The Therapeutic Alliance in the Digital Age: What the Research Actually Says

Key takeaways

• Bordin's Alliance Model: A Framework That Still Holds
• The Alliance-Outcome Relationship: What Flückiger's Meta-Analysis Shows
• Berger's Research on the Digital Therapeutic Alliance

Bordin's Alliance Model: A Framework That Still Holds

Edward Bordin's 1979 paper "The Generalizability of the Psychoanalytic Concept of the Working Alliance" remains the theoretical foundation for most alliance research. Bordin proposed that the therapeutic alliance consists of three interdependent components: agreement on the goals of therapy, agreement on the tasks that will achieve those goals, and the emotional bond between therapist and client. This tripartite model was deliberately pan-theoretical — Bordin argued that the working alliance operates across all forms of therapy, not just the psychoanalytic tradition where the concept originated.

The elegance of Bordin's model is its simplicity and clinical utility. Goals refer to the shared understanding of what therapy is trying to accomplish — symptom reduction, improved relationships, greater self-understanding, or whatever the client and therapist have agreed to work toward. Tasks refer to the specific activities of therapy — the interventions, techniques, homework, and in-session processes that constitute the therapeutic work. Bond refers to the quality of the personal connection — trust, warmth, mutual respect, and the felt sense that the therapist genuinely cares about the client's wellbeing.

What makes Bordin's model relevant to the digital tools conversation is its emphasis on agreement. The alliance is not something the therapist provides unilaterally — it is co-constructed through negotiation and mutual understanding. When a therapist introduces a digital tool into the therapy process, the question is not simply whether the tool is good or bad, but whether it is integrated into the shared goals, aligned with the agreed-upon tasks, and implemented in a way that preserves or enhances the bond. A tool that is imposed without discussion or that clients experience as replacing therapeutic contact will undermine the alliance. A tool that is introduced collaboratively and serves the shared therapeutic agenda can strengthen it.

Research has consistently supported the clinical importance of all three Bordin components. A meta-analysis by Tryon and Winograd found that goal consensus and collaboration are significantly related to treatment outcome (d = 0.72), while Horvath and colleagues have shown that bond strength is particularly predictive of outcome in longer-term therapies. The practical implication is that any change to the therapeutic frame — including the introduction of technology — needs to be evaluated through the lens of goals, tasks, and bond.

The Alliance-Outcome Relationship: What Flückiger's Meta-Analysis Shows

Christoph Flückiger and colleagues at the University of Zurich published the most comprehensive meta-analysis of the alliance-outcome relationship in 2018 in the journal Psychotherapy. Their analysis synthesized 295 independent studies covering more than 30,000 patients treated across a wide range of modalities, settings, and clinical populations. The overall correlation between alliance and outcome was r = .278 for face-to-face psychotherapy — a moderate effect size that has remained remarkably stable across decades of research.

To put that effect size in perspective, r = .278 corresponds to a Cohen's d of approximately 0.579, meaning that clients with stronger therapeutic alliances show meaningfully better outcomes than those with weaker alliances. The alliance accounts for approximately 7.7 percent of outcome variance — which may sound modest until you consider that specific treatment techniques typically account for a similar or smaller proportion. The alliance is not the only factor that matters, but it is among the most consistent and robust predictors across the entire psychotherapy research literature.

One of the most striking findings from Flückiger's analysis was that the alliance-outcome relationship was nearly identical in internet-delivered treatments (r = .275) compared to face-to-face treatments (r = .278). This finding directly challenges the assumption that technology inherently dilutes the therapeutic relationship. Clients receiving therapy through digital platforms formed alliances of comparable strength and those alliances predicted outcomes with comparable effect sizes. The medium of delivery appeared to matter less than the quality of the therapeutic interaction within that medium.

The meta-analysis also examined moderators and found that the alliance-outcome relationship was consistent across different types of therapy (CBT, psychodynamic, humanistic), different clinical populations (depression, anxiety, substance use), and different alliance measures. This consistency reinforces Bordin's original assertion that the alliance is a pan-theoretical construct — it matters regardless of what kind of therapy you practice. For clinicians considering digital tools, this means that the question is not whether the alliance can survive technology but whether specific tools and implementations preserve the conditions that support a strong alliance.

