What is Ika, the conversational voice layer built into IKI Health's BeHealthy programs?

Ika is the conversational voice AI assistant yeshcube developed for IKI Health Group on the Somia architecture and approved with Somia Within. The Ika brand belongs to IKI Health, which has deployed it in its BeHealthy pilots to support participants through short conversations: helping them make sense of how they are doing, reinforcing adherence and sustaining habit change through small, manageable steps. It debuted in the BeHealthy Calvià pilot, run by the Ajuntament de Calvià together with IKI Health, and then reached BeHealthy San Francisco, the model’s first pilot in the United States.

Short support between program sessions

Prevention programs work in sessions. Between one and the next, a gap of days or weeks opens up in which the person is left alone with their habits, and that is where adherence usually dissolves. Ika fills that space with a short voice conversation that picks up what was worked on and restores order.

The conversational design draws on the fundamentals of motivational interviewing and follows a stable structure: opening, listening, reformulation, brief exploration and closing. The opening invites the participant to express how she is doing. The reformulation organizes what she has said and returns it with clarity. The closing proposes a synthesis, a reflection or a manageable action. In Calvià, the most frequent themes were tiredness, stress, irregular sleep, lack of time, family overload, low mood and motivation to move. The conversation stays in the territory of everyday wellbeing and outside any clinical intervention.

Somia and Somia Within, the basis Ika runs on

Ika rests on the Somia™ architecture, yeshcube’s modular conversational artificial intelligence system for everyday emotional wellbeing, and uses its voice layer. Somia interprets the context of each interaction and adjusts the language and the flow of the conversation. When a situation falls outside the preventive framework, it activates a crisis protocol that recognizes sensitive scenarios and offers the appropriate response. Ika always makes its status as artificial intelligence explicit and does not replace professional care.

The integration was carried out under Somia Within, the framework with which yeshcube approves the incorporation of Somia into external products and services against ethical, technical, privacy and measurement criteria. In BeHealthy, that framework meant explicit consent from participants, processing in line with the GDPR, anonymisation and automated analysis with no human access to identifiable conversation content. The conversation enters the program with a specific function, defined limits and traceability.

What the Calvià pilot showed

BeHealthy Calvià was a six-month pilot for 150 women over 35, run by the Ajuntament de Calvià together with IKI Health. The BeHealthy app worked as the main digital environment, with structured sessions, device synchronization, health records and FeelTrack as an emotional self-tracking tool. Ika was added as a voice layer.

Initial demand was high: the places filled in under 72 hours. Subsequent take-up traced a pattern familiar in these programs. Low-effort milestones, such as answering the survey or enabling synchronization, reached roughly three in four participants. Deep use, measured in completed sessions and FeelTrack, concentrated in a core of 34.8%. Average retention observed was 49.8 days, with the sharpest drop in the first fortnight.

The reception of Ika was the clearest finding. Across 122 valid evaluations and 243 responses, the overall evaluation of the experience reached 92.8%, with perceived usefulness of 92.2% and perceived wellbeing of 93.4%. 86.4% of responses sat at the top value of the scale. The firmest indicator was the absence of negative wellbeing ratings in the valid sample. The only two negative responses concerned usefulness, which places the room for improvement in the practical clarity of some conversations. These results describe perception as reported by participants and their experience of use, with no diagnostic or clinical efficacy value.

Voice also acted as a spontaneous channel for questions about the product. Ika collected 26 functional queries, most of them about connecting wearable devices to the app. That double signal — support and unsought help desk — helps to understand how the program is used in real conditions.

In terms of maturity, yeshcube places this transfer of Somia on the ERL scale (Evidence Readiness Level), with which it measures whether a solution is ready to be transferred:

Voice and adherence: what the data showed

The core of the analysis compares activity in the app between those who spoke with Ika and those who did not. The differences were consistent across every indicator. Users with Ika (63) completed an average of 3.1 sessions against 0.9 for participants without Ika (95). They used FeelTrack in 68.3% of cases against 12.6%, and synchronized in 92.1% against 46.3%. Among recurring users the signal is stronger still: 70.6 days of average synchronization against 42.5.

These numbers should be read with care. Participation was voluntary, so the most motivated participants may have used the app more and also spoken with Ika more. The data describes an association between usage variables. Causal direction cannot be established with this design. What can be stated is that Ika use identifies the segment most engaged with the program, without attributing that greater adherence to Ika on its own.

From Calvià to San Francisco

The collaboration between yeshcube and IKI Health moved from a local pilot to a second program in another country. BeHealthy San Francisco takes the model to the United States with a three-month pilot in the Bay Area, for a cohort of 60 women aged 40 to 55 in perimenopause, under the direction of dietitian-nutritionist Lisa Booth alongside IKI Health’s medical and scientific team.

San Francisco brings a territorial, cultural and care context different from Calvià’s, and with it a new setting in which to study how voice conversation behaves within a structured prevention program. The incorporation of Somia into that pilot is detailed in the integration of Somia into BeHealthy San Francisco. yeshcube will present the results of this second integration once they are available.

What Ika is for and what it is not

Ika gives continuity to support between program sessions, within the preventive framework and through short conversations focused on habits and everyday wellbeing. It does not make diagnoses, does not provide medical or psychological services and does not replace professional care. Faced with a situation that exceeds that framework, it refers on through the crisis protocol.

Ika takes yeshcube’s work on the Somia architecture and its Somia Within integration framework into an external context.

Does your organization run prevention or wellbeing programs that could incorporate conversational technology?

Discover Somia Within →

Fill in our contact form →

The BeHealthy Calvià data comes from the results reports and the integration case study produced by Yeshcube Tech, S.L. It has been handled anonymously and in line with the GDPR, and it describes use, perception and experience, with no diagnostic, treatment or clinical efficacy value.