (case study email): A personalised email with a Japanese-language case study from a comparable mid-market deployment.
Day ten: A second touch with refreshed context, run with the eight day cooling off requirements for telemarketing contracts in view.
The AI is not replacing the partner's sales team. It is doing the volume work that decides whether the partner team gets to talk to Japanese mid-market accounts at all.
How to evaluate
Q1
Where do your buyers sit?
All in Japan, buying from Japanese vendors: the domestic logic favors PKSHA. Spread across India, Japan, the UK, or the US: one cross market system beats one per country.
Q2
Is the job customer communication or outbound sales?
PKSHA's center of gravity is enterprise contact and communication AI. If the job is filling a pipeline, the sequencing layer matters more than contact center pedigree.
Q3
Who owns the sequence across channels?
Separate voice and chat products mean your team owns the choreography between them. Decide whether that project belongs on your roadmap.
Q4
Who is tuning it in month six?
An implementation team rolls off, and a procurement cycle does not tune prompts. Name the owner before you sign, whichever side you pick.
KEY TAKEAWAYS
- This is a geography and ownership decision more than a feature decision. PKSHA Technology is the domestic Japanese enterprise answer; Kaigen Labs is the managed system for cross market outbound.
- Read the matrix by who runs each capability in month six. The half discs are where hidden headcount lives.
- PKSHA leads on Japanese depth and domestic trust. One orchestrated motion across voice, text, and email, with CRM write back and weekly tuning run for you, is what Kaigen Labs ships.
FAQ
Is PKSHA Technology a good choice for Japanese enterprises?
For a domestic Japanese deployment, yes. PKSHA holds roughly a quarter of the Japanese voice bot market, runs at institutions like Mizuho Bank, and brings keigo depth plus a domestic support relationship that international vendors struggle to match. This comparison exists for buyers whose motion crosses borders.
How long does a Kaigen Labs launch take?
Most pilots are live in two to three weeks: discovery in week one, build and quality checks in week two, soft launch on a focused slice of volume in week three. That cadence is the structural opposite of a multi-month enterprise implementation, which is exactly why the two models suit different buyers.
Can Kaigen Labs handle Japanese keigo properly?
For mid-market deployments through Japanese partners, yes. For domestic enterprise work where keigo depth is mission critical, at large banks for example, PKSHA and other domestic vendors remain the stronger fit, and this guide says so plainly.
Does Kaigen Labs replace a Japanese partner sales team?
No. In an India to Japan motion the AI runs the volume layer: researched first contact, disclosure compliant calls, qualification in keigo, and scheduling. The partner team takes every qualified conversation from there. The system decides how many of those conversations exist.
What if we run PKSHA Technology today and want to expand beyond Japan?
That is the expansion case we see most. The domestic deployment can stay put; Kaigen builds the international motion beside it, one orchestrated sequence across voice, WhatsApp, SMS, and email with CRM write back, and both systems feed the same pipeline. Nothing forces a rip and replace.
What happens when a voice provider has an outage?
Kaigen orchestrates across multiple voice, language model, and telephony providers, so traffic reroutes to a backup automatically and your callers do not notice. A stack built on a single vendor's runtime, however good that vendor is, has nowhere to fail over to.
The decision in one sentence
A Japanese enterprise running Japanese customer communication should shortlist PKSHA Technology; a company running an outbound sales motion across markets and channels, and wanting one team to build and run it, should talk to Kaigen Labs; both are strong answers to two different questions.
If you want to see what a Kaigen Labs build would look like for your motion, the next step is a twenty minute audit. Book a slot.
Weighing more than one platform? See how Kaigen Labs compares with PolyAI and Retell AI. And for the full map of the six layers any deployment runs on, start with the production voice AI stack.