Berger's Research on the Digital Therapeutic Alliance

Thomas Berger at the University of Bern has been one of the leading researchers on therapeutic alliance in internet-based interventions. His 2017 narrative review in Psychotherapy Research examined whether a meaningful therapeutic alliance can be established when therapy is delivered through digital channels — a question of considerable relevance as the field increasingly adopts technology-mediated treatment formats. His conclusion was cautiously optimistic: across diverse digital formats, client-rated alliance scores were high and roughly equivalent to ratings found in face-to-face therapy studies.

Berger's review found that this pattern held across different communication modalities (videoconference, text-based, even automated guided self-help), different diagnostic groups, and different amounts of therapist contact. Even in minimal-contact guided self-help programs — where the "therapist" role was limited to brief check-in messages — clients reported alliance levels that were comparable to traditional therapy. This suggests that clients are capable of forming meaningful therapeutic connections through digital channels, and that the alliance construct is more flexible and resilient than many clinicians assume.

More recently, Berger published a 2025 paper in World Psychiatry titled "Rethinking the Therapeutic Alliance in Digital Mental Health Interventions," which pushed the concept further. He argued that the alliance in digital contexts may involve different mechanisms than in face-to-face therapy — for example, the sense of being supported by a system rather than a single individual, or the feeling of autonomy that self-guided programs can provide. These are not lesser forms of alliance but qualitatively different ones that deserve their own conceptual frameworks.

For practicing therapists, Berger's work offers reassurance that introducing digital tools into an existing therapeutic relationship does not inherently weaken the alliance. The evidence suggests that what matters most is not the presence or absence of technology but the quality of the therapeutic interaction, the clarity of communication, and the client's perception that the technology serves their therapeutic goals. A between-session tracking tool that helps the therapist arrive better prepared and more attuned to the client's experience is likely to enhance rather than diminish the alliance.

How Digital Tools Can Strengthen the Alliance

The research literature suggests several specific mechanisms through which digital tools can enhance the therapeutic alliance rather than weaken it. The first is increased continuity. One of the inherent limitations of traditional weekly therapy is the gap between sessions — a gap during which clients face challenges, experience insights, and undergo emotional shifts that the therapist learns about only retroactively and incompletely. Digital tools that maintain a therapeutic thread between sessions — through journaling, mood tracking, or skill practice logging — create a sense of ongoing connection that can strengthen the bond component of the alliance.

The second mechanism is improved attunement. When a therapist reviews between-session data before a session, they arrive with a richer understanding of the client's recent experience. This enables more targeted, relevant session openings and demonstrates a level of attention that clients experience as caring and validating. A therapist who says "I noticed your mood dipped on Wednesday — what was happening then?" communicates a fundamentally different level of engagement than one who opens with the standard "So how has your week been?"

The third mechanism is collaborative transparency. Measurement-based care research, as discussed in Lambert's work, shows that sharing progress data with clients enhances their sense of agency and partnership in the therapeutic process. Digital tools that make data visible to both therapist and client — mood trend lines, journal entry summaries, skill practice records — create shared reference points for the therapeutic conversation. This aligns directly with Bordin's emphasis on mutual agreement about goals and tasks.

The fourth mechanism is reduced administrative burden. When between-session data collection is automated and organized by technology, therapists can spend more session time on actual therapeutic work rather than information gathering. Empath's approach of providing AI-generated session preparation summaries exemplifies this principle — the technology handles data aggregation so the therapist can focus on the relational and clinical work that requires human judgment and empathy.

Research by Clough and Casey confirmed that client receptivity to technology in therapy depends heavily on how the therapist frames it. When digital tools are presented as extensions of the therapeutic relationship — "This will help me understand your experience better between sessions so our work together is more focused and useful" — clients respond positively. When technology is imposed without explanation or perceived as replacing human contact, it damages the alliance. The therapist's intentionality in introducing and contextualizing digital tools is the critical variable.

When Technology Threatens the Alliance: Warning Signs

While the research is generally encouraging about technology's compatibility with the alliance, it is important to acknowledge scenarios where digital tools can genuinely harm the therapeutic relationship. The most significant risk is when technology substitutes for rather than supplements human connection. If a client experiences a between-session app as their primary therapeutic contact and in-session time as an afterthought, the alliance may suffer. Digital tools should amplify the in-session relationship, not compete with it.

Another risk is data surveillance anxiety. Clients who feel monitored rather than supported may withhold authentic emotional material — the opposite of what between-session data collection is intended to accomplish. This risk is particularly relevant for clients with histories of controlling relationships, institutional surveillance, or violations of privacy. Therapists need to assess each client's readiness for data-sharing tools individually and respect clients who prefer not to use them.

Over-reliance on data at the expense of clinical intuition is a third risk. If a therapist begins treating outcome scores as the sole arbiter of therapeutic progress, they may miss important qualitative shifts that the data does not capture — emerging self-compassion, subtle relationship improvements, or shifts in how a client relates to their own experience. Data informs clinical judgment; it does not replace it. The alliance is fundamentally a human relationship, and reducing it to metrics misunderstands its nature.

A fourth consideration is the digital divide. Not all clients have equal access to smartphones, reliable internet, or comfort with technology. Introducing digital tools without considering these factors can inadvertently communicate that the therapy is designed for a different kind of client, which undermines the bond component of the alliance. Offering alternatives and adapting the approach to each client's technological comfort and access is essential for maintaining an inclusive therapeutic relationship.

Alliance Rupture and Repair in Technology-Mediated Therapy

Jeremy Safran and Christopher Muran's influential work on alliance ruptures and repair processes provides an important lens for understanding technology's role in the alliance. They identified two types of ruptures: withdrawal ruptures (where the client disengages or becomes compliant rather than authentic) and confrontation ruptures (where the client expresses dissatisfaction or hostility toward the therapist or the therapy process). Both types can be triggered by the introduction of technology if it is experienced as misattuned or invalidating.

The encouraging finding from Safran and Muran's research is that repaired ruptures often lead to stronger alliances than relationships where ruptures never occurred. If a client expresses discomfort with a digital tool — "I feel like you care more about the app data than about what I am actually experiencing" — and the therapist responds with genuine curiosity and validation rather than defensiveness, the resulting conversation can deepen the alliance significantly. The technology itself becomes the occasion for a relationally meaningful exchange.

Research on routine outcome monitoring supports this dynamic. The Session Rating Scale, when it captures a dip in the client's alliance rating, provides an early warning system for ruptures. A therapist who notices a decreased SRS score and addresses it directly — "I see your rating of our session was lower today. I would really like to understand what was not working for you" — demonstrates the responsiveness that is central to rupture repair. Digital tools that surface these signals can actually make therapists more alert to alliance threats than they would be relying on clinical impression alone.

The practical lesson is that introducing digital tools into therapy is not a one-time decision but an ongoing relational process. Regular check-ins about how the client experiences the technology — whether it feels helpful, intrusive, neutral, or problematic — should be part of the therapeutic conversation. This ongoing negotiation is itself an expression of the alliance: it communicates that the client's experience matters, that the therapeutic frame is flexible, and that the therapist prioritizes the relationship above any specific tool or technique.

What Clients Actually Want From Digital Tools in Therapy

Research on client preferences for technology in therapy reveals patterns that should guide how we integrate digital tools. A systematic review by Lattie and colleagues found that clients generally prefer digital mental health tools that are personalized, easy to use, and clearly connected to their therapeutic goals. Generic wellness apps with no connection to the therapy process were viewed with indifference or skepticism. Tools that felt like an extension of the therapeutic relationship — carrying the therapist's voice and attention into daily life — were valued most highly.

Clients also consistently report wanting control over what they share and when. The ability to journal or track mood without being required to share every entry preserves a sense of autonomy that is important for both the alliance and the therapeutic process. Many clients appreciate having the option to flag specific entries or data points for their therapist's attention while keeping other reflections private. This selective sharing mirrors the natural process in therapy where clients choose what to bring to session.

Ease of use is a non-negotiable factor. Research on technology adoption in healthcare consistently shows that even clinically excellent tools fail if they create friction in the user experience. Clients will engage with a simple, intuitive interface that takes 30 seconds to open and use; they will abandon a clunky system that requires multiple logins, confusing navigation, or excessive data entry. The best digital tools for therapy are the ones that feel effortless enough to use in the moments when clients most need them — during emotional activation, after a difficult interaction, or before bed when reflection comes naturally.

Perhaps most importantly, clients want to know that their therapist is actually using the data they provide. Nothing undermines engagement with a between-session tool faster than the realization that the therapist is not reviewing the entries. When clients see that their between-session reflections inform the session — when the therapist references a journal entry or asks about a mood pattern — it validates the effort and reinforces the sense that the therapeutic relationship extends beyond the 50-minute hour.

Practical Guidelines for Preserving Alliance While Adopting Digital Tools

First, introduce any digital tool as a collaborative decision, not a unilateral clinical directive. Explain why you are suggesting it, how it connects to the therapeutic goals you have established together, and what the client can expect. Invite questions and concerns. Make it clear that the tool is optional and that declining to use it will not affect the therapeutic relationship or the quality of care. This approach aligns with Bordin's emphasis on goal and task agreement as foundations of the alliance.

Second, review the data with your client, not just about your client. Between-session data should be a shared resource that informs collaborative discussion, not private intelligence that the therapist uses to form interpretations without the client's input. When you reference between-session data in session, do it transparently: "I was looking at your mood tracking this week, and I noticed a pattern I wanted to explore with you." This positions the data as belonging to the therapeutic relationship rather than to the therapist alone.

Third, regularly check in about the client's experience with the digital tool. Is it helpful? Burdensome? Neutral? Does it feel like an extension of the therapy or a separate obligation? These check-ins serve dual purposes: they provide information about whether the tool is working, and they communicate that the client's experience and preferences matter more than any specific technological implementation. Be willing to modify or discontinue the tool based on client feedback.

Fourth, remember that the alliance is your primary instrument. If at any point a digital tool appears to be creating distance, resentment, or compliance rather than genuine engagement, address it directly and prioritize the relationship. No tool — no matter how sophisticated or evidence-based — is worth compromising the therapeutic alliance that research consistently identifies as the single most robust predictor of therapy outcomes. The technology serves the relationship, never the other way around.

Frequently asked questions

Does technology weaken the therapeutic alliance?

Not inherently. Flückiger and colleagues' 2018 meta-analysis found that alliance-outcome correlations in internet-delivered treatments (r = .275) were nearly identical to face-to-face therapy (r = .278). What matters is not the presence of technology but how it is introduced, framed, and integrated into the therapeutic relationship. Technology that enhances attunement and continuity can strengthen the alliance.

What is Bordin's model of the therapeutic alliance?

Edward Bordin's 1979 model proposes that the therapeutic alliance consists of three components: agreement on goals (what therapy aims to achieve), agreement on tasks (the specific activities and methods of therapy), and the emotional bond between therapist and client (trust, warmth, mutual respect). This pan-theoretical model remains the dominant framework in alliance research.

Can clients form strong therapeutic alliances through digital platforms?

Yes. Thomas Berger's research at the University of Bern found that client-rated alliance scores in internet-based interventions were roughly equivalent to those in face-to-face therapy, regardless of communication modality, diagnostic group, or amount of therapist contact. Clients are capable of forming meaningful therapeutic connections through diverse digital channels.

How should I introduce digital tools to clients without damaging the alliance?

Frame the tool as a collaborative decision that serves shared therapeutic goals. Explain the rationale, invite questions, and make it clear that the tool is optional. Present it as an extension of the therapeutic relationship — a way for you to understand their experience better between sessions — rather than as an administrative requirement or monitoring device.

What if a client does not want to use digital tools?

Respect their preference completely. Not all clients are comfortable with technology, and some may have specific concerns about privacy, surveillance, or the digital divide. The alliance is more important than any tool. Offer alternatives such as paper-based tracking or simply maintaining the standard session format, and reassure the client that their care will not be affected by their choice.

How much does the therapeutic alliance actually predict outcomes?

The alliance accounts for approximately 7.7 percent of outcome variance, with a correlation of r = .278 (equivalent to d = 0.579). While this may seem modest in percentage terms, it is among the most consistent and robust predictors in the entire psychotherapy research literature, comparable to or exceeding the effect of specific therapeutic techniques.

What are alliance ruptures and how do digital tools relate to them?

Alliance ruptures are breakdowns in the therapeutic relationship, classified by Safran and Muran as withdrawal (client disengages) or confrontation (client expresses dissatisfaction). Digital tools can trigger ruptures if experienced as misattuned or intrusive. However, ruptures that are recognized and repaired often lead to stronger alliances. Digital tools like the Session Rating Scale can actually help therapists detect ruptures earlier.